CADTH Reimbursement Review

Belzutifan (Welireg)

Sponsor: Merck Canada Inc.

Therapeutic area: von Hippel-Lindau disease–associated tumours

This multi-part report includes:

Clinical Review

Pharmacoeconomic Review

Ethics Review

Stakeholder Input

Clinical Review

Executive Summary

An overview of the submission details for the drug under review is provided in Table 1.

Table 1: Background Information of Application Submitted for Review

CNS = central nervous system; NOC = Notice of Compliance; pNET = pancreatic neuroendocrine tumour; RCC = renal cell carcinoma.

Sources: Sponsor’s Summary of Clinical Evidence¹ and product monograph.²

Introduction

Von Hippel-Lindau (VHL) disease, which is an inherited, autosomal-dominant neoplasia syndrome caused by a germline mutation and/or deletion of the VHL gene, is associated with a variety of neoplasms, such as hemangioblastomas of the central nervous system (CNS) and retina, renal cysts and clear-cell renal cell carcinomas (RCCs), pheochromocytomas, pancreatic cysts and neuroendocrine tumours (pNETs), epididymal and broad ligament cystadenomas, and endolymphatic sac tumours.^3,4 VHL disease affects 1 in 36,000 live births.^4,5 Approximately 20% of the cases are caused by de novo mutations and do not have a family history of VHL disease.^3,6 The prevalence is estimated to be 1 in 53,000 individuals.⁴ In Canada, the estimated number of cases is 727.¹ VHL disease is typically diagnosed through genetic testing to identify a germline-pathogenic variant in the VHL gene.⁷ People with VHL disease can have tumours involving multiple organs several times in their life and their symptoms will depend on the location and size of the tumours.^5,8,9 CNS hemangioblastomas develop in 60% to 84% of patients, pancreatic tumour or cysts in 35% to 70% (pNETs in 11% to 17%) and renal cysts and carcinomas in 69% (RCCs in 24% to 45%) of patients with VHL during their lifetime at mean ages of 29, 36, and 39 years, respectively.^3,6,10 Tumours associated with VHL disease have the potential to metastasize, and an RCC constitutes the most frequent metastasizing tumour.³

Active surveillance until treatment is indicated is currently the standard treatment for VHL disease–associated tumours. The goal of active surveillance is to find and remove tumours as early as possible before they affect the patient’s health. Surgical resection is indicated for tumours with a high symptom burden or those carrying a high risk of organ dysfunction or metastasis.^10,11 Certain tumours can be treated with radiation.¹² However, clinical experts consulted by CADTH for this review indicated that treatments, such as surgery and irradiation, can be morbid. Both patients and clinicians identified a need for a systemic treatment that can effectively treat VHL disease while causing less harm.

The objective of this report is to review and critically appraise the evidence submitted by the sponsor on the beneficial and harmful effects of belzutifan (40 mg oral tablets) in the treatment of adult patients with a nonmetastatic RCC, CNS hemangioblastoma, or nonmetastatic pNET, not requiring immediate surgery.

Stakeholder Perspectives

The information in this section is a summary of input provided by the patient and clinician groups who responded to CADTH’s call for input and from clinical experts consulted by CADTH for the purpose of this review.

Patient Input

The Canadian VHL Alliance, the Canadian Organization for Rare Disorders, Kidney Cancer Canada, Pancreatic Cancer Canada, and the Canadian Neuroendocrine Tumour Society provided 1 joint input for the treatment of adult patients with VHL disease who require therapy for associated nonmetastatic RCCs, CNS hemangioblastomas, and nonmetastatic pNETs not requiring immediate surgery. Patient input was gathered from online surveys and semistructured telephone interviews among patients living with VHL and their caregivers in December 2022. In total, 123 responses were gathered (72 from patients and 51 from caregivers), and 19 of those patients had experience with belzutifan.

Patients and caregivers described ongoing physical and psychological struggles due to VHL, such as dismissal or misdiagnosis for initial symptoms; not receiving a diagnosis until they had an advanced-stage tumour affecting vision, hearing, and walking; discomfort, pain, interference with daily activities; difficulties adhering to tumour-screening guidelines, scheduling tests, and travelling for tumour screenings; and out-of-pocket payment due to noncoverage by public or private health care insurance. Surgical resection was reported as the primary treatment for symptomatic lesions. The majority of 92 respondents described undergoing multiple surgeries at multiple sites, some with life-threatening risks and side effects. Out of 98 respondents, 18 (18.4%) reported having 10 or more surgeries, and the average number of surgeries reported was 5.3.

While evaluating the importance of outcomes of new treatments, patients from the survey emphasized the need for a treatment that can improve their physical condition by decreasing or stabilizing the size of tumours (weighted average rating = 4.8 on a scale from 1 not important to 5 extremely important), improving quality of life (weighted average rating = 4.63 on a scale from 1 not important to 5 extremely important), offering long-term stability or reduction of disease (weighted average rating = 4.86 on a scale from 1 not important to 5 extremely important), and offering the opportunity to avoid surgery (weighted average rating = 4.9 on a scale from 1 not important to 5 extremely important).

The patient groups suggested providing all VHL patients with access to belzutifan based on an individual informed decision between the treating physician and the patient and their family.

Clinician Input

Input From Clinical Experts Consulted by CADTH

The clinician input was provided by a panel of 3 clinical experts with expertise in treating VHL-associated RCC, CNS hemangioblastoma, and pNET from across Canada.

The clinical experts noted that prolonging survival and improving quality of life are critical goals for patients with VHL-associated nonmetastatic RCC, CNS hemangioblastoma, and/or nonmetastatic pNET. The current treatment paradigm for VHL disease involves genetic testing for VHL at diagnosis and active surveillance until treatment is indicated for associated tumours. Reactive treatments, such as surgery and radiation, can be morbid and are usually selected to respond to the conditions or symptoms developed. The clinical experts agreed that an effective systemic treatment would minimize the morbidity associated with surgical procedures in patients with VHL-associated nonmetastatic pNET and RCC, many of whom are younger. The clinical experts noted that belzutifan, if reimbursed, would be the first systemic treatment for VHL-associated tumours and would improve the current treatment paradigm by helping patients delay or avoid the need for local therapies (e.g., surgery and radiation).

The clinical experts indicated that VHL disease is rare and that all patients with VHL disease might benefit from belzutifan. The clinical experts did not specify a subset of the patient population that is in the most need or identify any prognostic factors that might cause differential treatment effects. The clinical experts noted that, before initiating treatment with belzutifan, genetic testing for VHL should be required. The clinicians also advised that a genetic counsellor should be involved in the diagnosis of VHL disease.

The clinical experts described several situations in which belzutifan may be discontinued, including intolerable adverse events (AEs), such as becoming transfusion-dependent due to anemia or clinical disease progression (i.e., worsening of symptoms). The clinical experts indicated that strict stopping criteria based on radiographic disease progression alone would not be reasonable if a patient is still experiencing a clinical benefit. The clinical experts noted that, due to the rarity of VHL disease, it is highly likely that only specialists working in large medical centres (e.g., tertiary and specialized referral centres) in Canada may encounter patients with VHL. Prescriptions therefore may be limited to specialists (medical oncologist and neuro-oncologists) working in these large centres.

Clinician Group Input

Clinician group input was received from the Ontario Health (Cancer Care Ontario) Genitourinary Cancer Drug Advisory Committee (7 clinicians), and a group of 25 clinical subspecialists in Canada involved caring for patients with VHL disease.

The clinician groups agreed with the clinical experts consulted by CADTH that belzutifan, the first systemic therapy option for VHL disease approved in Canada, fulfills an important unmet need for the treatment of patients with VHL and represents a shift in the current treatment paradigm. They also generally agreed on treatment goals, patient populations, assessing response and treatment-discontinuation criteria, and prescribing conditions.

While the clinical experts considered genetic testing to be a prerequisite for initiating treatment with belzutifan, neither clinician group indicated whether genetic testing for a VHL mutation or deletion was required. The clinician groups indicated that belzutifan should be discontinued if the patient is pregnant.

Drug Program Input

Input was obtained from the drug programs that participate in the CADTH reimbursement review process. The following were identified as key factors that could potentially affect the implementation of a CADTH recommendation for belzutifan:

• relevant comparators

• consideration for initiation of therapy

• consideration of discontinuation of therapy

• consideration for prescribing a therapy

• care provision issues

• system and economic issues.

The clinical experts consulted by CADTH provided advice on the potential implementation issues raised by the drug programs (Table 4).

Clinical Evidence

Pivotal Studies and Randomized Controlled Trial Evidence

Description of Study

One sponsor-conducted, phase II, single-arm, open-label trial (LITESPARK-004, N = 61) was identified from the systematic literature review (SLR) conducted by the sponsor. The primary objective of the LITESPARK-004 trial was to evaluate the efficacy of belzutifan (oral administration at a dosage of 120 mg once daily in 3 40 mg tablets until disease progression or unacceptable toxicity for the treatment of a VHL disease–associated nonmetastatic RCC as measured by objective response rate (ORR) and assessed by an independent review committee (IRC) following Response Evaluation Criteria in Solid Tumours Version 1.1 (RECIST 1.1). Secondary objectives included evaluation of the efficacy of belzutifan for the treatment of VHL disease–associated non-RCC tumours (CNS hemangioblastoma and nonmetastatic pNET), as well as assessment of the safety and tolerability (including AEs of special interest — anemia, hypoxia, secondary primary malignancies, and hepatic safety). In terms of efficacy end points, tumour response and durability of response were assessed by ORR and duration of response (DOR), respectively. Time-to-event outcomes, such as time to surgery (TTS), progression-free survival (PFS), and time to response (TTR) were also reported. The LITESPARK-004 trial also measured end points such as disease control rate (DCR), best overall response (BOR), linear growth rate (LGR), and the number of patients who developed metastases. No inferential statistical analyses were carried out in the LITESPARK-004 trial due to the single-arm study design, and data were summarized using descriptive statistics. The LITESPARK-004 trial is ongoing, and the data submitted by the sponsor to support this reimbursement request are based on a data cut-off date of April 1, 2022, as of which the median follow-up duration was 37.7 months (range = 4.2 to 46.1).

Participants eligible to be included in the LITESPARK-004 trial were required to be at least 18 years of age, have a diagnosis of VHL disease based on a germline VHL alteration, and have had at least 1 measurable RCC. Eligible patients could have other VHL disease–associated non-RCC tumours such as CNS hemangioblastoma and pNET. Patients who had an immediate need for surgical intervention for tumour treatment or evidence of metastatic disease were excluded. Efficacy results for RCC came from the total study population (n = 61), while efficacy results for CNS hemangioblastoma (n = 50) and pNET (n = 22) were from subsets of the total study population. At baseline, for the total study population (i.e., patients with RCC), the median age was 41.0 years (range = 19.0 to 66.0), with the majority being white (90.2%; 55 of 61), and the median age at time of VHL disease diagnosis was 32.0 years (range = 4.0 to 66.0). Characteristics were similar for the subpopulations of patients with CNS hemangioblastoma and pNET.

Efficacy Results

The efficacy end points that were noted to be important to patients and clinicians based on stakeholder input are summarized in Table 2. Results are as of the April 1, 2022, data cut-off date.

Time to Surgery

Median TTS was not reached for patients with VHL disease–associated nonmetastatic RCC, CNS hemangioblastoma, or nonmetastatic pNET at the data cut-off date. Seven patients with RCC (11.5%; 7 of 61), 1 patient with CNS hemangioblastoma (2.0%; 1 of 51), and none of the patients with pNET had surgery during the follow-up period.

Progression-Free Survival

The IRC-assessed median PFS was 39.2 months (95% confidence interval [CI], 38.5 to not evaluable) for patients with a VHL disease–associated nonmetastatic RCC. The median PFS was not reached for patients with a VHL disease–associated CNS hemangioblastoma and those with a nonmetastatic pNET by the data cut-off date.

IRC-assessed results showed that, among 61 patients with RCC at baseline, 11 (18.0%) had events (i.e., progressive disease or death), and 50 (82%) were censored, mostly due to no progression at the time of data cut-off or before the end of treatment (43; 70.5%).

IRC-assessed results showed that, among 50 patients with CNS hemangioblastoma at baseline, 11 (22.0%) had events (i.e., progressive disease or death), and 39 (78%) were censored (of which 34 [68%]) were due to no progression at the time of data cut-off.

All 22 patients with pNET were censored due to no progression at the time of data cut-off.

Objective Response Rate

At a median follow-up of 37.7 months, the IRC-assessed percentage of patients who had a complete response (CR) or partial response (PR) to belzutifan was 63.9% (39 of 61) among those with a VHL-associated nonmetastatic RCC, 44% (22 of 50) among those with a VHL-associated CNS hemangioblastoma, and 90.9% (20 of 22) among those with a VHL-associated nonmetastatic pNET.

Duration of Response

IRC-assessed median DOR was not reached for responders with a VHL disease–associated nonmetastatic RCC, CNS hemangioblastoma, or nonmetastatic pNET by the data cut-off date. For patients with RCC, 74.4% of responders (29 of 39) had a DOR of 18 months or longer, 56.4% (22 of 39) had a DOR of 24 months or longer, and 25.6% (10 of 39) had a DOR of 30 months or longer. For patients with a CNS hemangioblastoma, 63.6% of responders (14 of 22) had a DOR of 18 months or longer, 59.1% (13 of 22) had a DOR of 24 months or longer, and 54.5% (12 of 22) had a DOR of 30 months or longer. For patients with a pNET, 95.0% of responders (19/20) had a DOR of 18 months or longer, 75.0% (15 of 20) had a DOR of 24 months or longer, and 40.0% (8 of 20) had a DOR of 30 months or longer.

Among 39 RCC patients with a confirmed response, 32 (82.1%) were censored due to no progression at the time of data cut-off or before the end of treatment. Out of 22 patients with CNS hemangioblastoma who showed a confirmed response, 17 were censored due to no progression at the time of data cut-off or before the end of treatment. All 20 patients with a pNET who showed a confirmed response had progression at the time of data cut-off.

Harms Results

Adverse Events

Treatment-emergent adverse events (TEAEs) were reported in all 61 patients (100%) in the LITESPARK-004 trial. The most commonly reported TEAE was anemia (90.2%), followed by fatigue (73.8%), headaches (47.5%), dizziness (45.9%), and nausea (39.3%).

Serious Adverse Events

Serious adverse events (SAEs) were reported in 18 patients (29.3%).

Withdrawal Due to Adverse Events

Treatment discontinuation due to TEAEs (i.e., dizziness or intracranial hemorrhage) was reported in 4 patients, 2 of which were drug-related.

Mortality

Two patients died during the study, 1 due to acute toxic effects of fentanyl and 1 due to suicide.

Notable Harms

As of the data cut-off, 90.2% (55 of 61) patients had 229 episodes of anemia. The average number of episodes of anemia for each patient was 4.2. One (1.6%) and 2 patients (3.3%) developed hypoxia and secondary primary malignancies, respectively. None of the participants had drug-induced liver injury.

Table 2: Summary of Key Results from Pivotal Studies and RCT Evidence

CI = confidence interval; CNS = central nervous system; CR = complete response; IRC = independent review committee; NE = not evaluable or not estimable; NR = not reported; pNET = pancreatic neuroendocrine tumour; PR = partial response; RCC = renal cell carcinoma; RCT = randomized controlled trial.

^aAccording to the sponsor’s explanation, the “+” sign indicates that the objective response in the respective participant remained ongoing, with no progressive disease or death reported at the time of data cut-off. For example, “5.4+” means that, at the data cut-off, the respective participant had been in response for 5.4 months and had not experienced disease progression or death.

^bOne death was due to acute toxic effects of fentanyl, and 1 was due to suicide.

^cCriteria for drug-induced liver injury: alanine transaminase or aspartate transaminase levels equal to or greater than 3 times the upper limit of normal range, and bilirubin levels equal to or twice the upper limit of normal range, and alkaline phosphatase levels of less than twice the upper limit of normal range.

Source: LITESPARK-004 Clinical Study Report.¹³

Critical Appraisal

Internal Validity

The LITESPARK-004 trial was a phase II, single-arm, open-label clinical trial. Given the rarity of VHL disease and that active surveillance is the current stand of care for patients with VHL-associated nonmetastatic tumours, the single-arm design and small sample size were considered appropriate from the regulatory perspective to assess the efficacy and safety of belzutifan. However, the absence of an internal comparison group in the single-arm LITESPARK-004 trial hampered the ability to establish a causal link with the efficacy or safety outcomes observed in patients.

The LITESPARK-004 trial explicitly defined the hypotheses (i.e., a null hypothesis of an ORR of 15% or lower with an alternate hypothesis of ORR of 30% or higher), which were considered clinically meaningful by the clinical experts consulted by CADTH. The selection of ORR (defined as the sum of CR and PR according to RECIST 1.1) to measure antitumour activity and DOR (in patients with CR or PR) to determine the durability of tumour response were appropriate. Additional time-to-event end points (TTS and PFS) were employed in the LITESPARK-004 trial and were considered by the clinical experts to be critical outcomes to assess the efficacy of belzutifan. However, randomized controlled trials (RCTs) are preferred over single-arm studies for time-to-event end points such as PFS due to their sensitivity to baseline differences in patient, disease, and other clinical characteristics, and interpreting results without a randomized reference could be problematic.^14-16 The LITESPARK-004 trial also involved outcomes pre- and posttreatment with belzutifan (i.e., a change in LGR) to demonstrate the efficacy of belzutifan. However, without formal statistical analysis, the role of chance could not be ruled out. With respect to outcome measurement, in addition to study investigators, an IRC was used to assess radiographic outcomes and reduce the risk of bias in measuring outcomes for most of the efficacy end points in patients with a VHL-associated nonmetastatic RCC and those with a VHL-associated CNS hemangioblastoma.

Altogether, due to a lack of comparison groups and a lack of formal inferential statistical analyses, no definitive conclusions could be drawn from the LITESPARK-004 trial about the efficacy and safety of belzutifan in patients with VHL-associated nonmetastatic RCC, CNS hemangioblastoma, or nonmetastatic pNET, all of whom did not require immediate surgery.

External Validity

All participants in the LITESPARK-004 trial were required to have at least 1 RCC. The LITESPARK-004 trial therefore may not reflect results for participants with only CNS hemangioblastomas and/or pNETs. However, the clinical experts consulted by CADTH did not consider this a serious generalizability issue. According to the clinical experts, the inclusion and exclusion criteria of the LITESPARK-004 trial in general were aligned with selection criteria in the Canadian settings when identifying suitable candidates for belzutifan. However, the clinical experts noted that the requirement for patients to have an Eastern Cooperative Oncology Group Performance Status (ECOG PS) of 0 or 1 might not be necessary to initiate belzutifan in clinical practice because ECOG PS can be unstable and subjective. There were no study sites in Canada as the LITESPARK-004 trial was conducted only in Denmark, France, the UK, and the US. More than 90% of the patients in LITESPARK-004 were white, which is not representative of the racial profile of the Canadian patient population, according to the clinical experts. The dosing and administration of belzutifan in the LITESPARK-004 trial were consistent with the product monograph. The clinical experts commented that concomitant medications and procedures in the LITESPARK-004 trial were also appropriate and commonly used in the Canadian settings. The outcomes of TTS, PFS, ORR, and DOR are commonly used in clinical trials of anticancer therapy and are relevant to clinical practice, according to the clinical experts. These outcomes are also important to patients, who indicated they want treatments that offer the opportunity to avoid surgery, decrease or stabilize the size of tumours, and result in long-term stability or reduction of disease. However, LGRs have not been commonly adopted in clinical practice and may not correlate with clinical benefit. The clinical expert specializing in CNS hemangioblastoma noted that, in clinical practice, Response Assessment in Neuro-Oncology (RANO) criteria are used instead of RECIST 1.1 to assess tumour responses in patients with CNS hemangioblastomas. The LITESPARK-004 trial did not assess some outcomes that are important to patients, such as symptoms and health-related quality of life (HRQoL).

Indirect Comparisons

Description of Study

A sponsor-conducted indirect treatment comparison (ITC) was submitted to address the absence of comparative evidence of belzutifan for the treatment of adult patients with VHL disease in the LITESPARK-004 trial. The ITC compared a real-world, retrospective, noninterventional cohort study of existing medical records of VHL patients managed and treated at the National Cancer Institute (NCI) in the US or in Canada (the VHL Natural History Study) with the LITESPARK-004 trial. Patients with VHL-associated RCC from the VHL Natural History Study were reweighted to match the distribution of key baseline characteristics among patients with VHL-associated RCC in LITESPARK-004 and compared using a matching-adjusted indirect comparison (MAIC). The comparative treatment was active surveillance. The primary and only outcome assessed was time to RCC-related surgery.

Efficacy Results

The weekly exponential rate of RCC surgery was estimated at 0.00487 (standard error [SE] = 0.00034) in the matched Natural History Study sample versus 0.00071 (SE = 0.0003) in the LITESPARK-004 population.

Critical Appraisal

Findings from the sponsor-conducted ITC, which used the VHL Natural History Study as an external comparator for the LITESPARK-004 trial, were considered highly uncertain. Although the estimated decrease in the rate of surgeries in the LITESPARK-004 trial relative to the VHL Natural History Study was large, several major limitations decreased CADTH’s confidence in the results. First, the selection criteria that informed the VHL Natural History Study subcohort used in the ITC was intended to match with those from the LITESPARK-004 cohort but did not include some key criteria. Specifically, the Natural History Study cohort did not include restrictions on ECOG PS scores of 0 or 1; this difference created a risk of bias in effectiveness that may favour belzutifan. Second, it was difficult to assess the degree of heterogeneity between the included studies based on the sponsor-provided technical report because the reporting of the study design and patient characteristics were limited. It is likely that the underlying assumption of the unanchored MAIC — that all potential prognostic and effect-modifying factors were balanced across groups — was violated, which would result in a high risk of confounding. Third, the outcome definition of RCC surgery within the Natural History Study cohort is subject to potential measurement error. Specifically, the clinical experts indicated that renography and cyst removal are not definitive surgical interventions for the management of RCC, although these were considered an RCC surgery outcome in the VHL Natural History Study. The magnitude of bias due to measurement error is unknown but may overestimate the estimated relative rate of RCC surgeries in favour of belzutifan. Finally, the analysis did not provide information specific to VHL-associated CNS hemangioblastoma and pNET populations.

Studies Addressing Gaps in the Pivotal and Randomized Controlled Trial Evidence

The sponsor submitted the VHL Natural History Study¹⁷ to fill the gap in published clinical trials or observational data on the efficacy outcomes of the standard of care (active surveillance). This study provided active-surveillance efficacy data used for the ITC. The sponsor also provided a cross-sectional HRQoL survey to address the lack of HRQoL or utility values; it is summarized and critically appraised in Appendix 1. This study assessed the impact of VHL disease on HRQoL as measured by the EQ-5D questionnaire in patients with RCC, CNS hemangioblastoma, or pNET. A total of 220 patients completed the survey. Overall, patients with VHL-associated tumours had a mean EQ-5D score of 0.771. As. patients who participated in this study were not treated with belzutifan, this study does not provide information related to the effect of treatment with belzutifan on the HRQoL of patients with VHL disease.

VHL Natural History Study

Description of Study

The VHL Natural History Study,¹⁷ a retrospective real-world cohort study of growth kinetics and surgical patterns in patients with VHL disease and associated renal solid tumours, was conducted using data registered by the NCI in a hereditary database of patients with VHL syndrome. The primary study population consisted of US and Canadian patients treated at the NCI with confirmed VHL syndrome and ≥ 1 renal solid tumour with available measurement(s) during the study period (July 31, 2004, to June 30, 2020). Additional criteria were applied in an attempt to match the study population more closely to the population enrolled in the LITESPARK-004 trial.

Of 776 patients with VHL disease in the NCI hereditary database, 308 with at least 1 solid renal tumour met the eligibility criteria and were included in the primary study population. After applying additional eligibility criteria focusing on the tumour growth rate assessment, 247 patients (80.2%) were included in the trial population subgroup.

Subgroups of 131 patients and 114 patients in the primary study population and the trial population subgroup, respectively, had 3 or more serial measurements of at least 1 solid renal tumour during the study period that qualified them for inclusion in the LGR analysis subgroups to address the primary research objective.

Efficacy Results

Linear Growth Rate

The median tumour-level LGRs for the primary study population and trial population subgroup were 0.38 cm per year (interquartile range [IQR] = 0.30 to 0.49) and 0.37 cm per year (IQR = 0.29 to 0.47), respectively.

Frequency and Type of Tumour–Reduction Procedures

Of the 308 patients in the primary study population, 232 patients (75.3%) had at least 1 renal solid tumour–reduction procedure during the study period, including 225 patients (73.1%) with surgical procedures (96% of which were partial nephrectomies), 16 patients (5.2%) with ablation procedures, and 1 patient (0.3%) who received radiation. In the trial population subgroup, 184 patients (74.5%) underwent at least 1 renal solid tumour–reduction procedure. The median number of tumour-reduction procedures per patient in the trial population subgroup was 2 (range = 1 to 9).

Time to Tumour–Reduction Procedures

The 1-year, 2-year, 5-year, and 7-year intervention-free survival probabilities for the first tumour-reduction procedure were 79.0%, 69.9%, 38.3%, and 26.4%, respectively, for the trial population subgroup. The median time to first tumour-reduction procedure was 44.2 months (95% CI, 35.74 to 49.51) in the trial population subgroup.

Harm Results

Of the 217 patients (70.5%) in the primary study population with at least 1 partial nephrectomy, 413 partial nephrectomies were performed during follow-up, 124 (30.0%) of which were associated with complications. The median estimated blood loss, assessed among all surgical procedure types, was 1.5 L (IQR = 0.6 to 2.6). Two patients (0.9%) in the primary study population with at least 1 renal tumour-reduction procedure conducted at the National Institutes of Health died within 30 days of the procedure (1 nephrectomy and 1 biopsy).

Critical Appraisal

Internal Validity

This real-world retrospective cohort study did not evaluate the effect of belzutifan and did not provide evidence about the efficacy of the treatment. The presented analyses appeared to include no a priori protocol. The study was conducted using data registered by the NCI in a hereditary database of patients with VHL syndrome, but how these data were located (e.g., which search methods were used) or selected is not specified.

Some of the limitations of this study were the high level of missing data for some variables; the unavailability of longitudinal measures of tumour growth for all tumours in a systematic manner (i.e., differences may have arisen due to variation in measurement across observers) because the measures were extracted from the registry directly; incomplete documentation of metastasis; and the possibility of misclassification. There is also a possibility of loss to follow-up of patients who left the registry but for whom no relevant information was provided. Attempts were made to ensure that the trial population subgroup would be similar to the population of the LITESPARK-004 trial, based on the additional criteria that were applied. However, several criteria in the Natural History Study cohort are approximate to those of the pivotal trial due to insufficient access to information within the database. While demographics and clinical characteristics were similar for patients in the trial population and LGR analysis subgroups, many demographic characteristics (i.e., race or ethnicity and ECOG PS) were not reported in the Natural History Study report. One of the exclusion criterion was receiving systemic oncologic therapy or investigational therapy within 30 days on or before the patient-level index date. However, no specific information was given about the type of systemic or investigational therapy. Moreover, this exclusion criteria may have led to the exclusion of patients with a better prognosis, therefore affecting the results for the trial population subgroup.

External Validity

This study includes patients managed and treated at the NCI only. This may affect the overall generalizability of the study, as it may not be representative of all patients with VHL in Canada that fall within the indication. While this study presented data for renal tumours, patients included in the study also had other tumours associated with VHL. However, no additional information was provided for these tumours, which may have resulted in the exclusion of some important clinical outcomes associated with VHL syndrome. Because the indication under review includes nonmetastatic pNETs and CNS hemangioblastomas in addition to nonmetastatic RCCs, the unavailability of data related to these 2 tumours represents a gap in the evidence provided by this study. The authors of this study noted the potential risk of losing a substantial number of samples due to the eligibility criteria of recruiting patients with 3 or more serial measurements for the assessment of tumour growth rate patterns.

Conclusions

Evidence from the phase II, single-arm, open-label LITESPARK-004 trial suggests that treatment with belzutifan may have a beneficial effect in adult patients with VHL-associated nonmetastatic RCC with or without VHL-associated CNS hemangioblastoma and/or nonmetastatic pNET not requiring immediate surgery, based on the opinions of clinical experts. The clinical experts indicated that the efficacy results from the patient populations with a VHL disease–associated nonmetastatic RCC, RCC and CNS hemangioblastoma, RCC and nonmetastatic pNET, as well as an RCC, CNS hemangioblastoma, and pNET, in the LITESPARK-004 trial demonstrated a promising and durable tumour response as measured by ORR and DOR that is likely attributable to belzutifan, based on their clinical experience and expectations of the natural history of the disease. However, the absence of a comparator group in the pivotal study precluded definitive conclusions about the efficacy and safety of belzutifan, due to the potential for confounding. The findings for other supportive outcomes (e.g., TTS and PFS) are uncertain due to the lack of an internal comparator group and no inferential statistical analyses. The incidence of harms events, such as anemia and hypoxia, were acceptable, according to the clinical experts. Because VHL disease is a life-long condition and patients can receive belzutifan for an extended period, longer follow-ups would be needed to better understand the long-term benefits and harms of belzutifan, as well as the potential residual effects following discontinuation. Findings from an ITC suggest that treatment with belzutifan may have a beneficial effect on time to RCC-related surgery compared to active surveillance in patients with a VHL disease–associated nonmetastatic RCC. However, the certainty of the indirect evidence was low due to violation of the underlying assumptions of the MAIC approach that resulted in a high risk of residual confounding, and important differences in outcome definitions between the LITESPARK-004 trial and the VHL Natural History Study that would bias the findings in favour of belzutifan. Based on the evidence in the ITC analysis, the comparative efficacy of belzutifan versus active surveillance for VHL-associated CNS hemangioblastoma and nonmetastatic pNET was not assessed and is therefore unknown. The VHL Natural History Study and the VHL HRQoL survey did not provide evidence on the efficacy and safety of belzutifan as standalone studies because patients in these studies did not receive belzutifan. It is unknown whether treatment with belzutifan would improve patient quality of life.

Introduction

The objective of this report is to review and critically appraise the evidence submitted by the sponsor on the beneficial and harmful effects of belzutifan (40 mg tablets for oral administration) in the treatment of adult patients with VHL disease who require therapy for associated nonmetastatic RCCs, CNS hemangioblastomas, or nonmetastatic pNETs not requiring immediate surgery.

Disease Background

The contents in this section were informed by materials submitted by the sponsor and clinical expert input. The following summaries of observations were validated by the CADTH review team.

VHL disease is an inherited, autosomal-dominant neoplasia syndrome caused by a germline mutation and/or deletion of the VHL gene.¹ Abnormal or absent VHL tumour-suppressor function leads to dysregulation of the degradation of hypoxia-inducible factor (HIF), a key mediator of cellular adaptation to hypoxia.¹ In VHL disease, HIF accumulates and its activity increases, leading to abnormal growth of both nonmalignant and malignant tumours in multiple organ systems.¹¹ VHL disease is therefore a familial cancer-predisposition syndrome associated with a variety of neoplasms, such as hemangioblastomas of the CNS and retina, renal cysts and clear-cell RCCs, pheochromocytomas, pancreatic cysts and pNETs, epididymal and broad ligament cystadenomas, and endolymphatic sac tumours.^3,4

A single mutated copy of the VHL gene is not enough to cause VHL disease, and a mutation in the other copy of the gene must occur.⁵ However, almost everyone born with 1 VHL mutation will eventually acquire a mutation in the second copy of the gene and develop VHL disease.⁵ VHL disease affects males and females and all ethnic groups equally.⁵ Tumours associated with VHL disease typically appear first in young adulthood but are also seen in children.⁵ The mean age at diagnosis is 26 years and more than 90% of patients have developed symptoms by the age of 65 years.^4,10,18

VHL disease affects 1 in 36,000 live births.^4,5 Approximately 20% of the cases are caused by de novo mutations and therefore do not have a family history of VHL.^3,6 The prevalence is estimated to be 1 in 53,000 individuals.⁴ There are 200,000 cases of VHL disease worldwide and 10,000 cases in the US.^4,5 In Canada, the estimated number of cases is 727.¹

A diagnosis of VHL disease is typically established by identifying a germline-pathogenic variant of the VHL gene through genetic testing.⁷ Occasionally, VHL disease may be diagnosed based on clinical criteria, such as patients with 1 VHL-associated lesion and a family history of VHL, or patients with multiple VHL-associated tumours but who do not have access to genetic testing.⁷

People with VHL disease can have tumours involving multiple organs several times in their life and their symptoms will depend on the location and size of the tumours.^5,8,9 CNS hemangioblastomas develop in 60% to 84% of patients, pancreatic tumours or cysts in 35% to 70% (pNET in 11% to 17%), and renal cysts and carcinomas in 69% (RCC in 24% to 45%) of patients with VHL during their lifetime at mean ages of 29, 36 and 39 years, respectively.^3,6,10 The most common first manifestation of VHL disease is a retinal hemangioblastoma followed by CNS hemangioblastoma and RCC.⁶ An RCC is the first manifestation of VHL disease in approximately 10% of cases.¹⁸ Tumours associated with VHL disease have the potential to metastasize, and an RCC is the most frequent metastasizing tumour.³ CNS hemangioblastomas, even though nonmalignant, are a major cause of morbidity and mortality in VHL disease due to frequent and multiple recurrent and/or progressive tumours throughout the CNS, affecting vital structures by compression or as surgical sequelae.⁹

Standards of Therapy

Contents within this section have been informed by materials submitted by the sponsor and clinical expert input. The following summaries have been validated by the CADTH review team.

VHL disease has no cure and how it will progress varies by the individual. Because of the progressive, diverse nature, and high frequency of multiple neoplasms in various organ systems in patients with VHL disease, the management of a single tumour type is complicated by the presence of others.¹⁰ Active surveillance of VHL-associated tumours until treatment is the standard approach. The goal of active surveillance is to find, monitor, and remove tumours if resectable and clinically indicated before they affect the patient’s health. Surgical resection is indicated for tumours with a high symptom burden or those carrying a high risk of organ dysfunction or metastasis.^10,11 Certain tumours can be treated with radiation therapy.¹²

Localized CNS hemangioblastoma, RCC, and pNET masses in patients with VHL are usually monitored with active surveillance until surgical resection is indicated, often when the largest tumour reaches a size of 3 cm or larger, when tumours exhibit rapid radiologic progression, or in the presence of significant symptoms, according to the literature.^10,11,19 The clinical experts consulted by CADTH noted that the tumour location, the presence of symptoms, and tumour growth rate are usually more important than tumour size in determining when surgical resection is required.

Vascular endothelial growth factor receptor inhibitors (sunitinib, pazopanib, and dovitinib) were evaluated in small prospective clinical trials of patients with VHL disease to test their ability to alter the growth of tumours.^20-22 However, due to TEAEs, and limited responses in non-RCC tumours such as CNS hemangioblastomas, treatment with drugs that target vascular endothelial growth factor receptors is not recommended as standard of care in VHL disease, and none of these treatments received regulatory approval for VHL disease in Canada.¹¹ The clinical experts consulted by CADTH noted that belzutifan, if reimbursed, would be the first systemic treatment for VHL-associated tumours.

Drug Under Review

Key characteristics of belzutifan are summarized in Table 3.

Belzutifan is an inhibitor of the hypoxia-inducible factor 2 alpha subunit (HIF-2-alpha).^2,23-25 Belzutifan binds to HIF-2-alpha and, in conditions of hypoxia or impairment of VHL protein function, belzutifan blocks the interaction between HIF-2-alpha and the hypoxia-inducible factor 1 beta subunit, leading to reduced transcription and expression of HIF-2-alpha downstream target genes associated with tumourigenic effects.

Belzutifan was reviewed under Project Orbis and received a Notice of Compliance from Health Canada on July 11, 2022, for the treatment of adult patients with VHL disease who require therapy for associated nonmetastatic RCC not requiring immediate surgery. On July 11, 2023, Health Canada–approved an expanded indication for belzutifan for the treatment for adult patients with VHL disease who require therapy for associated CNS hemangioblastomas or nonmetastatic pNETs not requiring immediate surgery. To date, the FDA²⁶ and the Therapeutic Goods Administration of Australia²⁷ have approved the use of belzutifan for the treatment of adult patients with VHL disease who require therapy for associated nonmetastatic RCCs, CNS hemangioblastomas, or nonmetastatic pNETs not requiring immediate surgery.

This is the first review of belzutifan by CADTH. The sponsor’s reimbursement request was aligned with the indication. Belzutifan is supplied as 40 mg tablets for oral administration.² The recommended belzutifan dose is 120 mg (three 40 mg tablets) administered once daily.² The product monograph recommends that treatment should continue until disease progression or unacceptable toxicity occurs. No dose adjustment is needed in older patients nor in the presence of renal (mild to moderate) or hepatic (mild) impairment.

Table 3: Key Characteristics of Belzutifan

CNS = central nervous system; HIF-1-beta = hypoxia-inducible factor 1 beta subunit; HIF-2-alpha = hypoxia-inducible factor 2 alpha subunit; pNET = pancreatic neuroendocrine tumour; RCC = renal cell carcinoma; VHL = von Hippel-Lindau.

Source: Belzutifan (Welireg) draft product monograph.²⁸

Stakeholder Perspectives

Patient Group Input

This section was prepared by the CADTH review team based on the input provided by patient groups. The full original patient input received by CADTH is included in the stakeholder section at the end of this report.

The Canadian VHL Alliance, Canadian Organization for Rare Disorders, Kidney Cancer Canada, Pancreatic Cancer Canada, and Canadian Neuroendocrine Tumour Society provided a single joint input for the treatment of adult patients with VHL disease who require therapy for associated nonmetastatic RCCs, CNS hemangioblastomas, and nonmetastatic pNETs not requiring immediate surgery. Patient input was gathered from online surveys and semistructured telephone interviews among patients living with VHL disease and their caregivers in December 2022. In total 123 responses were gathered (72 from patients and 51 from caregivers), and 19 patients had experience with belzutifan.

Patients and caregivers described their ongoing physical and psychological struggles due to VHL, such as dismissal or misdiagnosis of initial symptoms; not receiving a diagnosis until they had an advanced-stage tumour affecting their vision, hearing, and walking; discomfort, pain, interference with daily activities; difficulties adhering to tumour-screening guidelines, scheduling tests and travelling for tumour screenings; and paying all the expenses not covered by public or private health care insurance. Patients from the survey reported worrying about potential adverse effects of frequent screenings and the stress caused by the uncertainty of tumour activity between scans. Patients also described their experiences with VHL as “excruciatingly painful and debilitating,” diminishing their ability to care for themselves, engage in family and social activities, plan for a family, and participate in work. Patients also reported that VHL disease affected their finances, feelings of self-worth, and engagement in meaningful activities. Overall, about 45% of respondents reported a negative impact of VHL on their ability to work, travel, exercise, perform household chores, spend time with family and friends, and fulfill family obligations.

Survey respondents reported that surgical resection was the primary treatment received for symptomatic lesions. When patients and caregivers were asked about how surgeries have affected their ability to work or quality of life, 92 respondents, some of whom were taking belzutifan, provided commentary. When respondents were asked about the number of surgeries they had undergone to remove VHL cysts and tumours, 98 individuals provided a response. Of these, 18 (18.4%) reported having 10 or more surgeries, and the average number of surgeries reported was 5.3.

While evaluating the importance (on a scale of 1 [not important] to 5 [extremely important]) of outcomes of new treatments, patients from the survey emphasized the need for a treatment that can improve their physical condition by decreasing or stabilizing the size of tumours (weighted average rating = 4.8), improving quality of life (weighted average rating = 4.63), offering long-term stability or reduction of disease (weighted average rating = 4.86), and offering the opportunity to avoid surgery (weighted average rating = 4.9). Patients also expressed their willingness to tolerate moderate-to-severe side effects of a treatment that would allow them to avoid surgery to remove VHL-related cysts and tumours (weighted average rating = 3.38).

While describing their experiences with belzutifan, the patient groups suggested that all patients with VHL disease should be given access to belzutifan based on an individual informed decision between the treating physician and the patient and their family. The patient groups also emphasized that the decision to take belzutifan should consider the clinical status of the patient; the location, severity, and prognosis of the tumours; the history of surgery and outcomes; the psychological and mental health status of the patient; and other factors.

Clinician Input

Input From Clinical Experts Consulted by CADTH

All CADTH review teams include at least 1 clinical specialist with expertise in the diagnosis and management of the condition for which the drug is indicated. Clinical experts are a critical part of the review team and are involved in all phases of the review process (e.g., assisting in the critical appraisal of clinical evidence, interpreting the clinical relevance of the results, and providing guidance on the potential place in therapy). In addition, as part of the belzutifan review, a panel of 3 clinical experts with expertise in treating VHL-associated RCCs, CNS hemangioblastomas, and pNETs from across Canada was convened to characterize unmet therapeutic needs, help identify and communicate situations in which there are gaps in the evidence that could be addressed by collecting additional data, promote the early identification of potential implementation challenges, gain further insight into the clinical management of patients living with VHL disease, and explore the potential place in therapy of belzutifan (e.g., potential reimbursement conditions). A summary of this panel discussion is presented in the sections that follow.

Unmet Needs

The clinical experts noted that prolonging survival and improving quality of life are critical goals for patients with a VHL-associated nonmetastatic RCC, CNS hemangioblastoma, and/or nonmetastatic pNET. The clinical experts also highlighted the need to improve function, preserve neurologic status, and address symptoms related to the burden of disease unique to patients with VHL-associated CNS tumours, as well as the need to delay progression to a metastatic RCC and increase the time to development of dialysis dependence for patients with VHL-associated RCC.

The clinical experts agreed that an effective systemic treatment for VHL is needed. According to the clinical experts, an effective systemic treatment would minimize the morbidity associated with surgical procedures in patients with a VHL-associated nonmetastatic pNET and RCC, many of whom are younger. An effective systemic treatment would also address neurologic disability and mortality related to growing symptomatic CNS lesions, which may be surgically inaccessible at recurrence and can only be re-resected and/or irradiated a limited number of times. The clinical experts noted that VHL disease is a multiorgan disease that is frequently highly symptomatic, requiring frequent surgical interventions, radiation, specialist review, and opioid analgesia, among other health care service interactions.

Place in Therapy

The clinical experts noted that the current treatment paradigm for VHL disease involves genetic testing for VHL at diagnosis and active surveillance until treatment is indicated for associated tumours. The clinical experts indicated that reactive treatments, such as surgery and radiation, can be morbid and are usually selected to respond to the conditions or symptoms developed. The clinical experts noted that belzutifan, if reimbursed, would be the first systemic treatment for VHL-associated tumours and would change the current treatment paradigm by helping patients delay or avoid the need for local therapies (e.g., surgery and irradiation).

Patient Population

The clinical experts indicated that VHL is a rare disease and that all patients with VHL might benefit from belzutifan. However, the clinical experts noted that the eligibility criteria of the pivotal LITESPARK-004 trial matched the patient population of the indication and reimbursement request, which is for patients with a nonmetastatic RCC, CNS hemangioblastoma, or nonmetastatic pNET not requiring immediate surgery. It was noted that all patients enrolled in the LITESPARK-004 trial would have at least 1 measurable RCC according to RECIST 1.1. The clinical experts did not specify any subset of the patient population with the greatest need or identify any prognostic factors that might cause differential treatment effects.

The clinical experts considered size to be an appropriate inclusion criterion of the LITESPARK-004 trial (i.e., no RCC tumours greater than 3 cm that necessitated immediate surgical intervention). The clinical experts advised against setting an upper limit for the size of the tumours for patients to be eligible for treatment with belzutifan, noting that a larger size is not the sole factor influencing management; tumour location, presence of symptoms, and tumour growth rate are usually equally important factors.

With respect to how VHL disease is diagnosed in Canada, the clinical experts indicated that several circumstances could trigger the need for screening and genetic testing for VHL, such as young patients with a pNET, patients with more than 1 cancer type (e.g., RCC, CNS hemangioblastoma, and pNET), and patients presenting with multiple CNS lesions. The clinical experts noted that genetic testing to confirm a VHL diagnosis should be required before initiating treatment with belzutifan. They also indicated that genetic testing for VHL is available in certain centres across Canada. The clinical experts identified multiple challenges related to genetic testing for VHL. Living in rural or remote areas may be a barrier to accessing the test, and greater awareness is needed among patients without an established family history of VHL disease so that these patients can be tested. The clinicians also noted that a genetic counsellor should be involved in the diagnosis of VHL disease.

Assessing the Response Treatment

The clinical experts noted that response assessments for CNS tumours in clinical practice typically involve the measurement of target lesions over time to determine if there is a CR, PR, stable disease, or disease progression using the RANO criteria.²⁹ The measurement is usually performed on the brain and spine using MRI every 3 months for symptomatic patients and up to yearly for patients under surveillance. Ideally, neurologic function is also assessed. The clinical experts reported that response assessments for pNET involve CT measurements every 2 or 3 months; however, they noted that clinical response or symptoms may be difficult to assess in patients with nonmetastatic disease. For RCC, response is assessed with CT and/or MRI every 3 to 6 months, depending on patient conditions as defined by RECIST 1.1.

Discontinuing Treatment

The clinical experts referred to several situations in which belzutifan may be discontinued, including intolerable AEs (e.g., becoming transfusion-dependent due to anemia), clinical disease progression (e.g., worsening of symptoms), and radiographic disease progression. The clinical experts indicated that they might continue treatment with belzutifan in some patients who have experienced radiographic progression without clinical progression if in their opinion the patient is still deriving clinical benefit. The clinical experts noted that patients who have remained in a stable status for an extended period may base a decision to discontinue belzutifan and be actively monitored on a personal preference.

Prescribing Considerations

The clinical experts noted that, due to the rarity and complex needs of patients with VHL disease, it is highly likely that only specialists working in large medical centres (e.g., tertiary and specialized referral centres) in Canada may encounter patients with VHL. Prescription may therefore be limited to specialists (e.g., medical oncologist and neuro-oncologists) working in such centres.

Clinician Group Input

This section was prepared by the CADTH review team based on the input provided by clinician groups. The full original clinician group inputs received by CADTH are included in the stakeholder section at the end of this report.

Clinician group input was received from the 7-member Ontario Health (Cancer Care Ontario) Genitourinary Drug Advisory Committee and 25 clinical subspecialists in Canada involved in VHL care.

The clinician groups agreed with the clinical experts consulted by CADTH that belzutifan, as the first systemic therapy option for VHL disease approved in Canada, fulfills an important unmet need for the treatment of patients with VHL and represents a shift in the current treatment paradigm. They also generally agreed on treatment goals, patient populations, response treatment-discontinuation criteria, and prescribing conditions.

While the clinical experts considered genetic testing to be a prerequisite for initiating treatment with belzutifan, neither clinician group indicated whether genetic testing for a VHL mutation or deletion was required. The clinician group indicated that belzutifan should be discontinued if the patient is pregnant.

Drug Program Input

Table 4: Summary of Drug Plan Input and Clinical Expert Responses

CNS = central nervous system; ECOG PS = Eastern Cooperative Oncology Group Performance Status; pERC = CADTH pan-Canadian Oncology Review Expert Review Committee; VHL = von Hippel-Lindau.

Clinical Evidence

The objective of CADTH’s Clinical Review Report is to review and critically appraise the clinical evidence submitted by the sponsor on the beneficial and harmful effects of belzutifan (40 mg per tablet, oral administration) in the treatment of adult patients with VHL disease who require therapy for associated nonmetastatic RCC, CNS hemangioblastomas, or nonmetastatic pNET not requiring immediate surgery. The focus will be placed on comparing belzutifan to relevant comparators and identifying gaps in the current evidence.

A summary of the clinical evidence included by the sponsor in the review of belzutifan is presented in 3 sections, and CADTH’s critical appraisal of the evidence is included after each section. The first section, the systematic review, includes pivotal studies and RCTs that were selected according to the sponsor’s systematic review protocol. The second section includes indirect evidence from the sponsor. The third section includes additional studies that were considered by the sponsor to address important gaps in the pivotal and RCT evidence.

Included Studies

Clinical evidence from the following is included in the CADTH review and appraised in this document:

• One pivotal phase II, single-arm, open-label trial (LITESPARK-004)

• One ITC

• One additional study addressing gaps in evidence (VHL Natural History Study)

In addition, a cross-sectional survey submitted by the sponsor that assessed HRQoL in patients with VHL disease who did not receive treatment with belzutifan is summarized and appraised in Appendix 1.

Pivotal Studies and Randomized Controlled Trial Evidence

Contents within this section have been informed by materials submitted by the sponsor. The following have been summarized and validated by the CADTH review team.

Description of Studies

Only 1 study (LITESPARK-004),¹³ which was conducted by the sponsor, met the inclusion criteria for the sponsor-submitted systematic review. Characteristics of the LITESPARK-004 trial are summarized in Table 5.

The LITESPARK-004 trial was a phase II, single-arm, open-label study of the efficacy and safety of belzutifan in 61 patients with VHL disease–associated nonmetastatic RCC across 11 centres in Denmark, France, the UK, and the US. The primary objective was to evaluate the efficacy of belzutifan for the treatment of VHL disease–associated nonmetastatic RCCs as measured by ORR. Evaluation of the efficacy of belzutifan for the treatment of VHL disease–associated nonmetastatic non-RCC tumours (e.g., CNS hemangioblastomas and pNET) as well as assessment of the safety and tolerability of belzutifan were among the secondary objectives. Evaluation of non-RCC tumours was not part of the original protocol, but was added after the completion of enrolment, using a retrospective review of existing images.

The LITESPARK-004 trial is ongoing. Data submitted by the sponsor to support this reimbursement request are based on a cut-off date of April 1, 2022. The database lock date was June 2, 2022.

Table 5: Details of Pivotal Studies and RCT Evidence Identified by the Sponsor

CNS = central nervous system; ECOG = Eastern Cooperative Oncology Group; HIF-2-alpha = hypoxia-Inducible factor 2 alpha subunit; IRC = independent review committee; pNET = pancreatic neuroendocrine tumour; RCC = renal cell carcinoma; RECIST 1.1 = Response Evaluation Criteria in Solid Tumors Version 1.1; RCT = randomized controlled trial; VEGF = vascular endothelial growth factor; VHL = von Hippel-Lindau.

Note: Details from the table have been taken from the sponsor’s Summary of Clinical Evidence.¹

^aEnd points including number of patients who develop a metastatic CNS hemangioblastoma or pNET, change in linear growth rate among patients with a VHL-associated nonmetastatic RCC, CNS hemangioblastoma, or pNET, and disease control rate were reported in the LITESPARK-004 Clinical Study Report but not explicitly listed as efficacy parameters in the LITESPARK-004 trial protocol. CADTH considered these exploratory end points.

Sources: LITESPARK-004 Clinical Study Report,¹³ LITESPARK-004 trial protocol,³³ and Jonasch et al. (2021).^30,34

Populations

Inclusion and Exclusion Criteria

Participants eligible to be included in the LITESPARK-004 trial were required to be at least 18 years of age and be diagnosed with VHL disease based on a germline VHL alteration, with at least 1 measurable RCC and an ECOG PS 0 or 1. Eligible patients could also have other VHL disease–associated tumours, including a CNS hemangioblastoma and pNET. Patients who had an immediate need for surgical intervention for tumour treatment or evidence of metastatic disease were excluded.

Interventions

In the single-arm, open-label LITESPARK-004 trial, belzutifan was provided centrally by the sponsor to patients and administered orally at a dosage of 120 mg (in 3 40 mg tablets) once daily until disease progression or unacceptable AEs. The tablets were taken approximately 24 hours apart. Patients who required surgery for nonrenal lesions may have had their belzutifan interrupted in the perioperative period, if necessary. Administration of belzutifan was resumed when patients could swallow and tolerate oral medications after surgery.

Patients were not allowed to receive any approved or additional investigational antineoplastic drug during the study. Pretreatment administration of prophylactic antiemetics was also not allowed but could be given if needed during study treatment. There were no constraints on the use of growth factors, including erythropoietin, during treatment. However, prophylactic use was discouraged and adherence to the American Society of Clinical Oncology or European Society for Medical Oncology guidelines was recommended. Patients should have received all necessary supportive care, including blood products, transfusions, antibiotics, pain medications, bisphosphonates, and replacement hormonal therapies (insulin, thyroid hormones, estrogen, or progesterone).

Outcomes

A list of efficacy end points assessed in this clinical review report is provided in Table 6. These end points are further summarized in the following section. Summarized end points are based on those included in the sponsor’s Summary of Clinical Evidence¹ as well as any identified as important to this review by stakeholders (clinical experts, clinician groups, or patient groups). The distribution of all tumour-reduction procedures before and after treatment initiation for individual patients was also reported in the Clinical Study Report, even though this outcome was not prespecified in the study protocol.

Table 6: Outcomes Summarized From Pivotal Studies and RCT Evidence Identified by the Sponsor

BOR = best overall response; CNS = central nervous system; DCR = disease control rate; DOR = duration of response; IRC = independent review committee; LGR = linear growth rate; ORR = objective response rate; PFS = progression-free survival; pNET = pancreatic neuroendocrine tumour; RCC = renal cell carcinoma; RCT = randomized controlled trial; TTR = time to response; TTS = time to surgery; VHL = von Hippel-Lindau.

^aUnless otherwise specified, outcomes assessed by investigators were not available in the LITESPARK-004 Clinical Study Report

^bThese outcomes were not explicitly listed as end points in the LITESPARK-004 trial protocol, but were considered by CADTH to be exploratory end points.

Sources: LITESPARK-004 Clinical Study Report,³⁵ LITESPARK-004 trial protocol,^33,34 LITESPARK-004 statistical analysis plan,³⁶ and Jonasch et al. (2021).³⁰

Efficacy Outcomes

Patients were assessed radiologically about 12 weeks after initiation of treatment and every 12 weeks thereafter while continuing on belzutifan for a minimum of 3 years. After 3 years, patients were evaluated radiologically every 24 weeks, or more frequently if clinically indicated.

Patients who discontinued belzutifan for any reason underwent long-term follow-up every 6 months until the patient either withdrew consent for follow-up or died. If patients discontinued belzutifan without documented disease progression and had not met any of the discontinuation criteria for scanning (i.e., IRC-verified disease progression as defined by RECIST 1.1, the start of a new anticancer treatment, pregnancy, death, withdrawal of consent, or end of study), every effort was to be made to monitor disease status by acquiring tumour scans 6 months after the last scan and at 6-month intervals for a minimum of 2 years, or more frequently if indicated. After 2 years, scans were collected every 12 months or more frequently if clinically indicated.

Time to Surgery

TTS for VHL disease–associated nonmetastatic RCC, CNS hemangioblastomas, or pNET was defined as the interval from the start of study treatment to the date of surgery.³⁴ Patients who did not undergo surgery were censored as of the last known alive date. Surgery was defined as any tumour-reducing intervention including partial nephrectomy, radical nephrectomy, ablative procedure (e.g., cryoablation, thermal ablation, radioablation), and tumour-debulking surgeries, but excluding radiation therapy. The data cut-off date for TTS to support this reimbursement request was April 1, 2022.

Progression-Free Survival

PFS was measured for VHL disease–associated nonmetastatic RCCs, CNS hemangioblastomas, and pNETs. PFS was defined as the interval from the start of study treatment until the earlier of the first documentation of disease progression determined by RECIST 1.1 or death from any cause.³⁴

Several censoring rules for PFS were applied. First, if progressive disease or death was documented after no more than 1 missed disease assessment and before new anticancer therapy, PFS was considered as an event at the date of documented progressive disease or death. Second, if progressive disease or death was documented immediately after 2 or more consecutive missed disease assessments or after new anticancer therapy, the patient was censored at the last adequate disease assessment before the 2 consecutive missed disease assessments and new anticancer therapy. Third, if there was no progressive disease or death, and a new anticancer therapy was initiated, the patient was censored at the last disease assessment documented before the new anticancer therapy. Fourth, if there was no progressive disease, death, or newly initiated anticancer therapy, the patient was censored at the last adequate disease assessment. Fifth, if no adequate posttreatment tumour assessments were obtained for a patient, the patient was censored at day 1 (i.e., zero duration).³⁶ The data cut-off date for PFS to support this reimbursement request was April 1, 2022.

Objective Response Rate

In the LITESPARK-004 trial protocol,³⁴ ORR was defined as the proportion of patients with a best confirmed response of a CR or PR as determined by RECIST 1.1 for each VHL-associated organ system. ORR for a VHL disease–associated nonmetastatic RCC measured by IRC was the only primary end point for the LITESPARK-004 trial. All other ORR outcomes were considered secondary end points. The data cut-off date for ORR to support this reimbursement request was April 1, 2022.

Duration of Response

DOR was measured for VHL disease–associated nonmetastatic RCCs, CNS hemangioblastomas, and pNETs. DOR was defined as the interval from the first documentation of response, as determined by RECIST 1.1, to the earlier of the first documentation of disease progression or death from any cause and calculated for patients with a best confirmed response of a CR or PR.³⁴ Start time was the first instance of a CR or PR that was subsequently confirmed, not the confirmatory date itself.

Four censoring rules for DOR were applied. First, if progressive disease or death was documented after no more than 1 missed disease assessment and before new anticancer therapy, the outcome was considered as an event and DOR was documented as progressive disease or death. Second, if progressive disease or death was documented immediately after 2 or more consecutive missed disease assessments or after new anticancer therapy, the outcome was considered a non-event and the patient was censored at the earlier date of the last adequate disease assessment before 2 or more missed adequate disease assessments and new anticancer therapy. Third, if there was no progressive disease or death, and a new anticancer therapy was initiated, the outcome was considered a nonevent and the patient was censored at the last adequate disease assessment before the new anticancer therapy. Fourth, if there was no progressive disease, death or newly initiated anticancer therapy, the outcome was considered a nonevent and the patient was censored at the last adequate disease assessment.³⁶ The data cut-off date for DOR to support this reimbursement request was April 1, 2022.

Disease Control Rate

In the sponsor-submitted clinical evidence summary,¹ DCR was defined as the proportion of patients with a best confirmed response of a CR, PR, or stable disease by an IRC as defined by RECIST 1.1. In the sponsor’s statistical analysis plan (version 3.0; February 24, 2022),³⁶ DCR was defined as the proportion of patients who have a BOR of a CR, PR, or stable disease. However, DCR was not a predefined end point of interest in the LITESPARK-004 trial protocol (version 10.0; February 15, 2022).³⁴ The sponsor noted that this was added post hoc to report on the proportion of patients with stable disease. The data cut-off date for DOR to support this reimbursement request was April 1, 2022.

Best Overall Response

BOR, which summarized the number of patients who had a CR, PR, stable disease, progressive disease, or a status not evaluable, was not explicitly listed as an end point and defined in the LITESPARK-004 trial protocol (version 10.0; February 15, 2022).³⁴ The data cut-off date for BOR to support this reimbursement request was April 1, 2022.

Time to Response

TTR for VHL disease–associated nonmetastatic RCC, CNS hemangioblastomas, or pNET was defined as the interval from the start of study treatment to the first documentation of a response, as determined by RECIST 1.1, and calculated for patients with a best confirmed response of a CR or PR.³⁴ The data cut-off date for TTR to support this reimbursement request was April 1, 2022.

Change in Linear Growth Rate

Change in LGR was not explicitly listed as an end point in the LITESPARK-004 trial protocol (version 10.0; February 15, 2022).³⁴ According to the sponsor’s statistical analysis plan (version 3.0; February 24, 2022),³⁶ LGRs for target tumours before and after study drug treatment were calculated for patients with at least 3 scans before treatment, including the screening scan. The data cut-off date for change in LGR to support this reimbursement request was April 1, 2022.

Number of Patients who Develop Metastases

Although not explicitly defined in the protocol, the number of patients who developed metastatic RCC and number of patients who developed metastatic pNET were reported. The number of patients who developed metastatic CNS hemangioblastomas was not applicable because CNS hemangioblastomas are nonmalignant tumours by histology and do not metastasize, according to the sponsor’s explanation, which was validated by the clinical experts consulted by CADTH.

Statistical Analysis

In the LITESPARK-004 trial, the null hypothesis was that the ORR is 15%, while the alternative hypothesis was that the ORR is 30%. The sample size was estimated at approximately 50 patients based on the following assumptions: A sample size of 50 patients would provide greater than 80% power to reject the null under the alternative hypothesis using a 1-sided test at a 0.05 level of significance.³⁴

No inferential statistical analyses were performed as this was a single-arm trial.³⁶ Data were summarized using descriptive statistics. For continuous variables, the number of patients with nonmissing data, mean, standard deviation (SD), median, and range were presented. For categorical data, the frequency and percent distribution were presented.

Estimates of the ORR along with the associated 95% exact binomial CIs (Clopper-Pearson method) were provided. The binomial probability that the observed RCC ORR was greater than 0.15 was calculated. BOR was summarized. Kaplan-Meier methods were used to estimate the distribution of DOR, PFS, and TTS. Quartiles including the median were estimated by the Kaplan-Meier method along with their 95% CIs (Brookmeyer-Crowley method). Kaplan-Meier cumulative estimators along with their associated 95% CIs were provided for RCC DOR rates at 26, 39, and 52 weeks. TTR was calculated and summarized descriptively. LGRs at tumour level pre- and posttreatment were calculated for tumours using at least 3 scans (including the screening scan) and linear regression was applied to the tumour size with time as a continuous variable. LGRs at the patient level pre- and posttreatment were calculated for patients with at least 3 scans (including the screening scan) and linear regression was applied to the tumour sizes with time as a continuous variable and individual tumour as a categorical variable. LGRs were derived as the coefficient of time. The number and proportion of patients developing metastatic RCC or pNET were summarized descriptively.

Tumour responses were measured by CT or MRI and assessed according to RECIST 1.1. CRs and PRs were confirmed by a second assessment at least 4 weeks after the initial response was documented. Where confirmed responses were available, the date of the first assessment was used in all analyses. If a confirmation assessment was not available, the assessment at the next scheduled time point was used for confirmation. If there were no confirmation assessments, the response was taken to be unconfirmed and considered a nonresponse. Patients who did not have a tumour assessment for any reason were considered nonresponders and included in the denominator when calculating the response rate.

AEs by maximum severity were summarized by system organ class and preferred term, according to the number and percentage of patients with at least 1 AE, SAE, death, or AE leading to drug withdrawal, or dose reduction or interruption.

Missing data were not estimated or carried forward for any of the other summaries. If only a partial date was available and was required for a calculation of whether a medication is concomitant or an AE is treatment-emergent, the start and stop dates were imputed according to a predefined standard. No other imputation of missing data was performed.

Analysis populations of the LITESPARK-004 trial are summarized in Table 7. All participants in the LITESPARK-004 trial should have had a VHL disease–associated nonmetastatic RCC to meet the eligibility criteria. Efficacy results for VHL disease–associated nonmetastatic RCCs therefore came from the efficacy analysis set, while efficacy results for VHL disease–associated CNS hemangioblastomas and nonmetastatic pNETs were from subsets of the efficacy analysis set that were defined according to the study protocol³⁴ by the site of primary tumours other than RCCs at screening.

Protocol Amendments and Deviations

Between November 10, 2017, and February 15, 2022, 10 versions of the LITESPARK-004 trial protocol were identified. A summary of key protocol amendments implemented in the LITESPARK-004 trial is presented in Table 8.

Table 7: Analysis Populations in the LITESPARK-004 Trial

Source: LITESPARK-004 trial protocol.^33,34

Table 8: Summary of Key Protocol Amendments in LITESPARK-004

^aRevisions were highlighted in bold.

Source: LITESPARK-004 trial protocol.³⁴

As of April 1, 2022, 6 important protocol deviations were reported: 1 patient was given an unapproved drug to treat fatigue and headaches; 1 patient was administered a higher belzutifan dosage of 120 mg once a day compared with the originally prescribed dose of 80 mg once a day; 1 patient discontinued belzutifan between June 2019 and August 2019 due to abdominal pain (exceeding the 3-week off-drug interval allowed by the study protocol); 1 patient had an ECOG PS of 2, which did not meet the study eligibility criteria; 1 patient was off belzutifan for more than 3 weeks due to lack of access; 1 patient was removed from the LITESPARK-004 trial on October 22, 2020, but the date of the last dose of belzutifan was November 16, 2020.

Results

Patient Disposition

A summary of patient disposition in the LITESPARK-004 trial is provided in Table 9. Of 67 screened participants, 9.0% (6) were screen failures, all of which were due to failure to meet eligibility criteria. Because the LITESPARK-004 trial was a single-arm trial, all 61 eligible participants received the treatment of belzutifan. As of the data cut-off date of April 1, 2022, 9.8% of the study participants (5 of 61) discontinued from the study. About 37.7% of the participants (23 of 61) discontinued study treatment as of April 1, 2022, among which discontinuation due to patient decisions (n = 11; 18.0%) was the most common reason, followed by disease progression (n = 6; 9.8%). Reasons for discontinuing belzutifan due to patient decisions included: study drug side effects (i.e., 3 patients due to fatigue and 1 patient due to feeling down and disappointed), personal or logistical reasons (4 patients due to difficulty in travelling, 1 patient was not able to afford medical expenses, and 1 patient started a family), and unknown reasons (i.e., 1 patient was unable to return to the trial site and had been noncommunicative).

For this reimbursement request, efficacy data for patients with VHL disease–associated RCC were from all 61 study patients, efficacy data for patients with VHL disease–associated CNS hemangioblastoma were from 50 participants, and efficacy data for patients with VHL disease–associated pNET came from 22 patients.

Table 9: Summary of Patient Disposition From Pivotal Studies and RCT Evidence Submitted by the Sponsor

RCT = randomized controlled trial.

Source: LITESPARK-004 Clinical Study Report.¹³

Baseline Characteristics

Baseline characteristics of the LITESPARK-004 trial are shown in Table 10. In the total study population (i.e., patients with RCC, N = 61), the median age was 41.0 years (range = 19.0 to 66.0), and the median age at time of VHL disease diagnosis was 32.0 years (range = 4.0 to 66.0). Approximately half of the patients were male (52.5%) and half were female (47.5%). Most patients were white (90.2%), had an ECOG PS of 0 (82.0%), and had a previous surgery or ablative procedure (96.7%). The median number of previous procedures was 5.0 (range = 1.0 to 15.0). In the total study population, 77.1% of patients had a previous surgery for an RCC, 75.4% had a previous surgery for a CNS hemangioblastoma, and 14.8% had previous pancreas surgery.

Median age was similar in the CNS hemangioblastoma and pNET subpopulations. For patients with a CNS hemangioblastoma (n = 50), the median age was 40.5 years (range = 19.0 to 65.0). For patients with a pNET (n = 22), the median age was 42.0 years (range = 19.0 to 66.0). Similar to the total study population, most patients in the CNS hemangioblastoma and pNET subpopulations were white (88.0% and 81.8%, respectively) and had an ECOG PS of 0 (78.0% and 90.9%, respectively).

Exposure to Study Treatments

At the data cut-off date of April 1, 2022, the median duration of exposure to belzutifan was 37.3 months (range = 1.9 to 46.1) and 44 patients (72%) had received treatment for at least 36 months. A total of 50 patients (82%) had a dose interruption over the course of the study; dose interruption due to AEs was reported in 26 patients (43%). Eleven patients (18%) had a dose reduction to 80 mg daily and 4 patients (7%) had a dose reduction to 40 mg daily (Table 11).

Table 10: Summary of Baseline Characteristics of Pivotal Studies and RCT Evidence Submitted by the Sponsor

BMI = body mass index; CNS = central nervous system; ECOG = Eastern Cooperative Oncology Group; NA = not applicable; NR = not reported; pNET = pancreatic neuroendocrine tumour; RCC = renal cell carcinoma; SD = standard deviation; VHL = von Hippel-Lindau; vs. = versus.

Sources: LITESPARK-004 Clinical Study Report.¹³

Table 11: Summary of Patient Exposure from Pivotal Studies and RCT Evidence Submitted by the Sponsor

RCT = randomized controlled trial; SD = standard deviation.

Source: LITESPARK-004 Clinical Study Report.³⁵

Concomitant Medications

All concomitant medication(s) used during the study and within 28 days before the start of study drug administration were recorded. Almost all (98.4%; 60 of 61) of the participants received concomitant medications during the LITESPARK-004 trial. The most common concomitant medications (in at least 20% of total study population) included acetaminophen (39.3%), COVID-19 vaccines (36.1%), ibuprofen (23.0%), and iron (21.3%). The most frequently reported concomitant treatments were analgesics (62.3%), followed by antianemic preparations (41.0%), vaccines (39.3%), ophthalmologicals (37.7%), and psycholeptics (37.7%). One patient received radiation therapy during treatment of belzutifan for a CNS hemangioblastoma.

Efficacy

Key efficacy results are presented in Table 12. All efficacy outcomes were obtained from the data with a cut-off date of April 1, 2022. The median follow-up duration was 37.7 months (range = 4.2 to 46.1 months). IRC- and investigator-assessed efficacy outcomes were reported for most of the patients, with VHL disease–associated nonmetastatic RCC and those with CNS hemangioblastoma, whereas only IRC-assessed efficacy results were available for patients with VHL disease–associated nonmetastatic pNET. Unless otherwise specified, results described in this report were IRC-assessed.

Time to Surgery

Median TTS was not reached for patients with VHL disease–associated nonmetastatic RCCs, CNS hemangioblastomas, or nonmetastatic pNETs as of the data cut-off date (April 1, 2022). Seven (11.5%; 7 of 61) patients with an RCC, 1 patient (2.0%; 1 of 51) with a CNS hemangioblastoma, and none of the patients with a pNET had surgery during the follow-up period. (Table 12). The Kaplan-Meier plot of TTS was only presented for patients with RCC in the LITESPARK-004 trial (Figure 1).

Table 12: Summary of Key Efficacy Results from Pivotal Evidence

BOR = best overall response; CI = confidence interval; CNS = central nervous system; CR = complete response; DCR = disease control rate; DOR = duration of response; LGR = linear growth rate; NA = not applicable; NE = not evaluable or not estimable; NR = not reported; ORR = objective response rate; PFS = progression-free survival; pNET = pancreatic neuroendocrine tumour; PR = partial response; RCC = renal cell carcinoma; SD = standard deviation; TTR = time to response; TTS = time to surgery.

^aAccording to sponsor’s explanation, the “+” sign means that the objective response in the respective participant remained ongoing no progressive disease or death was reported at the time of data cut-off. For example, “5.4+” means that, at the data cut-off, the respective participant had been in response for 5.4 months and had not experienced disease progression or death.

Source: LITESPARK-004 Clinical Study Report.¹³

Figure 1: Kaplan-Meier Plot of TTS for VHL Disease–Associated Nonmetastatic RCC

CI = confidence interval; NE = not evaluable or not estimable; RCC = renal cell carcinoma; TTS = time to surgery; VHL = von Hippel-Lindau.

Source: LITESPARK-004 Clinical Study Report.¹³

Progressive-Free Survival

The IRC-assessed median PFS was 39.2 months (95% CI, 38.5 to not evaluable) for patients with a VHL disease–associated nonmetastatic RCC, whereas the investigator-assessed median PFS was not reached. Median PFS was also not reached for patients with a VHL disease–associated CNS hemangioblastoma and those with a nonmetastatic pNET at the data cut-off date (April 1, 2022). Kaplan-Meier plots of PFS for patients with a VHL disease–associated nonmetastatic RCC, CNS hemangioblastoma, or nonmetastatic pNET assessed by IRC are in Figure 2, Figure 3, and Figure 4, respectively.

Details regarding the events and censoring of PFS are provided in Table 13. IRC-assessed results showed that, among 61 patients with RCC at baseline, 11 (18.0%) had PFS events (i.e., progressive disease, death), and 50 (82%) patients were censored, mostly due to no progression at the time of data cut-off or before the end of treatment (n = 43; 70.5%). Investigator-assessed results were similar to IRC-assessed results.

IRC-assessed results showed that, among 50 patients with CNS hemangioblastoma at baseline, 11 (22.0%) had PFS events (i.e., progressive disease or death), and 39 patients (78%) were censored (among which 34 [68%] were due to no progression at the time of data cut-off). However, the investigator-assessed results found that, out of 51 patients with CNS hemangioblastoma at baseline, 48 patients (94.1%) were censored, among which 46 (90.2%) were due to no progression at the time of data cut-off.

All 22 patients with pNET were censored due to no progression at the time of data cut-off.

Figure 2: Kaplan-Meier Plot of PFS for VHL-Associated Nonmetastatic RCC — IRC Assessed

CI = confidence interval; NE = not evaluable or not estimable; IRC = independent review committee; PFS = progression-free survival; RCC = renal cell carcinoma; VHL = von Hippel-Lindau.

Source: LITESPARK-004 Clinical Study Report.¹³

Objective Response Rate

As of the data cut-off date of April 1, 2022, with a median follow-up of 37.7 months, the IRC-assessed percentage of patients who had a CR or PR to belzutifan was 63.9% (39 of 61) among those with a VHL-associated nonmetastatic RCC, 44% (22 of 50) among those with a VHL-associated CNS hemangioblastoma, and 90.9% (20 of 22) among those with a VHL-associated nonmetastatic pNET (Table 12). IRC- and investigator-assessed ORRs were consistent among patients with a nonmetastatic RCC or a CNS hemangioblastoma.

Figure 3: Kaplan-Meier Plot of PFS for VHL-Associated CNS Hemangioblastoma — IRC Assessed

CNS = central nervous system; CI = confidence interval; NE = not evaluable or not estimable; IRC = independent review committee; PFS = progression-free survival; VHL = von Hippel-Lindau.

Source: LITESPARK-004 Clinical Study Report.¹³

Figure 4: Kaplan-Meier Plot of PFS for VHL-Associated Nonmetastatic pNET — IRC Assessed

CI = confidence interval; NE = not evaluable or not estimable; IRC = independent review committee; PFS = progression-free survival; pNET = pancreatic neuroendocrine tumour; VHL = von Hippel-Lindau.

Source: LITESPARK-004 Clinical Study Report.¹³

Duration of Response

The investigator-assessed median DOR for patients with VHL-associated CNS hemangioblastoma (the only estimable DOR) was 33.8 months (95% CI, 33.8 to not evaluable). The IRC-assessed median DOR was not reached for responders with a VHL disease–associated nonmetastatic RCC, CNS hemangioblastoma, or nonmetastatic pNET at the data cut-off date of April 1, 2022. For patients with an RCC, 74.4% (29 of 39) of responders had a DOR of at least 18 months, 56.4% (22 of 39) had a DOR of at least 24 months, and 25.6% (10 of 39) had a DOR of at least 30 months. For patients with CNS hemangioblastoma, 63.6% (14 of 22) of responders had a DOR of at least 18 months, 59.1% (13 of 22) had a DOR of at least 24 months, and 54.5% (12 of 22) had a DOR of at least 30 months. For patients with pNET, 95.0% (19 of 20) of responders had a DOR of at least 18 months, 75.0% (15 of 20) had a DOR of at least 24 months, and 40.0% (8 of 20) had a DOR of at least 30 months.

Table 13: Details on Events and Censoring of PFS

CNS = central nervous system; NR = not reported; PFS = progression-free survival; pNET = pancreatic neuroendocrine tumour; RCC = renal cell carcinoma; VHL = von Hippel-Lindau.

Source: LITESPARK-004 Clinical Study Report.¹³

Kaplan-Meier plots of DOR for patients with a VHL disease–associated nonmetastatic RCC, CNS hemangioblastoma, or nonmetastatic pNET assessed by IRC are provided in Figure 5, Figure 6, and Figure 7, respectively.

Details regarding the events and censoring of DOR are provided in Table 14. Among 39 RCC patients with a confirmed response, 32 (82.1%) were censored due to no progression at the time of data cut-off or before end of treatment. Out of 22 patients with CNS hemangioblastoma who showed a confirmed response, 17 were censored due to no progression at the time of data cut-off or before the end of treatment. All 20 patients with a pNET who had a confirmed response were censored due to no progression at the time of data cut-off.

Disease Control Rate

As of the data cut-off date of April 1, 2022, the IRC-assessed percentages of patients who had a CR, PR, or stable disease were 98.4% (60 of 61) among those with a VHL-associated nonmetastatic RCC, 90% (45 of 50) among those with a VHL-associated CNS hemangioblastomas, and 100% (22 of 22) among those with a VHL-associated nonmetastatic pNET (Table 12). Similar percentages were found between IRC- and investigator-assessed DCRs among patients with a nonmetastatic RCC or a CNS hemangioblastoma.

Best Overall Response

As of the data cut-off date of April 1, 2022, the IRC-assessed percentages of patients with a VHL disease–associated nonmetastatic RCC who had a CR, PR, stable disease, progressive disease, or not evaluable while taking belzutifan were 6.6% (4 of 61), 57.4% (35 of 61), 34.4% (21 of 61), 0% (0 of 61), and 1.6% (1 of 61), respectively. IRC- and investigator-assessed BORs were consistent.

Figure 5: Kaplan-Meier Plot of DOR for VHL-Associated Nonmetastatic RCC — IRC Assessed

CI = confidence interval; DOR = duration of response; NE = not evaluable or not estimable; IRC = independent review committee; RCC = renal cell carcinoma; VHL = von Hippel-Lindau.

Source: LITESPARK-004 Clinical Study Report.¹³

Figure 6: Kaplan-Meier Plot of DOR for VHL-Associated CNS Hemangioblastoma — IRC Assessed

CNS = central nervous system; CI = confidence interval; DOR = duration of response; NE = not evaluable or not estimable; IRC = independent review committee; VHL = von Hippel-Lindau.

Source: LITESPARK-004 Clinical Study Report.¹³

Figure 7: Kaplan-Meier Plot of DOR for VHL Disease Nonmetastatic pNET — IRC Assessed

CI = confidence interval; DOR = duration of response; NE = not evaluable or not estimable; IRC = independent review committee; pNET = pancreatic neuroendocrine tumour; VHL = von Hippel-Lindau.

Source: LITESPARK-004 Clinical Study Report.¹³

Table 14: Details on Events and Censoring of Duration of Response

CNS = central nervous system; DOR = duration of response; IRC = independent review committee: pNET = pancreatic neuroendocrine tumour; RCC = renal cell carcinoma; VHL = von Hippel-Lindau.

Source: LITESPARK-004 Clinical Study Report.¹³

The IRC-assessed percentages of patients with a VHL disease–associated CNS hemangioblastoma who had a CR, PR, stable disease, progressive disease, or not evaluable while taking belzutifan were 8% (4 of 50), 36% (18 of 50), 46% (23 of 50), 6% (3 of 50), and 4% (2 of 50), respectively. Investigator-assessed BORs were higher in patients with stable disease compared with IRC-assessed BORs (62.7% [32 of 50] versus 46% [23 of 50]).

The IRC-assessed percentages of patients with a VHL disease–associated nonmetastatic pNET who had a CR, PR, stable disease, progressive disease, or not evaluable while taking belzutifan were 31.8% (7 of 22), 59.1% (13 of 22), 9.1% (2 of 22), 0% (0 of 22), and 0% (0 of 22), respectively (Table 12).

Time to Response

As of the data cut-off date of April 1, 2022, the IRC-assessed median TTRs were 11.1 months (range = 2.7 to 30.5) among those with a VHL-associated nonmetastatic RCC, 5.4 months (range = 2.3 to 33.1) among those with a VHL-associated CNS hemangioblastomas, and 8.2 months (range = 2.5 to 16.4) among those with a VHL-associated nonmetastatic pNET (Table 12). IRC- and investigator-assessed TTRs were similar among patients with a nonmetastatic RCC or CNS hemangioblastoma.

Linear Growth Rate

The IRC-assessed LGR at both the patient and lesion levels showed a decrease in lesion LGR from before to after treatment with belzutifan in patients with a VHL-associated nonmetastatic RCC and in those with a pNET. Only LGRs after treatment with belzutifan were analyzed for patients with a VHL-associated CNS hemangioblastomas as no pretreatment CNS scans were collected in the LITESPARK-004 trial (Table 12).

Number of Patients who Developed Metastases

As of the data cut-off date of April 1, 2022, only 1 patient (1.6%) developed a metastatic RCC, while none of the patients with a VHL-associated nonmetastatic pNET developed metastatic disease.

Distribution of All Tumour–Reduction Procedures Before and After Treatment Initiation

As of the data cut-off date of April 1, 2022, fewer surgeries or tumour-reduction procedures occurred after patients received belzutifan treatment compared to the same period immediately before treatment initiation (Figure 8): 11 surgeries (7 renal, 2 CNS, and 2 retinal) were performed in 9 patients, whereas 73 surgeries in 44 patients had been performed in the 3 years before initiation of belzutifan treatment.

Figure 8: Distribution of All Tumour–Reduction Procedures Before and After Treatment Initiation for Individual Patients (Safety Analysis Set)

CNS = central nervous system; RCC = renal cell carcinoma.

Notes: Procedures included adrenalectomy, craniotomy, cryoablation, cryotherapy, eye removal, intradural resection, laser ablation, laser surgery, laminectomy, laser photocoagulation, pancreatectomy, partial nephrectomy, radiofrequency ablation retinal surgery, total nephrectomy, tumour enucleation, and ventriculoperitoneal shunt placement. Two patients did not undergo tumour resection procedures at any time before entering the study or while enrolled in the study. The 2 CNS surgical procedures reported after belzutifan initiation were in the same patient; 1 patient underwent radiation therapy but was not included in this figure.

Source: LITESPARK-004 Clinical Study Report.¹⁴

Harms

A summary of harms in the LITESPARK-004 trial (data cut-off date: April 1, 2022) is shown in Table 15.

Table 15: Summary of Harms — Pivotal and RCT Evidence

^aCriteria for drug-induced liver injury: alanine transaminase or aspartate transaminase 3 or more times the upper limit of normal range, and bilirubin 2 or more times the upper limit of normal range, and alkaline phosphatase less than twice the upper limit of normal range.

Source: LITESPARK-004 Clinical Study Report.¹³

Adverse Events

TEAEs were reported in all 61 patients (100%) in the LITESPARK-004 trial. The most commonly reported TEAE was anemia (90.2%), followed by fatigue (73.8%), headaches (47.5%), dizziness (45.9%), and nausea (39.3%). Other TEAEs that were reported in more than 20% of the study participants included dyspnea (26.2%), myalgia (24.6%), constipation (23.0%), arthralgia (21.3%), and blurred vision (20.3%).

Serious Adverse Events

SAEs were reported in 18 patients (29.3%). The most common SAEs were embolism (3.3%) and intracranial hemorrhage (3.3%).

Mortality

Two patients died during the study due to acute toxic effects of fentanyl (n = 1) and suicide (n = 1).

Withdrawal due to Adverse Events

Treatment discontinuation due to TEAE was reported in 4 patients, of which 2 (dizziness and intracranial hemorrhage) were considered drug-related.

Notable Harms

As of April 1, 2022, 90.2% (55 of 61) patients had 229 episodes of anemia. The average number of episodes of anemia for each patient was 4.2.

One patient (1.6%) developed hypoxia. Two patients (3.3%) developed secondary primary malignancies, 1 of whom had vulval cancer and 1 non–small cell lung cancer. None of the patients had drug-induced liver injury.

Critical Appraisal

Pivotal Evidence

Internal Validity

The LITESPARK-004 trial, the only eligible study identified from the sponsor-conducted SLR, was a phase II, single-arm, open-label clinical trial that enrolled a small number of patients (N = 61). Given the rarity of VHL disease and that active surveillance is the current standard of care for patients with VHL-associated nonmetastatic tumours, the single-arm trial design and small sample size were considered adequate from the regulatory perspective to assess the efficacy and safety of belzutifan. No serious concerns were identified in the conduct of the LITESPARK-004 trial, including protocol amendments and protocol deviations. However, the absence of an internal comparison group in the single-arm LITESPARK-004 trial is a key limitation that increases the uncertainty of the findings from the pivotal study. The lack of comparative data means that causal inferences about the efficacy and safety of belzutifan cannot be established with certainty.

As consecutive enrolment of eligible patients was not mentioned, it is unclear whether all eligible patients would have been entered into the trial, or whether selection may have been biased to include patients who were most likely to benefit. However, the clinical experts consulted by CADTH indicated that the patient population appeared to be representative of what they would expect to see in clinical practice. Almost all patients used concomitant treatments during the trial, but these were not expected to affect the reported efficacy results.

The LITESPARK-004 trial explicitly defined the hypothesis, which was a null hypothesis of an ORR of 15% or lower with an alternate hypothesis of an ORR of 30% or higher. These thresholds were considered clinically meaningful by the clinical experts consulted by CADTH. The selection of ORR (defined as sum of CR and PR as defined by RECIST 1.1) as a key efficacy end point was appropriate for the single-arm monotherapy trial design because the ORR, when defined in this manner, is generally viewed as a direct measure of antitumour activity as well as an objective and clinically relevant end point. The LITESPARK-004 trial also assessed DOR to provide evidence on the durability of tumour response (i.e., the durability of CR or PR), which is an important end point in single-arm trials in oncology.¹⁴ DCR was not a protocol-specified end point but was reported in the LITESPARK-004 Clinical Study Report, with the sponsor noting the importance of considering stable disease a potential response. However, the clinical experts consulted by CADTH also noted that tumours do not grow at a steady rate, and periods of stability might be expected as part of the natural history of the disease.

In addition to DOR, the LITESPARK-004 trial employed other time-to-event end points (TTS and PFS) to address the potential limitations of ORR (i.e., ORR does not always capture the effects of a treatment on patient survival and may not always correlate with symptoms or function).¹⁴ The clinical experts consulted by CADTH considered these the most important outcomes to assess the efficacy of belzutifan. Generally speaking, RCTs are preferred over single-arm studies for time-to-event end points such as PFS due to their sensitivity to baseline differences in patient, disease, and other clinical characteristics, and interpreting results without a randomized reference can be problematic.^14-16 In the context of the LITESPARK-004 trial, medians of most of the time-to-event data were not reached, but reasons for censoring appeared to support the efficacy of belzutifan, primarily due to the lack of progression at the time of the data cut-off.

The LITESPARK-004 trial involved outcomes pre- and posttreatment with belzutifan (i.e., change in LGR) to demonstrate the efficacy of belzutifan. However, without formal statistical analysis, the role of chance could not be ruled out, despite the presence of a trend in which the growth of tumours was decreasing after treatment with belzutifan. With respect to outcome measurement, in addition to study investigators, an IRC was also involved to assess radiographic outcomes to reduce the risk of bias in measurements of the outcome for most of the efficacy end points in patients with a VHL-associated nonmetastatic RCC and those with a VHL-associated CNS hemangioblastoma.

External Validity

All participants in the LITESPARK-004 trial were required to have at least 1 RCC. The LITESPARK-004 trial therefore may not reflect results for participants with only a CNS hemangioblastoma and/or pNET. However, the clinical experts consulted by CADTH did not consider this a serious generalizability issue, and they indicated that results from the pivotal trial could still be extrapolated to VHL patients with CNS hemangioblastoma and/or nonmetastatic pNET.

According to the clinical experts consulted by CADTH, the inclusion and exclusion criteria of the LITESPARK-004 trial in general were aligned with selection criteria in the Canadian settings when identifying suitable candidates for belzutifan. However, the clinical experts noted that the requirement for patients with an ECOG PS of 0 or 1 might not be necessary in clinical practice to initiate belzutifan because ECOG PS can be unstable and subjective. The clinical experts indicated that a clinician may determine that a patient with an ECOG PS greater than 1 could benefit from belzutifan and prescribe belzutifan.

There were no study sites in Canada as the LITESPARK-004 trial was conducted in Denmark, France, the UK, and the US. More than 90% of the patients in LITESPARK-004 were white, which does not reflect the racial profile in the Canadian patient population; given the rarity of the disease, it is unclear whether this might affect the natural history or response to treatment. The dosing and administration of belzutifan in LITESPARK-004 were consistent with the product monograph. The clinical experts described the concomitant medications and procedures in the LITESPARK-004 trial as appropriate and commonly used in the Canadian settings. The outcomes of TTS, PFS, ORR, and DOR are commonly used in clinical trials of anticancer therapies and relevant to clinical practice, according to the clinical experts. However, LGR is not commonly adopted in clinical practice. The clinical expert specializing in CNS hemangioblastoma noted that, in clinical practice, RANO criteria are used instead of RECIST 1.1 to assess tumour response in patients with CNS hemangioblastoma. The LITESPARK-004 trial did not include some outcomes important to patients, such as symptoms and HRQoL.

Overall survival, which is usually viewed as the most reliable and preferred end point in oncology trials,¹⁴ was not used as an end point in the LITESPARK-004 trial. The clinical experts consulted by CADTH commented that the absence of overall survival data was acceptable given that overall survival might not be a feasible outcome in this patient population and indication because death due to VHL disease could take years to occur and that overall survival could be confounded by multiple lines of therapy. The clinical experts also pointed out that TTS and PFS are more clinically important to VHL patients with nonmetastatic disease given that delaying morbid local therapies (e.g., surgery, radiation) was a main treatment goal.

The median length of follow-up in the LITESPARK-004 trial was 37.7 months, which was adequate to observe response, but a longer follow-up would be needed to understand the full extent of the efficacy of belzutifan on time-to-event outcomes. The clinical experts consulted by CADTH noted that this length of follow-up would be adequate to capture adverse effects. However, given that the life expectancy of patients with VHL could be very long (e.g., a recent study suggested over 60 years³⁷) and the draft product monograph recommends that treatment should continue until disease progression or unacceptable toxicity occurs, a median of 37.7 months might not be long enough to determine the long-term effects of treatment with belzutifan. In addition, the long-term effects following treatment discontinuation (e.g., due to patient preference with stable disease, need for surgery) are unknown.

Indirect Evidence

Contents within this section have been informed by materials submitted by the sponsor. The following have been summarized and validated by the CADTH review team.

Objectives for the Summary of Indirect Evidence

The pivotal evidence submitted by the sponsor was a single-arm trial³⁸ and therefore does not provide evidence of efficacy compared to active surveillance in the Canadian health care setting. The sponsor identified no published evidence for the efficacy of the standard active surveillance therapeutic approach used in the Canadian setting that could be used for an indirect comparison with the belzutifan efficacy outcomes assessed in the LITESPARK-004 trial. The sponsor therefore conducted the VHL Natural History Study³⁹ and submitted an ITC to supplement the lack of comparative evidence of belzutifan for the treatment of adult patients with VHL disease.⁴⁰ The sponsor noted that a MAIC was chosen, rather than other potentially relevant approaches, because the external research group conducting the ITC had access to individual patient data from the comparator study but not from the LITESPARK-004 trial.

Description of Indirect Comparison — MAIC

The submitted ITC compared the VHL Natural History Study,³⁹ a real-world, retrospective, noninterventional cohort study of existing medical records of VHL patients managed and treated at the NCI in the US or Canada, to the pivotal trial cohort from the LITESPARK-004 trial. The Studies Addressing Gaps in the Pivotal and RCT Evidence section of this report provides additional details on the VHL Natural History Study. Patients in the VHL Natural History Study cohort were reweighted to match the distribution of key baseline characteristics among patients in the LITESPARK-004 trial and compared using a MAIC.⁴¹ The comparative treatment was active surveillance, defined as enrolment into the database. The primary (and only) outcome of interest was time to RCC-related surgery.

Objectives

The objective of the ITC was to estimate the comparative treatment effect between belzutifan and active surveillance on the rate of RCC surgery in participants with VHL disease–associated tumours using a MAIC.⁴⁰

Study Selection Methods

A targeted literature review conducted by the sponsor did not identify any published study that could be used to generate comparator data that would provide a benchmark against the results of the LITESPARK-004 trial in the ITC. Due to this gap in evidence, the VHL Natural History Study was conducted, and unpublished data from this study were used in the MAIC. After the initiation of the Natural History Study, an SLR with a search date of June 2022 was conducted by the sponsor. This SLR confirmed the finding of the targeted literature search conducted before the Natural History Study. There is no indication about how data were extracted for use in the MAIC, and no risk-of-bias assessment of the Natural History Study was performed.

Indirect Treatment Comparison Analysis Methods

The population considered for LITESPARK-004 was the efficacy analysis set, which included all participants who met the eligibility criteria and received at least 1 dose of the study intervention. The VHL Natural History Study cohort consisted of patients with nonmetastatic VHL disease from the US or Canada and treated at the NCI with at least 1 solid renal tumour and available measurement(s) during the study period (July 31, 2004, to June 30, 2020). The eligibility criteria described in Table 16 were applied to the VHL Natural History Study in attempt to match the selection criteria more closely to those of the LITESPARK-004 trial. Most inclusion and exclusion criteria overlapped between cohorts, namely, inclusion criteria based on age, VHL diagnosis, RCC tumour, and exclusion criteria based on prior anticancer therapy, metastatic disease progression, immediate prior surgery, or prior HIF-2-alpha inhibitor use. Several criteria in the Natural History Study cohort are approximate to those of the pivotal trial due to insufficient access to information within the database. For example, the Natural History Study cohort excluded patients who had surgery within 60 days of the index date to align with the LITESPARK-004 criteria to only include RCC tumours that do not necessitate surgery. The LITESPARK-004 cohort included additional inclusion and exclusion criteria that selected patients at low risk of AEs; these are included in Table 5.

A MAIC analysis was conducted based on individual patient data from the VHL Natural History Study and aggregate data from the LITESPARK-004 trial. The 260 patients included in the external cohort were reweighted to match the distribution of key baseline characteristics using the method-of-moment approach among patients in the LITESPARK-004 trial.⁴⁰ The baseline variables used for weighting adjustment were as follows: age at patient-level index date, sex, number of renal surgeries with therapeutic intent before patient-level index date, and tumour size of the largest renal solid tumour at the patient-level index date. The variables were selected based on input from clinical experts on patient baseline characteristics that were considered prognostic of transition probabilities starting from the presurgery state, or that may modify the effect of belzutifan on these transition probabilities. The consulted experts indicated that VHL type, VHL gene alteration type, and the number of measured tumours were important effect-modifying or prognostic factors, but these were not included in the matching procedure because they were either unavailable or affected by a high proportion of missing data.

Within the Natural History Study cohort, the index date was defined as the earliest date that a measurable renal solid tumour was detected during the study period (July 31, 2004, to June 30, 2020), in contrast to the LITESPARK-004 trial, which defined the index date as the start of study treatment. In the LITESPARK-004 trial, the outcome was defined as RCC surgery from any procedure, excluding radiation, which led to a reduction in RCC tumour size. Patients who did not undergo surgery during the follow-up period were censored at the date they were last known to be alive. For the Natural History Study, the outcome was a surgical procedure defined as: complete nephrectomy (bilateral, left, right), partial nephrectomy (bilateral, left, right), ablation (not otherwise specified), microwave ablation, radiofrequency ablation, cryotherapy, renography, resection, surgery (not otherwise specified), and other-cyst removal. Patients were censored at the end of follow-up or at the time of experiencing a competing event type (metastatic disease progression or death).

The cause-specific hazard of RCC surgery for belzutifan was parametrically modelled using the observed time from the baseline visit until the first postbaseline renal surgery in the overall LITESPARK-004 trial population (N = 61). Similarly, for active surveillance, the rate of RCC surgery was estimated using patient-level time-to-event data from the VHL Natural History Study, which were reweighted to match the baseline characteristics of the LITESPARK-004 trial population. Specifically, the exponential distribution was selected due to the small number of observed events in the LITESPARK-004 trial and based on visual inspection and statistical goodness-of-fit, which was based on the minimized mean squared error for the VHL Natural History Study.

No sensitivity or subgroup analyses were conducted. No assessment of validity of method assumptions, potential unmeasured confounding, or transportability were provided. No mention was made of missing data or if any patients in the VHL study were lost to follow-up before the final calendar date.

Results of Sponsor-Submitted Matching-Adjusted Indirect Comparison

Summary of Included Studies

The distribution of baseline characteristics before and after reweighting is displayed in Table 17. After reweighting, the effective sample size from the Natural History Study decreased by 65% (from 260 to 92.2), as patients had fewer prior surgeries on average compared with their corresponding LITESPARK-004 trial samples before matching. There appear to be similarities in the balance of sex and age at time of enrolment. While the number of prior surgeries and size of tumour differed in the original sample, the reweighting balanced these covariates. Other potentially relevant baseline characteristics were not presented in the sponsor-submitted report.

Table 16: Sample Selection Process of the Natural History Study Population Used to Match LITESPARK-004

HIF-2-alpha = hypoxia-inducible factor 2 alpha; VHL = von Hippel-Lindau.

Source: Indirect Treatment Comparison Report (2022).⁴⁰

Table 17: Baseline Characteristics of the Natural History Study and LITESPARK-004 Trial Populations Before and After Matching

^aEffective sample size is computed as the square of the summed weights divided by the sum of the squared weights.

Source: Indirect Treatment Comparison Report (2022).⁴⁰

Results

In the LITESPARK-004 cohort, 7 RCCs surgeries occurred during the follow-up; 2 patients were censored for death. In the VHL Natural History Study, 196 tumour-reduction surgeries events were observed, and the numbers of patients censored due to metastatic disease or death were 2 and 0, respectively. The weekly exponential rate of RCC surgery was estimated at 0.00487 (SE = 0.00034) in the matched Natural History Study sample versus 0.00071 (SE = 0.0003) in the LITESPARK-004 population. Figure 9 illustrates the difference in estimated time to RCC surgery distribution using the LITESPARK-004 and reweighted Natural History Study data.

Figure 9: Cause-Specific Hazard Using Exponential Distribution of RCC Surgery Based on Data From LITESPARK-004 (Belzutifan) and the VHL Natural History Study (Routine Surveillance)

Source: Indirect Treatment Comparison Report (2022).⁴⁰

Critical Appraisal of VHL Natural History Study Matching-Adjusted Indirect Comparison

Internal Validity

The sponsor-submitted MAIC was informed by patient-level data from the VHL Natural History Study and the aggregate data of the LITESPARK-004 cohort discussed in the Pivotal Studies section. There are discrepancies between the studies in the definition of outcome events and index dates. The outcome definition of RCC surgery within the Natural History Study cohort is subject to potential measurement error. Specifically, the clinical experts indicated that renography and cyst removal are not definitive surgical interventions for management of RCC, although these were considered an RCC surgery outcome in the VHL Natural History Study. The magnitude of bias due to measurement error is unknown but may overestimate the estimated relative rate of RCC surgeries in favour of belzutifan. The clinical experts noted that the definition of RCC tumour surgery used in the LITESPARK-004 trial was more restrictive compared to the Natural History Study. In the LITESPARK-004 trial the index date is treatment assignment, whereas the Natural History Study index date is the date of identification of an RCC tumour. Beyond the differences in the outcome, there is potential endogeneity in surgery event rates between populations due to unmeasured factors related to tumour identification and measurement. The clinical experts noted that the decision to perform an RCC tumour surgery is dependent on the anatomic location of resections, baseline renal function after prior surgeries, comorbidities, and other factors. In addition, 77% of the LITESPARK-004 trial population had a prior RCC surgery (Table 10), whereas the Natural History Study cohort consisted of patients at their first identifiable RCC tumour. The populations may be heterogenous, and it is unclear if the estimated rates of RCC surgery are informative of comparative efficacy.

The use of a cause-specific hazard model requires including the effect modifiers of the active treatment on each outcome. The sponsor noted that “Belzutifan is expected to improve OS mainly by preventing surgeries and distant metastases, but also by reducing the risk of transitioning directly from presurgery to death in some cases.” By only including a few potential prognostic factors of surgery events there is likely heterogeneity in the estimates of metastatic and death that result in a biased estimate of the RCC surgery event rate. Additionally, some important effect-modifying or prognostic factors were left out of the matching and reweighting because they were either unavailable or affected by a high proportion of missing data. A key underlying assumption of the unanchored MAIC approach is that all prognostic and effect-modifying factors are balanced across treatment arms. Given that this assumption was likely to have not been met and there was no information on the potential degree of residual confounding, the estimates of RCC surgery rates may be substantially biased. It is unclear why the estimated SEs are so similar (0.0003 versus 0.00034) when there is a notable difference in the number of events between groups. The LITESPARK-004 cohort had 7 events whereas the Natural History Study had 196 (Figure 9). An extrapolation from a sample of 7 events provides little reliable evidence of the treatment’s comparative efficacy on surgery rates. Because the estimated trial event rate of 7 RCC surgeries is low relative to the amount of data collected (n = 61, median follow-up 37.7 months) and the effective sample size of the VHL Natural History Study was nearly halved (n = 92.2), the robustness of the reported findings is uncertain. The choice of an exponential model was based on visual inspection and a mean squared error criteria; these are inadequate for the purpose of measuring comparative effectiveness and likely overestimate the precision and effect size within treatment groups.⁴²

External Validity

The MAIC analysis was conducted after a literature search was unable to identify existing studies with which to contrast the efficacy of belzutifan. The cohort from the VHL Natural History Study was recruited from 2004 to 2020, and the clinical experts indicated that a diagnosis of VHL, treatment regimens for VHL, and the characteristics of patients undergoing VHL-associated RCC tumour surgeries would be similar throughout this time frame. The selection criteria that informed the Natural History Study subcohort was intended to match with those from the LITESPARK-004 cohort but did not include some key criteria. The Natural History Study cohort did not include restrictions on ECOG PS scores of 0 or 1, and this difference may have biased the effectiveness in favour of belzutifan. It was difficult to assess the degree of heterogeneity between the included studies based on the sponsor-provided technical report because reporting of study design and patient characteristics was limited. Key covariates from the pivotal trial, such as ECOG PS, were not reported, and it is therefore unclear whether these would have been balanced after the matching. For active surveillance in the VHL-associated CNS hemangioblastoma and pNET populations, an important limitation of the VHL Natural History Study data was the inability to identify patients who had CNS hemangioblastoma and pNET tumours at the patient-level index date. The ITC results therefore do not contribute to the evidence for the efficacy of belzutifan in VHL patients who had CNS hemangioblastoma and pNET tumours.

Studies Addressing Gaps in the Pivotal and Randomized Controlled Trial Evidence

Contents within this section have been informed by materials submitted by the sponsor. The following have been summarized and validated by the CADTH review team.

The sponsor submitted 1 real-world, retrospective cohort study³⁹ to address the lack of published clinical trials or observational data on the efficacy outcomes of the standard of care (active surveillance) that could be used for an indirect comparison with the belzutifan efficacy outcomes assessed in the LITESPARK-004 trial. This study provided the active surveillance efficacy data used for the ITC discussed in the previous section (Table 18).

This section provides a summary of a real-world VHL Natural History Study³⁹ based on a retrospective cohort of VHL patients managed and treated at NCI. The study reported on the growth kinetics and surgical patterns within these patients.

The sponsor also provided a cross-sectional HRQoL survey⁴³ to address the lack of HRQoL or utility values, and this is summarized and appraised in Appendix 1. This study assessed the impact of VHL disease on HRQoL as measured using the EQ-5D questionnaire in patients with RCC, CNS hemangioblastoma, or pNET. A total of 220 patients completed the survey. Overall, patients with VHL-associated tumours had a mean EQ-5D score of 0.771. As patients who participated in this study did not receive belzutifan, this study does not provide information related to the effect of treatment with belzutifan on the HRQoL of patients with VHL disease.

Populations

The primary study population consisted of US and Canadian patients treated at the NCI with confirmed VHL syndrome and at least 1 renal solid tumour with available measurement(s) during the study period (July 31, 2004, to June 30, 2020). The patient-level index date was defined as the first date that a patient had a radiology report showing at least 1 measurable renal solid tumour. Additional criteria were applied in attempt to match the study population more closely to the population enrolled in the LITESPARK-004 trial (i.e., to remove patients with investigational therapy, oncologic therapy, or renal tumour-reduction procedures within 30 days before the index date). This was referred to as the “trial population subgroup.” Patients were followed until their last clinical encounter or date of death, whichever occurred first.

This was a noninterventional study with patients treated with usual care (active surveillance). No systemic therapies, including belzutifan, were allowed at study entry.

Table 18: Summary of Studies Addressing Gaps in the Evidence

IQR = interquartile range; NCI = National Cancer Institute; RCC = renal cell carcinoma; VHL = von Hippel-Lindau.

Note: Details from the table have been taken from the sponsor’s Summary of Clinical Evidence.

Source: von Hippel-Lindau Natural History Study.³⁹

Outcomes

The primary end point was the LGR of renal solid tumours among patients who had at least 3 serial assessments of renal tumour diameter (1 initial measurement and at least 2 subsequent measurements) for unique tumour(s) before the first of: date of death or last clinical encounter, initiation of investigational or oncologic therapy, or first renal tumour-reduction procedure affecting the tumour. Patient-level and tumour-level LGR for renal solid tumours were assessed. Tumour size metrics were collected throughout follow-ups.

Tumour size was defined as the longest diameter of the tumour in centimetres and was ascertained from the structured database. Tumour size was ascertained for all individual renal solid tumours with a serial measurement from the tumour-level index date until the end of tumour-level follow-up. The date of the renal solid tumour measurement was also collected.

Secondary end points included an assessment of the proportion of patients with tumour-reduction procedures, time from first tumour detection in the study period to first tumour-reduction procedure, frequency and type of tumour-reduction procedures, time between tumour-reduction procedures, and kidney function.

Occurrence of chronic kidney disease, anemia and patients qualifying for dialysis at key time points were also assessed as other secondary outcomes.

Statistical Analysis

In the primary analysis, LGRs for renal solid tumours were calculated using a linear mixed (multilevel) model regressing tumour size on time since the tumour-level index date, accounting for dependency of measurements within tumour and patient. The multilevel model described LGRs at 3 levels: unique patients, unique tumours nested within unique patients, and tumour measurements nested within unique tumours. LGRs were presented continuously and categorically at the patient, tumour, and population level. A secondary LGR analysis was also performed using simple linear regression models used in the LITESPARK-004 trial.

Descriptive characteristics of the number and proportion of patients with at least 1 and at least 2 tumour-reduction procedures (overall and among surgical procedures only), number of procedures during follow-up, and surgical complications (including estimated blood loss and perioperative mortality) were described. Time from the patient-level index date to first renal tumour-reduction procedure (overall and separately among surgical procedures only), and between patients’ first and second renal tumour–reduction procedures were described using Kaplan-Meier metrics and curves.

Results

Patient Disposition

Of 776 VHL patients in the NCI hereditary database, 308 patients with at least 1 solid renal tumour met the eligibility criteria and were included in the primary study population. After applying additional eligibility criteria mentioned previously (focusing on the tumour growth rate assessment), 247 patients (80.2%) were included in the trial population subgroup.

Subgroups of 131 patients and 114 patients in the primary study population and the trial population subgroup, respectively, had at least 3 serial measurements of at least 1 solid renal tumour during the study period that qualified them for inclusion in the LGR analysis subgroups to address the primary research objective.

Baseline Characteristics

Patients in the primary study population had a median age of 41 years (range = 18 to 79) at the patient-level index date, and 56% were male. The median length of follow-up was 107.2 months (approximately 9 years; range = 0.03 to 190.2 months), with 88% and 71% having at least 2 and 5 years of follow-up, respectively. At the patient-level index date, other clinical manifestations of the VHL syndrome included renal cysts (92% of patients), brain and spine hemangioblastomas (79% and 71%, respectively), pancreatic cysts (77%), retinal hemangioblastomas (57%), pancreatic solid tumours (34%), epididymal cystadenomas (26% of males), and endolymphatic sac tumours (9%). The median number of renal tumour–reduction procedures before the index date was 1 (range = 0 to 9). Demographics and clinical characteristics were similar for patients in the trial population subgroup and the subpopulation of patients eligible for the LGR analysis (i.e., with at least 3 serial measurements). However, many demographic characteristics, such as race and ethnicity and ECOG PS, were not reported in the Natural History Study report.

Efficacy

Linear Growth Rate

The median tumour-level LGRs for the primary study population and trial population subgroup were 0.38 cm per year (IQR = 0.30 to 0.49) and 0.37 cm per year (IQR = 0.29 to 0.47), respectively. In a secondary analysis using a simple linear model and multilevel modelling, median tumour-level LGRs for the primary study population and the trial population subgroup were 0.36 cm per year (IQR = 0.25 to 0.57) and 0.36 cm per year (IQR = 0.24 to 0.56), and 0.38 cm per year (IQR = 0.30 to 0.49) and 0.37 cm per year (IQR = 0.29 to 0.47), respectively.

Frequency and Type of Tumour–Reduction Procedures

Of the 308 patients in the primary study population, 232 patients (75.3%) had at least 1 renal solid tumour–reduction procedure during the study period, including 225 patients (73.1%) with surgical procedures (96% of which were partial nephrectomies), 16 patients (5.2%) with ablation procedures, and 1 patient (0.3%) who received radiation. Similar tumour resection patterns were observed in the trial population subgroup, and among the 184 patients (74.5%) who underwent at least 1 renal solid tumour–reduction procedure; 62% had at least 2 procedures during the overall follow-up, including 11 of 71 patients (15.5%) and 34 of 138 patients (24.6%) with at least 2 procedures within 2 years and 5 years of follow-up, respectively. The median number of tumour-reduction procedures per patient in the trial population subgroup was 2 (range = 1 to 9). At 2 years and 5 years of follow-up for patients in the trial population subgroup, the median numbers of procedures were 1 (range = 1 to 3) and 1 (range = 1 to 4), respectively.

Time to Tumour–Reduction Procedures

The 1-year, 2-year, 5-year, and 7-year intervention-free survival probabilities for the first tumour-reduction procedure were 79.0%, 69.9%, 38.3%, and 26.4%, respectively, for the trial population subgroup. The 25th percentile of the time to first tumour-reduction procedure was 17.5 months (95% CI, 9.13 to 24.08 months) in the trial population subgroup, with a median time-to-event of 44.2 months (95% CI, 35.74 to 49.51 months).

Harms

Of the 217 patients (70.5%) in the primary study population with at least 1 partial nephrectomy, 413 partial nephrectomies were performed during follow-up, 124 (30.0%) of which were associated with complications. The median estimated blood loss, assessed among all surgical procedure types, was 1.5 L (IQR = 0.6 to 2.6). The maximum estimated blood loss observed was 12 L in the primary study population. Two patients (0.9%) in the primary study population with at least 1 renal tumour–reduction procedure conducted at the National Institutes of Health died within 30 days of the procedure (1 nephrectomy and 1 biopsy)

Critical Appraisal

Internal Validity

The VHL Natural History Study was a real-world retrospective cohort study focusing on the natural history and progression of renal solid tumours associated with VHL syndrome. However, this study did not evaluate the effect of belzutifan and did not provide evidence about the efficacy of the treatment. No a priori protocol for the analyses appeared to be present. The study was conducted using data registered by the NCI in a hereditary database of patients with VHL syndrome, but how these data were located (e.g., which search methods were used) or selected is not specified.

Some of the limitations of this study were high levels of missing data for some variables; unavailability of longitudinal measures of tumour growth for all tumours in a systematic manner (i.e., differences may have arisen due to variation in measurement across observers), as the measures were extracted from the registry directly; incomplete documentation of metastasis; and the possibility of misclassification. There is also a possibility of loss to follow-up of patients who left the registry, but no relevant information was provided. Attempts were made to ensure that the trial population subgroup would be similar to the population of the LITESPARK-004 trial, based on the additional criteria that were applied. However, several criteria in the Natural History Study cohort approximate those of the pivotal trial due to insufficient access to information within the database. While demographics and clinical characteristics were similar for patients in the trial population subgroup and patients in the LGR analysis subgroups, many demographic characteristics (i.e., race and ethnicity and ECOG PS) were not reported in the Natural History Study report.

Receiving systemic oncologic therapy or investigational therapy within 30 days on or before the patient-level index date was part of the exclusion criteria. However, no specific information was provided about the type of systemic or investigational therapy. Moreover, this exclusion criteria may have led to the exclusion of patients with a better prognosis, which would have affected the results for the trial population subgroup.

External Validity

This study includes patients managed and treated at the NCI only. This may affect the overall generalizability of the study, as it may not be representative of all patients with VHL disease in Canada that fall within the indication. Patients selected for this study might have received a higher quality of care compared with patients treated at other centres, which might have led to lower treatment complication rates and overall morbidity associated with standard of care at the NCI compared with that generally expected at other clinical centres.

While this study presented data for renal tumours, patients included in the study also had other tumours associated with VHL disease, (e.g., renal cysts, brain, and spine hemangioblastomas, pancreatic cysts, retinal hemangioblastomas, and pancreatic solid tumours). However, no additional information was provided for these tumours, and some important clinical outcomes associated with VHL syndrome may have been excluded. As the indication under review includes patients with nonmetastatic pNETs and CNS hemangioblastomas in addition to nonmetastatic RCCs, the unavailability of data related to these 2 tumours represents a gap in the evidence provided by this study.

The authors of this study noted the potential risk of losing substantial numbers of samples due to the eligibility criteria of at least 3 serial measurements for the assessment of tumour growth rate patterns. They added that, although patients with at least 3 measurements did have longer follow-up times, on average, no substantial differences in baseline demographic and clinicopathologic characteristics were observed compared to patients with fewer than 3 measurements. Considering the substantial variation in diagnostic techniques and treatment approaches over time, the authors also raised concerns regarding whether the growth rate is generalizable to the current patient populations.

Discussion

Summary of Available Evidence

One ongoing, phase II, single-arm, open-label trial (LITESPARK-004, N = 61) was included in the SLR conducted by the sponsor. The primary objective of the LITESPARK-004 trial was an evaluation of the efficacy of belzutifan for the treatment of VHL disease–associated nonmetastatic RCC as measured by ORR. Secondary end points included ORR for patients with CNS hemangioblastoma (n = 50) and nonmetastatic pNET (n = 22); TTR, DOR, TTS, and PFS; and safety, including notable harms (i.e., anemia, hypoxia, secondary primary malignancies, and hepatic safety). The trial enrolled adults with VHL disease based on a germline VHL alteration, with at least 1 measurable RCC but without a need for immediate surgery or evidence of metastasis. Patients could have other VHL disease–associated non-RCC tumours. Efficacy analyses were conducted for the full (RCC) population, as well as for subsets of the population with a CNS hemangioblastoma or nonmetastatic pNET. Patients in the total study population (i.e., those with RCC) had a median age of 41.0 years (range = 19.0 to 66.0), and about half were male (52.5%) and half were female (47.5%). Most patients were white (90.2%), had an ECOG PS of 0 (82.0%), and had undergone a previous surgery or ablative procedure (96.7%). Reported baseline characteristics for CNS hemangioblastoma and pNET subpopulations were comparable to those of the full population.

An ITC was submitted by the sponsor to provide comparative evidence of belzutifan versus active surveillance for the treatment of adult patients with VHL disease in the LITESPARK-004 trial. The ITC compared a real-world, retrospective, noninterventional cohort study (the VHL Natural History Study) with the LITESPARK-004 trial. Patients with VHL-associated RCC from the VHL Natural History Study were reweighted in an attempt to match the distribution of key baseline characteristics among patients with VHL-associated RCC in LITESPARK-004 and compare them using an unanchored MAIC analysis. The primary and only outcome assessed in the MAIC was time to RCC-related surgery.

The VHL Natural History Study, which provided the active surveillance efficacy data used for the ITC, was also submitted by the sponsor as a standalone study addressing gaps in the pivotal and RCT evidence. In addition, the sponsor provided a cross-sectional HRQoL survey, which assessed the impact of VHL disease on HRQoL as measured by an EQ-5D questionnaire given to patients with an RCC, CNS hemangioblastoma, or pNET.

Interpretation of Results

Efficacy

Belzutifan is a first-in-class HIF inhibitor and the first systemic treatment indicated for the treatment of VHL disease. Overall, belzutifan may be a safe treatment that can provide added clinical benefit for patients with a VHL-associated nonmetastatic RCC, those with a VHL-associated CNS hemangioblastoma, and those with a VHL-associated nonmetastatic pNET, none of whom required immediate surgery. Given the rarity of VHL disease and that active surveillance is currently the standard of care for patients with VHL-associated nonmetastatic tumours, the small sample size and single-arm trial design adopted in the LITESPARK-004 trial was considered acceptable from a regulatory perspective to demonstrate the efficacy of belzutifan, but drawing definitive conclusions is challenging due to the potential for confounding. This review’s interpretation of the efficacy findings in the LITESPARK-004 trial were based on the opinions of the clinical experts consulted by CADTH and the judgment of the CADTH clinical review team. Most importantly, this interpretation should not serve as confirmation of the efficacy of belzutifan; the efficacy of belzutifan remains uncertain until a phase III RCT is conducted.

ORR and DOR observed in single-arm studies may serve as substantial evidence when there is no therapy available and presuming that major tumour regressions can be attributed to the tested treatment.¹⁴ Patients placed importance on the need for a treatment that can improve their physical condition by decreasing or stabilizing the size of tumours as well as offering long-term stability or reduction of disease. In the LITESPARK-004 trial, the magnitude of the ORR (i.e., the percentage of patients who had a CR or PR in response to belzutifan) was greater than the prespecified alternative hypothesis of 30% and considered clinically meaningful for patients with VHL-associated nonmetastatic RCC, those with VHL-associated CNS hemangioblastoma, and those with VHL-associated nonmetastatic pNET. The LITESPARK-004 trial demonstrated a greater tumour response rate when patients who had stable disease were added to those with CR and PR in the DCR end point: a DCR of 90% or greater was observed across patients with an RCC, those with a CNS hemangioblastoma, and those with a pNET. However, it is not possible to determine whether stable disease can be attributed to belzutifan or the natural history of the disease. IRC- and investigator-assessed ORR and DCR were consistent among patients with nonmetastatic RCC and those with CNS hemangioblastoma. Moreover, based on the DOR, the clinical experts consulted by CADTH considered the tumour response to be durable: among 39 patients with RCC and a confirmed response, 32 (82.1%) were censored due to no progression at the time of data cut-off or before end of treatment. Of 22 patients with a CNS hemangioblastoma who showed confirmed response, 17 were censored due to no progression at the time of data cut-off or before the end of treatment. All 20 patients with a pNET who had a confirmed response had no progression at the time of data cut-off.

In addition to ORR and DOR, the clinical experts considered TTS and PFS to be important and clinically relevant outcomes in the assessment of the efficacy of belzutifan in patients with VHL-associated nonmetastatic lesions because prolonging survival and avoiding morbid local therapies such as surgeries and radiation as much as possible are among the most critical treatment goals in clinical practice. Similarly, patients indicated that the opportunity to avoid surgery is important to them, and they expressed their willingness to tolerate moderate-to-severe side effects if a treatment would allow them to avoid surgery to remove VHL-associated cysts and tumours. In the LITESPARK-004 trial, medians of TTS and PFS were not reached, yet the efficacy of belzutifan appeared to be supported because the majority of patients were censored for PFS due to no progression during the follow-up period. The clinical experts consulted by CADTH indicated that the rates of PFS and TTS appeared to be more promising than might be expected with standard of care during the time period of follow-up (i.e., median 37.7 months). However, the follow-up time was not long enough to reveal the full extent of the efficacy of belzutifan on time-to-event outcomes. Given that the life expectancy of patients with VHL could be long (e.g., a recent study suggested > 60 years³⁷), the length of follow-up in the trial was insufficient to determine the long-term effects of treatment with belzutifan. In addition, potential treatment effects following discontinuation (e.g., due to patient preference with stable disease and need for surgery) are unknown.

The outcome — distribution of all tumour-reduction procedures before and after treatment initiation — was not prespecified in the protocol of the LITESPARK-004 trial. Fewer tumour-reduction procedures occurred after patients received belzutifan, compared to the same period immediately before treatment initiation. This finding suggests the potential of belzutifan to delay the need for tumour-reduction procedures among patients with VHL. However, this finding, which involved VHL-associated tumours not limited to RCCs, CNS hemangioblastomas, and pNETs, was beyond the scope of the present reimbursement request.

Findings from the sponsor-conducted ITC, which used the VHL Natural History Study to provide an external comparator for the LITESPARK-004 trial, were considered highly uncertain. Although the results for the estimated number of surgeries were in favour of belzutifan compared with active surveillance and considered clinically meaningful, several major limitations decreased CADTH’s confidence in the evidence. First, the selection criteria that informed the VHL Natural History Study subcohort that was intended to match the LITESPARK-004 cohort did not include some key criteria, and the Natural History Study cohort did not include restrictions on ECOG PS scores of 0 or 1, which creates a risk of bias in effectiveness that favours belzutifan. Second, it was difficult to assess the degree of heterogeneity between the included studies based on the sponsor-provided technical report as reporting of study design and patient characteristics was limited. Third, key covariates from the pivotal trial, such as ECOG PS, were not reported, and the potential for residual confounding was not assessed and can be considered high. Fourth, the outcome definition of RCC surgery within the Natural History Study cohort is subject to potential measurement error. Specifically, the clinical experts indicated that renography and cyst removal are not definitive surgical interventions for management of a RCC, although these were considered RCC surgery outcomes in the VHL Natural History Study. The magnitude of bias due to measurement error is unknown but may overestimate the estimated relative rate of RCC surgeries in favour of belzutifan. Finally, the MAIC analysis did not report on surgeries related to CNS hemangioblastomas and pNETs at the patient-level index date.

Findings from the VHL Natural History Study and the HRQoL survey, which were submitted by the sponsor as studies addressing gaps in the pivotal and RCT evidence, did not provide information on the efficacy or safety of belzutifan as standalone studies because patients did not receive belzutifan in either of the 2 studies.

Patients placed a high value on new treatments that could improve their quality of life. HRQoL was not assessed in the LITESPARK-004 trial, the VHL Natural History Study, or the ITC. This gap in the evidence was acknowledged by the sponsor in its submission, and the sponsor submitted a cross-sectional HRQoL survey in response. Given that patients who participated in the HRQoL survey did not receive belzutifan, no information on the effects of belzutifan on HRQoL is available, and the effect of belzutifan on HRQoL in patients with VHL disease is unknown.

Harms

The challenge posed by the lack a comparator group in the LITESPARK-004 trial also applies to the interpretation of harms outcomes of belzutifan and decreases CADTH’s confidence in the findings. Almost all patients (95%) in the trial experienced anemia, and about one-third of these required treatment with either iron supplements or an erythropoiesis-stimulating drug, while blood transfusion was uncommon (2%). According to the clinical experts consulted by CADTH, the incidence of harms outcomes, including AEs of clinical interest (i.e., anemia, hypoxia, secondary primary malignancies, and hepatic events), within a median of the 37.7-month follow-up period was generally acceptable.

Regarding long-term safety, the clinical experts noted that the current follow-up time of the LITESPARK-004 trial appeared to be sufficient to observe most side effects. However, given the expectation that patients may take belzutifan for years, its long-term safety with respect to delayed toxicities or uncommon AEs is uncertain. Moreover, the harms caused by the treatments used for belzutifan-related side effects may also cause harms among patients. For example, erythropoiesis-stimulating drugs, which are used to treat anemia, may increase the risk of developing secondary malignancies.⁴⁴ To further substantiate the safety of belzutifan, more data are required from the ongoing LITESPARK-004 trial (with 38 patients on the treatment of belzutifan) and other sources, such as phase III RCTs and real-world safety data.

Neither the ITC nor the studies addressing gaps in the pivotal and RCT evidence provided evidence about the safety of belzutifan. No harms outcomes were reported in the ITC, and belzutifan was not investigated in the VHL Natural History Study or the HRQoL study.

Conclusion

Based on the opinions of clinical experts, evidence from the phase II, single-arm, open-label LITESPARK-004 trial suggests that treatment with belzutifan may have a beneficial effect in adult patients with VHL-associated nonmetastatic RCC with or without a VHL-associated CNS hemangioblastoma and/or nonmetastatic pNET not requiring immediate surgery. The clinical experts indicated that the efficacy results from the patient populations with a VHL disease–associated nonmetastatic RCC, RCC and CNS hemangioblastoma, RCC and nonmetastatic pNET, as well as RCC, CNS hemangioblastoma, and pNET in the LITESPARK-004 trial indicate a promising and durable tumour response as measured by ORR and DOR, which is likely attributable to belzutifan, based on their clinical experience and expectations of the natural history of the disease. However, the absence of a comparator group in the pivotal study precluded definitive conclusions about the efficacy and safety of belzutifan, due to the potential for confounding. There was uncertainty in the findings for other supportive outcomes (e.g., TTS, PFS) due to limitations such as the lack of an internal comparator group and no inferential statistical analyses. The incidence of harms, such as anemia and hypoxia, were acceptable according to the clinical experts. Because VHL is a life-long condition and patients can receive belzutifan for an extended period, a longer follow-up would be needed to better understand the long-term benefits and harms of belzutifan, as well as the potential residual effects following discontinuation. Findings from an ITC suggest that treatment with belzutifan may have a beneficial effect on time to RCC-related surgery compared to active surveillance in patients with VHL disease–associated nonmetastatic RCC. However, the certainty of the indirect evidence was low due to a violation of the underlying assumptions of the MAIC approach that resulted in a high risk of residual confounding, and important differences in outcome definitions between the LITESPARK-004 trial and the VHL Natural History Study that would bias the findings in favour of belzutifan. In the ITC analysis, the comparative efficacy of belzutifan versus active surveillance for VHL-associated CNS hemangioblastoma and nonmetastatic pNET was not assessed and is therefore unknown. The VHL Natural History Study and the VHL HRQoL survey did not provide evidence describing the efficacy and safety of belzutifan as standalone studies because patients in these studies did not receive belzutifan. It is unknown whether treatment with belzutifan would improve patient quality of life.

References

1.Sponsor Summary of Clinical Evidence [internal sponsor's report]. In: Drug Reimbursement Review sponsor submission: Belzutifan, 40 mg and oral tablet. Kirkland (QC): Merck Canada Inc.; 2023 Jan 23.

2.Welireg (belzutifan): 40 mg oral tablets [product monograph]. Kirkland (Québec): Merck Canada Inc.; 2022 Dec 13.

3.Crespigio J, Berbel LCL, Dias MA, et al. Von Hippel-Lindau disease: a single gene, several hereditary tumors. J Endocrinol Invest. 2018;41(1):21-31. PubMed

4.Orphanet. Von Hippel-Lindau Disease. 2023; https://www.orpha.net/consor/cgi-bin/OC_Exp.php?Lng=EN&Expert=892#:~:text=Disease%20definition%20Von%20Hippel-Lindau%20disease%20%28VHL%29%20is%20a,hemangioblastoma%2C%20renal%20cell%20carcinoma%20%28RCC%29%2C%20and%20pheochromocytoma.%20ORPHA%3A892. Accessed 2022 Jun 3.

5.National Organization for Rare Disorders (NORD). National Organization for Rare Disorders. Von Hippel-Lindau Disease. https://rarediseases.org/rare-diseases/von-hippel-lindau-disease/. Accessed 2022 Nov 8.

6.National Cancer Institute. Von Hippel-Lindau Disease (PDQ®)–Health Professional Version. 2023; https://www.cancer.gov/types/kidney/hp/renal-cell-carcinoma-genetics/vhl-syndrome. Accessed 2022 Jun 2.

7.Plon S, Jonasch E. Clinical features, diagnosis, and management of von Hippel-Lindau disease. In: E. A, Firth H, Perrone R, Gajjar A, Geffner M, eds. UpToDate. Waltham (MA): UpToDate; 2022.

8.National Institutes of Health. National Institute of Neurological Disorders and Stroke. Von Hippel-Lindau Disease (VHL). 2023; https://www.ninds.nih.gov/health-information/disorders/von-hippel-lindau-disease-vhl?search-term=VHL. Accessed 2022 Nov 8.

9.Takami H, Graffeo CS, Perry A, et al. Presentation, imaging, patterns of care, growth, and outcome in sporadic and von Hippel-Lindau-associated central nervous system hemangioblastomas. J Neurooncol. 2022;159(2):221-231. PubMed

10.Lonser RR, Glenn GM, Walther M, et al. von Hippel-Lindau disease. The Lancet. 2003;361(9374):2059-2067. PubMed

11.Hasanov E, Jonasch E. MK-6482 as a potential treatment for von Hippel-Lindau disease-associated clear cell renal cell carcinoma. Expert Opin Investig Drugs. 2021;30(5):495-504. PubMed

12.Louise MBM, Smerdel M, Borgwadt L, et al. von Hippel-Lindau disease: Updated guideline for diagnosis and surveillance. Eur J Med Genet. 2022;65(8):104538. PubMed

13.Clinical Study Report: P004V02MK6482. An Open Label Phase 2 Study to Evaluate PT2977 for the Treatment of von Hippel Lindau Disease-Associated Renal Cell Carcinoma [internal sponsor's report]. Toronto (Ontario): Merck Canada Inc.; 2021 Jan 7.

14.FDA. Clinical trial endpoints for the approval of cancer drugs and biologics - guidance for industry. 2018: https://www.fda.gov/media/71195/download. Accessed 2023 Mar 16.

15.European Medicines Agency. Guideline on the evaluation of anticancer medicinal products in man. 2017: https://www.ema.europa.eu/en/documents/scientific-guideline/guideline-evaluation-anticancer-medicinal-products-man-revision-5_en.pdf. Accessed 2023 Mar 16.

16.Agrawal S, Arora S, Amiri-Kordestani L, et al. Use of Single-Arm Trials for US Food and Drug Administration Drug Approval in Oncology, 2002-2021. JAMA Oncol. 2023;9(2):266-272. PubMed

17.Natural History Report: MK6482. Von Hippel-Lindau Natural History Study [internal sponsor's report]. Kenilworth (New Jersey): Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc.; 2021 Jan 6.

18.Maher ER, Yates JR, Harries R, et al. Clinical features and natural history of von Hippel-Lindau disease. Q J Med. 1990;77(283):1151-1163. PubMed

19.Coco D, Leanza S. Von Hippel-Lindau Syndrome: Medical Syndrome or Surgical Syndrome? A Surgical Perspective. J Kidney Cancer VHL. 2022;9(1):27-32. PubMed

20.Jonasch E, McCutcheon IE, Gombos DS, et al. Pazopanib in patients with von Hippel-Lindau disease: a single-arm, single-centre, phase 2 trial. Lancet Oncol. 2018;19(10):1351-1359. PubMed

21.Jonasch E, McCutcheon IE, Waguespack SG, et al. Pilot trial of sunitinib therapy in patients with von Hippel-Lindau disease. Ann Oncol. 2011;22(12):2661-2666. PubMed

22.Pilié P, Hasanov E, Matin SF, et al. Pilot study of dovitinib in patients with von Hippel-Lindau disease. Oncotarget. 2018;9(34):23390-23395. PubMed

23.Merck. Welireg prescribing information. 2022; https://www.welireghcp.com/mechanism-of-action/. Accessed 2022 Nov 10.

24.American Society of Health-System Pharmacists. AHFS firstReleaseTM. Belzutifan. Am J Health-Syst Pharm. 2021;78(24):2185-2187. PubMed

25.Pelle E, Al-Toubah T, Morse B, Strosberg J. Belzutifan in a Patient With VHL-Associated Metastatic Pancreatic Neuroendocrine Tumor. J Natl Compr Canc Netw. 2022;20(12):1285-1287. PubMed

26.Food and Drug Administration. FDA approves belzutifan for cancers associated with von Hippel-Lindau disease. 2022; https://www.fda.gov/drugs/resources-information-approved-drugs/fda-approves-belzutifan-cancers-associated-von-hippel-lindau-disease. Accessed 2023 Jan 25.

27.Therapeutic Goods Administration. Welireg belzutifan 40 mg film-coated tablet bottle (355338). 2022; https://www.ebs.tga.gov.au/servlet/xmlmillr6?dbid=ebs/PublicHTML/pdfStore.nsf&docid=355338&agid=%28PrintDetailsPublic%29&actionid=1. Accessed 2023 Jan 25.

28.Welireg (belzutifan): 40 mg oral tablets [proposed product monograph]. City (PROV): Merck Canada Inc.; 2023 Jul 10.

29.Lin NU, Lee EQ, Aoyama H, et al. Response assessment criteria for brain metastases: proposal from the RANO group. Lancet Oncol. 2015;16(6):e270-278. PubMed

30.Jonasch E, Donskov F, Iliopoulos O, et al. Belzutifan for Renal Cell Carcinoma in von Hippel-Lindau Disease. N Engl J Med. 2021;385(22):2036-2046. PubMed

31.Jonasch E, Iliopoulos O, Rathmell WK, et al. Abstract #4546, Poster Presentation: LITESPARK-004 (MK-6482-004) Phase 2 Study Of Belzutifan, An Oral Hypoxia-Inducible Factor 2α Inhibitor (HIF-2α), For Von Hippel-Lindau (VHL) Disease: Update With More Than Two Years Of Follow-Up Data. American Society for Clinical Oncology 2022 Annual Meeting; 2022; Chicago, IL, USA.

32.Srinivasan R, Iliopoulos O, Rathmell WK, et al. LBA69 Belzutifan, a HIF-2α Inhibitor, for von Hippel-Lindau (VHL) disease-associated neoplasms: 36 months of follow-up of the phase II LITESPARK-004 study. Annals of Oncology. 2022;33:S1433-S1434.

33.Jonasch E, Donskov F, Iliopoulos O, et al. Belzutifan for renal cell carcinoma in von Hippel–Lindau disease: Supplement. N Engl J Med. 2021;385.

34.Clinical Study Protocol: PT2977-202/MK-6482-004-16. An Open Label Phase 2 Study to Evaluate PT2977 for the Treatment of von Hippel Lindau Disease-Associated Renal Cell Carcinoma [internal sponsor's report]. Dallas (TX): Peloton Therapeutics, Inc.; 2017 Nov 10.

35.Merck Sharp & Dohme Corp. Clinical Study Report P004V02MK6482 (Data cutoff date of April 1, 2022) [internal sponsor's report]. 2022 Nov 15.

36.Statistical Analysis Plan: PT2977-202/MK-6482-004-16. An Open Label Phase 2 Study to Evaluate PT2977 for the Treatment of von Hippel Lindau Disease-Associated Renal Cell Carcinoma [internal sponsor's report]. Dallas (TX): Peloton Therapeutics, Inc.; 2020 Jul 12. Accessed Nov 10, 2022.

37.Binderup ML, Jensen AM, Budtz-Jørgensen E, Bisgaard ML. Survival and causes of death in patients with von Hippel-Lindau disease. J Med Genet. 2017;54(1):11-18. PubMed

38.Merck Sharp & Dohme Corp. Clinical Study Report P004V02MK6482 (Data cutoff date of April 1, 2022) [internal sponsor's report]. 2022 Nov 15.

39.Merck Sharp & Dohme Corp. Von Hippel-Lindau Natural History Study. Final Study Report [internal sponsor's report]. 2021.

40.Merck Canada. Von Hippel-Lindau associated non-metastaticnonmetastatic RCC, CNS hemangioblastomas or non-metastaticnonmetastatic pNET. Indirect Treatment Comparisons Natural Histry Study vs. LS-004 [internal sponsor's report]. 2022.

41.Phillippo D, Ades T, Dias S, Palmer S, Abrams KR, Welton N. NICE DSU Technical Support Document 18: Methods for population-adjusted indirect comparisons in submissions to NICE. Bristol (GB): University of Bristol; 2016: https://research-information.bris.ac.uk/ws/portalfiles/portal/94868463/Population_adjustment_TSD_FINAL.pdf. Accessed 2023 Aug 28.

42.Cawley GC, Talbot NL. On over-fitting in model selection and subsequent selection bias in performance evaluation. The Journal of Machine Learning Research. 2010;11:2079-2107.

43.Adelphi Real World. VHL-related disease impact on quality of life and work productivity in patients with renal cell carcinoma, central nervous system hemangioblastoma, or pancreatic neuroendocrine tumor manifestation. Version 1.0 ed. Rahway (NJ): Merck Sharp & Dohme Corp; 2022 Sep 23.

44.Fallah J, Brave MH, Weinstock C, et al. FDA Approval Summary: Belzutifan for von Hippel-Lindau Disease-Associated Tumors. Clin Cancer Res. 2022;28(22):4843-4848. PubMed

Appendix 1: HRQoL and VHL Disease Burden Study

Note that this appendix has not been copy-edited.

This section summarizes the HRQoL and VHL Disease Burden Study⁴³ assessing the real-world VHL disease impact on quality of life and work productivity in patients with RCC, CNS hemangioblastoma, or pNET (both metastatic and nonmetastatic) using a noninterventional international, cross-sectional survey of 220 patients with these VHL-associated tumours. The sponsors submitted this summary to address the gap of lack of HRQoL or utility values to assess the cost-effectiveness (i.e., cost-utility) of belzutifan compared to active surveillance. However, it should be noted that it does not provide information related to the HRQoL of patients treated with belzutifan.

This study was designed to understand the patient experience and burden of disease with current management practices in the treatment of VHL disease and the treatment preferences of patients with RCC, CNS hemangioblastomas, or pNET manifestations. Specifically, the objectives were to assess the patients’ HRQoL and the patients’ work productivity loss and activity impairment. The data collected in this study included patient perceptions of the burden of VHL and their surgical experiences. The study included patient-reported outcome tools to help describe the burden (direct and indirect) of the condition, including the EQ-5D and its Visual Analogue Scale. Patient demographics, disease status (i.e., self-reported CR, PR, stable disease, progressive disease, or unknown), time since most recent surgery, current tumour type, metastatic status, prescribed medication for VHL-associated cancer and presence of key comorbidities such as chronic pain, chronic kidney disease, neurologic issues, and diabetes, were also assessed.

Data collection was conducted by Adelphi Real World through a patient advocacy group, the VHL Alliance who operates in the US, and screened online. The study sampled adult patients (age 18 years and older) from the US, Canada, UK, France, and Germany with self-reported health care professional (HCP)-diagnosed VHL disease with at least 1 kidney, brain or spinal cord, or pancreas manifestation. More than 200 patients were asked to complete an online structured survey, designed to be completed in 1 sitting between December 2021 and June 2022.

Descriptive statistics were reported depending on the type of variable being assessed:

• Numeric – count, mean, standard deviation, median, 25th percentile, 75th percentile, minimum and maximum.

• Categorical – frequency of responses in each category and percentage of responses in each category.

A total of 220 patients completed the survey (response rate not reported): 108 from the US, 37 from Canada, 21 from UK, 3 from France and 51 from Germany. Mean patient age was 42.5 years, 68.2% of patients were female and the majority of patients were white (88.6%). Time since diagnosis was 210.4 months, 66.4% of patients had RCC, 55.5% of patients had pNET, and 86.4% of patients had CNS hemangioblastoma and 35% had all 3 manifestations.

Overall, patients with VHL-associated tumours had a mean EQ-5D score of 0.771 (SD = 0.21). Patients with metastatic disease (n = 16) had a mean EQ-5D score of 0.667 (SD = 0.28) and patients without metastatic disease (n = 195) had a mean EQ-5D score of 0.783 (SD = 0.20). Patients with VHL-associated tumours had a mean percent overall work impairment of 30.8% (SD = 42.42). Patients with metastatic disease had a mean percent overall work impairment of 40.6% (SD: 43.97) and patients without metastatic disease had a mean percent overall work impairment of 29.3% (SD: 42.04). Patients who had their most recent surgery within the last 6 months had a mean EQ-5D of 0.727 (SD = 0.20) and patients who had their most recent surgery more than 6 months ago had a mean EQ-5D of 0.777 (SD = 0.21). Patients who underwent a surgery within the last 6 months had a mean percent overall work impairment of 53.8% (SD = 51.89) compared to 29.1% (SD = 41.71) for patients who had their most recent surgery more than 6 months ago. Patients with more than 1 tumour type had a mean EQ-5D of 0.776 (SD = 0.21) and patients with only 1 tumour type had a mean EQ-5D of 0.832 (SD = 0.09).

As this was a noninterventional study, it does not provide evidence on the effects of belzutifan on HRQoL. Because the EQ-5D is not specifically validated in VHL, the of the outcomes is uncertain for VHL population. As this was a cross-sectional study, the information captured from the patient surveys was at a single point in time and from a select group of patients who responded to the survey. It may not be representative of all patients with VHL, and it is likely that HRQoL could vary over time.

Pharmacoeconomic Review

Executive Summary

The executive summary comprises 2 tables (Table 1 and Table 2) and a conclusion.

Table 1: Submitted for Review

CNS = central nervous system; NOC = Notice of Compliance; pNET = pancreatic neuroendocrine tumour; RCC = renal cell carcinoma; VHL = von Hippel-Lindau.

Table 2: Summary of Economic Evaluation

CNS = central nervous system; ICER = incremental cost-effectiveness ratio; NOC = Notice of Compliance; pNET = pancreatic neuroendocrine tumour; QALY = quality-adjusted life-year; RCC = renal cell carcinoma; RDI = relative dose intensity; TTS = time to surgery; ToT = time on treatment; VHL = von Hippel-Lindau.

Conclusions

Evidence from the LITESPARK-004 trial and feedback from clinical experts suggest that, in adult patients with a von Hippel-Lindau (VHL) disease–associated nonmetastatic renal cell carcinoma (RCC) with or without a VHL disease–associated central nervous system (CNS) hemangioblastoma and/or nonmetastatic pancreatic neuroendocrine tumour (pNET), none of whom required immediate surgery, belzutifan demonstrated a promising and durable tumour response as measured by the objective response rate (ORR) and duration of response, which would likely not be otherwise attributable to active surveillance. However, the CADTH clinical review identified uncertainty with the LITESPARK-004 trial findings for other key supportive outcomes (e.g., time to surgery [TTS] and progression-free survival [PFS]) due to major limitations, such as the lack of an internal comparator and no inferential statistical analyses. The sponsor’s indirect treatment comparison (ITC) suggested that treatment with belzutifan may have a beneficial effect on time to RCC-related surgery compared to active surveillance in the RCC cohort; however, due to several key methodological limitations that bias the findings in favour of belzutifan, the comparative effectiveness estimates are uncertain. Given the lack of ITC evidence in the CNS hemangioblastoma and pNET cohorts, the comparative effect of belzutifan in these populations is unknown. Due to the absence of robust clinical evidence, it is unclear whether the modelled survival and quality-of-life benefits associated with belzutifan would be observed in clinical practice.

The CADTH base-case analysis resulted in an incremental cost-effectiveness ratio (ICER) of $360,193 per quality-adjusted life-year (QALY) gained (incremental costs = $975,998; incremental QALYs = 2.71) for belzutifan versus active surveillance for the VHL-RCC cohort. The CADTH base-case analysis results were similar to the sponsors’ results, where belzutifan was not considered cost-effective relative to active surveillance at a willingness-to-pay (WTP) threshold of $50,000 per QALY. A price reduction of approximately 83% is required for belzutifan to be considered cost-effective at a WTP threshold of $50,000 per QALY for the broader RCC population. Based on a price reduction of this magnitude, the cost per 28 days of treatment would be approximately $3,050 per patient. Price reductions for the CNS hemangioblastoma and pNET subgroups were similar.

The cost-effectiveness of belzutifan was sensitive to assumptions surrounding time on treatment (ToT) and residual benefit after treatment discontinuation. Specifically, the CADTH base-case analysis found no residual benefit after belzutifan treatment discontinuation beyond the maximum follow-up period of the LITESPARK-004 trial, which resulted in a smaller difference in QALYs between belzutifan and active surveillance (the sponsor reported a gain of 3.19 QALYs with belzutifan in the RCC cohort while CADTH estimated 2.71 QALYs). The majority of the incremental QALYs associated with belzutifan was driven by an assumed survival benefit. Based on the deterministic results of the CADTH base-case analysis, approximately 90% of the incremental QALYs for belzutifan accrued after the observed median follow-up period in the LITESPARK-004 trial. This benefit should be interpreted with caution given the lack of long-term survival data to confirm the modelled benefit. Uncertainty remains regarding the long-term treatment effect of belzutifan, and the relative efficacy of belzutifan compared to active surveillance in the modelled population. Furthermore, the treatment effect of belzutifan for patients with a VHL-associated CNS hemangioblastoma or pNET who do not have RCC is unknown, as the LITESPARK-004 trial population was narrower than the Health Canada–indicated population. Should RCC have no impact, the subgroup analyses conducted may be generalizable to patients with only VHL-associated CNS or pNET. As CADTH was unable to fully assess the uncertainty around long-term treatment efficacy, comparative efficacy, and missing population due to lack of data, a greater price reduction may be warranted.

Stakeholder Input Relevant to the Economic Review

This section is a summary of the feedback received from the patient groups, registered clinicians, and drug plans that participated in the CADTH review process.

Patient input was received from the Canadian VHL Alliance, which coordinated feedback with the Canadian Organization for Rare Disorders, Kidney Cancer Canada, Pancreatic Cancer Canada, and the Canada Neuroendocrine Tumour Society. Patient input was gathered from 123 patients and caregivers via an online survey. Patients reported that living with VHL disease is associated with ongoing physical and psychological struggles that can have a significant impact on their quality of life. When asked about the number of surgeries experienced, respondents reported having gone through many surgeries. Patients expressed a need for a treatment that would decrease or stabilize the size of tumours, reduce pain, improve breathing, offer longer-term stability or reduce disease burden, and allow them to avoid VHL-related surgery. Input was gathered from 19 patients who had experience with belzutifan for the treatment of VHL disease. These patients noted that belzutifan was moderately to extremely effective in controlling their VHL disease, with side effects being moderately tolerable to very tolerable; however, respondents acknowledged that the current evidence that belzutifan can shrink or eliminate tumours is limited, and suggested all patients receiving belzutifan should be enrolled in a patient-monitoring program to help develop treatment guidelines and protocols. The patient input noted that the use of belzutifan does not require companion diagnostic testing, although respondents pointed out that diagnostic testing is associated with a significant psychological burden, and expressed concerns regarding travel and testing costs, particularly when tests are not covered by the public health system funders. Overall feedback from patient input expressed interest in belzutifan for the treatment of patients with VHL-associated tumours.

Clinician input was received from Canadian subspecialists involved in VHL care and the Ontario Health Cancer Care Oncology Genitourinary Cancer Drug Advisory Committee. Clinicians noted that there is no definitive curative treatment options for VHL disease. Instead, the management of VHL disease involves close clinical and radiographic monitoring and appropriate timing of local strategies, with the aim of reducing the risk of further spread of disease and delaying symptomatic progression. Clinicians noted that belzutifan would fill a large unmet need, cause a shift in disease management, and allow patients to benefit from systemic therapy to address the impact of VHL disease across multiple organs. Clinicians noted that all adult patients (aged 18 years or older) with VHL disease who require treatment for VHL-associated tumours such as an RCC, CNS hemangioblastoma, or pNET without immediate need for surgery should be considered for belzutifan. Clinicians noted that belzutifan treatment should only be received under the guidance of a medical oncologist or other specialist physicians experienced in the use of systemic therapies in a cancer care centre, and that many factors may result in treatment discontinuation, including radiographic disease progression, toxicity and/or unacceptable tolerance, and pregnancy. Last, clinicians indicated that belzutifan may reduce the number of surgical interventions and hospitalizations associated with VHL disease and benefit multiple organs within VHL patients.

Drug plan input identified concerns with drug wastage due to dose interruptions or treatment discontinuation after a supply of belzutifan is dispensed, referencing the number of dose interruptions, reductions, and discontinuations observed in the LITESPARK-004 trial. It also noted that VHL disease usually appears in young adults and can be found in children, and raised the question of whether belzutifan would be considered for use in the pediatric population.

Several of these concerns were addressed in the sponsor’s model:

• Severe treatment related adverse events (AEs) associated with belzutifan, health related quality of life, and surgical and hospitalization costs were included.

In addition, CADTH addressed some of these concerns as follows:

• Diagnostic testing costs were considered in a scenario analysis.

CADTH was unable to address the following concerns raised from stakeholder input:

• Drug wastage was not incorporated into the submitted economic model and could not be addressed by CADTH.

• The use of belzutifan in patients aged less than 18 years was not incorporated into the submitted economic model and could not be addressed by CADTH due to the lack of evidence for its use in this population.

Economic Review

The current review is for belzutifan (Welireg) for the treatment of adult patients with VHL disease who require therapy for an associated nonmetastatic RCC, CNS hemangioblastoma, or nonmetastatic pNET not requiring immediate surgery.

Economic Evaluation

Summary of Sponsor’s Economic Evaluation

Overview

The sponsor submitted a cost-utility analysis (CUA) to assess the cost-effectiveness of belzutifan compared with active surveillance for the treatment of adult patients with VHL disease who require therapy for associated nonmetastatic RCC, CNS, or nonmetastatic pNET not requiring immediate surgery. The model population aligns with the Health Canada indication.¹ The sponsor noted that a subset of patients within each cancer cohort may have 1 or both of the other 2 tumour types, and due to varying rates of surgery that depend on the primary tumour type, it was not possible to present a single ICER combining results from 3 target populations. As a result, the sponsor modelled the 3 subgroup populations separately based on tumour location using similar model structures.²

Belzutifan is available as a 40 mg oral tablet.² The recommended dosage of belzutifan is 120 mg (three 40 mg tablets) administrated orally once daily, with or without food, until unacceptable toxicity or disease progression.³ The submitted price of belzutifan is $213.33 per 40 mg tablet, resulting in a cost of $17,920 per 28 days.² The comparator for this analysis was active surveillance, defined as no systemic treatments, as the sponsor noted there are no currently recommended treatments for this population in Canada.²

The analysis was conducted from the Canadian health care system perspective. Outcomes of the model included QALYs and life-years over a lifetime horizon of 59 years.² Discounting at 1.5% per year was applied to both costs and outcomes.² A weekly cycle length was used with a half-cycle correction applied.²

Model Structure

The sponsor submitted a Markov model consisting of the following mutually exclusive health states: presurgery, metastatic disease, first surgery, event-free after first surgery, metastatic disease after first surgery, second surgery, event-free after second surgery, metastatic disease after second surgery, 3 or more surgeries, event-free after 3 or more surgeries, metastatic disease after 3 or more surgeries, and death (the inter-relationships of these states are depicted in Figure 1).²

All patients entered the model in the presurgery health state (although they may have had an earlier surgery), informed by the baseline visit of LITESPARK-004, and remained in the presurgery state until they underwent their first tumour-reduction surgery for their primary tumour type, developed distant metastases, or died due to any cause.² Patients who transitioned to the first-surgery health state remained in that health state for 1 cycle before either transitioning to the event-free-after-surgery health state or dying.² After undergoing and surviving surgery, patients reside in the event-free-after-surgery state during model cycles in which they are alive, metastases-free, and not requiring immediate surgery for the primary tumour. During this time, patients are at risk of entering subsequent surgery tunnel states (modelled similarly to the first surgery health state), developing metastasis, or dying.² Those in the metastatic-disease health states were noted to encompass RCC and pNET metastases, and patients remained in the metastatic-disease health state until death.² Patient survival within the metastases state was expected to depend on the efficacy of the specific first-line systemic treatment received in the advanced and/or metastatic RCC or pNET setting (depending on the origin tumour).² Each premetastatic health state included 3 different substates defined by patients ORR for their primary tumour to account for the impact of the overall response rate on patient’s health-related quality of life. In the sponsor’s base case, ToT was modelled using on/off treatment health states in which patients could transition from a given on-treatment health state to the corresponding off-treatment state. Clinical efficacy in the off-treatment health state for belzutifan were assumed to converge toward active surveillance over time.

Model Inputs

Patient baseline characteristics in the model were informed by the reported characteristics of the LITESPARK-004 trial, which enrolled adult patients with VHL-associated RCC (N = 61), many of whom also had a CNS hemangioblastoma (82%) and/or pNET (36%). As noted in the overview, the sponsor modelled 3 cohorts of patients separately based on primary tumour type.² As a result, the RCC population includes all patients in the LITESPARK-004 trial, while the CNS hemangioblastoma and pNET populations represent subgroups of the LITESPARK-004 trial and broader RCC population. Clinical data were based on the April 1, 2022, data cut-off from the LITESPARK-004 trial.

The transition from presurgery to first surgery for belzutifan in the VHL-RCC and VHL-CNS hemangioblastoma cohorts were modelled using the exponential distribution fitted to time from baseline visit until the first postbaseline renal or CNS surgery from LITESPARK-004, respectively.² Data informing the VHL-CNS hemangioblastoma population was based on a subset of LITESPARK-004 patients who had CNS hemangioblastoma tumours at baseline.² The transition rate from presurgery to first surgery for belzutifan for the VHL-pNET cohort was set to equal the hazard rate (HR) of presurgery to first surgery with belzutifan versus active surveillance adjusted by (1 − ORR_pNET)/(1 − ORR_RCC) due to the limited number of events that occurred in patients with pNET.² The sponsor assumed that the HR of presurgery to metastatic disease with belzutifan (versus active surveillance) was equal to the HR of presurgery to first surgery in each target population.² In all cohorts, the sponsor assumed that belzutifan would only reduce the risk of deaths attributable to CNS hemangioblastoma progression.² In the sponsor’s base-case analysis, the proportion of patients with deaths attributable to CNS hemangioblastoma (by tumour type) was informed by the VHL Natural History Study, in which each cohort was reweighted to match key baseline characteristics of the corresponding LITESPARK-004 trial population.² Alternative inputs informed by the literature were explored in a scenario analysis.^4,5

All transition rates from the presurgery health state for the VHL-RCC population on active surveillance were informed by the exponential function fitted to the observed data from VHL Natural History Study cohort, reweighted to match the LITESPARK-004 population.² Due to the inability to identify CNS hemangioblastoma and pNET patients at the patient-level index date from the VHL Natural History Study, the pretreatment period of the LITESPARK-004 trial was used to inform the presurgery-to-first-surgery transitions for patients with VHL-CNS hemangioblastoma and VHL-pNET on active surveillance.² Transitions from presurgery to metastatic disease or death for the CNS hemangioblastoma and pNET cohorts were informed using an approach analogous to that used for the VHL-RCC cohort on active surveillance. Adjustment factors were applied to HRs of surgery and metastatic disease to account for real-world standard of care.²

Transitions from any surgery tunnel state (i.e., first, second, or third or subsequent surgeries) to death were assumed to be equal between the belzutifan and active surveillance treatment arms with the perioperative mortality risk per surgery informed by published literature.^2,4,6,7 The sponsor assumed that the mortality risk for all surgery tunnel states would remain the same.²

The probability of transitioning from an event-free state after surgery to the next surgery, or from event-free after surgery to metastatic disease in the VHL-RCC cohort on belzutifan was estimated by multiplying the ratio of the HR of presurgery to first surgery for belzutifan versus active surveillance to the corresponding HR of event-free after first surgery to next surgery or metastatic disease for active surveillance.² For the VHL-CNS hemangioblastoma and VHL-pNET populations on belzutifan, it was assumed that the rates to next surgery and metastatic disease were the same as the corresponding transitions from presurgery to first-surgery and metastatic-disease health states.² The transition probabilities from event-free after first surgery to death for belzutifan were assumed to be the same for each cohort.²

For patients on active surveillance, the transition rates from event-free after surgery to next surgery, metastatic disease, or death in the VHL-RCC population were informed by reweighted patient-level time-to-event data from the VHL Natural History Study.² In the sponsor’s base-case analysis, exponential models were used to inform each transition.² For the VHL-CNS hemangioblastoma and VHL-pNET cohorts, cause-specific hazards of the event-free-after-surgery to next-surgery health states were assumed to be the same as presurgery to first surgery for each respective cohort.² The same approach was taken for the cause-specific hazards of event-free after surgery to metastatic disease and event-free after surgery to death.²

The transition from metastatic disease to death were modelled as a function of the distribution of origin tumour (i.e., RCC or pNET; informed by the VHL Natural History Study), the distribution of first-line treatments for metastatic RCC and pNET (informed by market research and clinical experts input), and the efficacy of the included first-line treatments as informed by published literature.^8-10 It was assumed that the transition probabilities from metastatic disease to death were equal between belzutifan and active surveillance.² Deaths from presurgery or event-free after surgery due to all other causes were assumed to be the same for both arms and informed by the general mortality rate from Statistics Canada.¹¹

The sponsor’s base case included costs and disutilities of surgeries for nonprimary tumours (retinal hemangioblastoma, adrenal lesions, endolymphatic sac tumours, and epididymal cystadenomas). Incidence rates were informed by the LITESPARK-004 trial and the VHL Natural History Study for the RCC cohort on belzutifan and active surveillance, respectively.² Incidence rates for the CNS hemangioblastoma and pNET cohorts were based on the subgroup of patients in LITESPARK-004 with CNS hemangioblastoma and pNET at baseline using data collected after treatment initiation or before treatment initiation, for belzutifan and active surveillance, respectively. The observed incidence rates of nonprimary tumour surgeries were adjusted downward to account for the fact that all patients in the CNS hemangioblastoma and pNET subgroups from the LITESPARK-004 trial also had an RCC.²

In the sponsor’s base case, the risk of AEs (those occurring in ≥ 5% of patients [all grades] and in > 0% of patients at grades 3 to 5 among belzutifan-treated patients) and mean durations were informed by the LITESPARK-004 trial (data cut-off date of April 1, 2022).² The sponsor assumed that patients treated with active surveillance had no AEs.² Both short-term and long-term surgical complications associated with all surgeries were incorporated in the model based on third-party claims data from the US.² Patients who developed long-term surgical complications were assumed to have those complications for the remainder of their lifetime.² The base-case analysis assumed that the risks of surgical complications were equal between belzutifan and active surveillance arms.²

Belzutifan ToT was informed by the Weibull parametric distribution fitted to time-to-treatment-discontinuation data from the LITESPARK-004 trial in the sponsor’s base case.² It was assumed that, following discontinuation of belzutifan treatment, the clinical efficacy inputs for belzutifan would converge gradually over a 2.71-year period toward those of active surveillance (calculated as the amount of time required for the largest RCC tumour to grow back to baseline size assuming the pre-treatment linear growth rate) from the end of the trial.²

Health-state utility values in the model were informed by the Quality of Life Disease Burden Study and the KEYNOTE-564 trial.^5,10 For the nonmetastatic states, a health-state utility value was calculated as a weighted average by objective response (complete response = 0.887, partial response = 0.814, stable disease = 0.814, progressive disease = 0.746) with the relative weights informed by the overall response rate results from the LITESPARK-004 trial for belzutifan and patient-reported responses in the Disease Burden Study for active surveillance.² Utility values in the metastatic disease states were informed by the Disease Burden Study (preprogression metastatic disease = 0.615 and postprogression metastatic disease = 0.570).⁵ Disutilities associated with short- and long-term surgical complications were incorporated as a one-time or recurrent decrement following tumour-reduction surgery, respectively.^5,12-19 AE disutilities were applied as one-off events upon model entry, informed by the utility data from KEYNOTE-564.²⁰ In the sponsor’s base case, a caregiver disutility was included and informed by published literature.^21-23

The sponsor’s base case included drug acquisition costs, AE management costs, surgery and surgical complication costs, disease-management costs, and terminal-care costs. Drug acquisition costs were taken from DeltaPA.²⁴ The drug acquisition cost of belzutifan was provided by the sponsor.² Patients on belzutifan were modelled to have a fixed dosage of 120 mg once daily with a mean relative dose intensity (RDI) of 92.5% as observed in the trial, resulting in a 28-day cost of $16,576.² Unit cost per AE event, tumour-reduction surgery, and costs of both short- and long-term surgical complications were obtained from either the Ontario Case Costing Initiative or published literature.^25-32 Unit costs used to inform health state and disease-management costs were obtained from the Ontario Schedule of Benefits, the Québec Directory and Measurement System for Medical Biology Procedures and published literature, and frequency of use was informed by clinical surveillance guidelines and clinical expert feedback.^2,33-36 Terminal-care costs obtained from published literature were applied as one-off costs associated with patients who transition to death.³⁷ All costs were expressed in 2022 Canadian dollars.

Summary of Sponsor’s Economic Evaluation Results

All analyses were run probabilistically (1,000 iterations). Submitted deterministic analyses were aligned with the probability results. The probabilistic findings are presented below.

Base-Case Results

Results were reported separately for the VHL-RCC, VHL-CNS hemangioblastoma, and VHL-pNET cohorts. The sponsor’s probabilistic base-case analysis found that belzutifan was associated with an additional 3.19 QALYs at an additional cost of $887,036 versus active surveillance for the VHL-RCC cohort. The ICER of belzutifan compared to active surveillance was therefore $278,259 per QALY gained. Similarly, belzutifan was associated with 3.10 and 3.61 more QALYs at an additional cost of $871,975 and $879,508, in the VHL-CNS hemangioblastoma and VHL-pNET cohorts, respectively. The ICERs for VHL-CNS hemangioblastoma and VHL-pNET relative to active surveillance were therefore $281,170 and $243,505 per QALY gained, respectively. Based on deterministic results, the majority (approximately 90%) of the incremental QALYs for belzutifan were found to be accrued during the extrapolation period (after the 37.8-month observed median follow-up period from the LITESPARK-004 trial).

Table 3: Summary of Sponsor’s Economic Evaluation Results

CNS = central nervous system; ICER = incremental cost-effectiveness ratio; pNET = pancreatic neuroendocrine tumour; QALY = quality-adjusted life-year; RCC = renal cell carcinoma; VHL = von Hippel-Lindau.

Source: Sponsor’s pharmacoeconomic submission.²

Sensitivity and Scenario Analysis Results

The sponsor conducted several scenario analyses incorporating different time horizons, perspectives, efficacy and transition probabilities associated with presurgery to surgery, exclusion of adjustments for real-world standard of care, different ToT distributions for belzutifan, alternative sources to inform the proportion of patients with deaths attributed to CNS hemangioblastoma, exclusion of RDI, inclusion of vial sharing, and health-state utilities and disutilities. Results from the scenarios indicated that the ICER was most sensitive to changes associated with belzutifan ToT and efficacy: the ICER decreased in all 3 cohorts (VHL-RCC: $65,643 per QALY gained, VHL-CNS hemangioblastoma: $57,861 per QALY gained, and VHL-pNET: $52,989 per QALY gained) when no treatment-effect waning after discontinuation was assumed. The highest ICERs were associated with a scenario that did not adjust the rates of surgeries or metastases to account for suboptimal real-world surveillance practices (VHL-RCC: $358,314 per QALY gained, VHL-CNS hemangioblastoma: $378,469 per QALY gained, and VHL-pNET: $472,506 per QALY gained).

CADTH Appraisal of the Sponsor’s Economic Evaluation

CADTH identified several key limitations to the sponsor’s analysis that have notable implications on the economic analysis:

• The long-term efficacy of belzutifan for patients with VHL disease–associated nonmetastatic RCC, CNS hemangioblastoma, or nonmetastatic pNET is unknown. Treatment efficacy for belzutifan in the sponsor’s base case was primarily informed by parametric extrapolations fitted to patient-level data from the phase II, single-arm LITESPARK-004 trial. TTS was the key clinical input used to inform transitions from the presurgery health state in the economic model, in which 7 patients (11.5%) with RCC, 1 (2.0%) with CNS hemangioblastoma, and none of the patients with pNET had surgery during the follow-up period. However, as noted in the CADTH clinical report, median TTS was not reached for patients with a VHL disease–associated nonmetastatic RCC, CNS hemangioblastoma, or nonmetastatic pNET at the April 1, 2022, data cut-off date. No definitive conclusion could therefore be drawn about the long-term benefits of belzutifan on TTS. Clinical expert feedback received by CADTH noted that, while treatment with belzutifan may have an impact on TTS and the number of surgeries, the magnitude of any impact is uncertain as a result of the identified limitations with the available data. Based on the sponsor’s model structure, the assumptions around TTS resulted in a survival benefit for patients on belzutifan. As described in the CADTH clinical report, investigator-assessed median PFS was not reached for metastatic RCC, median PFS was not reached for VHL-CNS hemangioblastoma or VHL-pNET patients, and overall survival was not captured in the LITESPARK-004 trial (data cut-off of April 1, 2022). Therefore, the assumed survival benefit of belzutifan modelled by the sponsor cannot be validated based on currently available information and is associated with uncertainty.

  CADTH also identified concerns regarding the generalizability of the LITESPARK-004 trial to clinical practice in Canada. As noted in the CADTH clinical report, despite the number of organs that VHL disease can affect, only a small number of patient groups (i.e., RCC only, RCC with CNS hemangioblastoma, RCC with pNET, and RCC with both CNS hemangioblastoma and pNET) were assessed in the LITESPARK-004 trial. The trial population represents a narrower population than the Health Canada–indicated population. While the clinical expert feedback received by CADTH noted that, in general, the study selection criteria were aligned with patients who may receive belzutifan in the Canadian setting, there may be limitations due to the lack of Canadian sites in the studies, and that greater than 90% of patients in the study were white, which does not reflect the racial profile in Canada.

  • CADTH was unable to resolve this issue due to limitations in data availability. The impact of this limitation on the cost-effectiveness of belzutifan is uncertain.

• The relative efficacy of belzutifan compared with active surveillance is uncertain. In the absence of comparative clinical data, the sponsor used patient-level time-to-event data from the VHL Natural History Study (2021) cohort, reweighted to match the baseline characteristics of the LITESPARK-004 trial population, to inform the efficacy of the comparator, active surveillance. However, as noted in the CADTH clinical report, there were limitations in utilization of the VHL Natural History Study relating to the selection criteria intended to match the VHL Natural History Study’s cohort with that from the LITESPARK-004, key covariates were missing from the pivotal trial, and no adjustments were made for important effect-modifying or prognostic factors, such as VHL gene alteration type and the number of measured tumours. These made it difficult to assess the degree of heterogeneity between the populations in the LITESPARK-004 trial and VHL Natural History Study, and the potential for residual confounding was considered high. Furthermore, the sponsor was unable to identify patients who had CNS hemangioblastomas and/or pNET tumours in the VHL Natural History Study cohort. Clinical expert feedback received by CADTH noted that, despite the limitations of the available comparative evidence, belzutifan appears to have a promising tumour response compared with the experts’ experiences with the natural history of the disease. However, the clinical experts indicated that the magnitude of benefit associated with belzutifan is unknown.

  • CADTH was unable to resolve this issue due to limitations in data availability. As a result, the estimate of the ICER for belzutifan compared with active surveillance is associated with uncertainty.

• The ToT for patients receiving belzutifan is uncertain. In the sponsor’s base case, the ToT for patients receiving belzutifan was informed by parametric curves fitted to patient-level data on the time-to-treatment-discontinuation data from LITESPARK-004. Although the Gompertz curve was the best-fitting distribution based on Akaike information criterion and Bayesian information criterion, the sponsor selected the Weibull distribution for its base-case analysis (patients received treatment for up to 14 years), stating that the longer tail was more plausible because a proportion of patients are expected to remain on treatment beyond 6 years (i.e., the point at which the Gompertz curve predicted that all patients would have discontinued). Clinical expert feedback received by CADTH noted that, given the follow-up period for the LITESPARK-004 trial (median = 37.8 months; range = 4.2 to 46.1 months), the sponsor’s ToT assumption for patients receiving belzutifan is associated with uncertainty. However, the clinical experts noted that the ToT estimates resulting from the Weibull distribution would likely overestimate the ToT for patients receiving belzutifan expected to be seen in clinical practice in Canada.

  • CADTH explored the impact of selecting the Gompertz curve to inform time on belzutifan treatment in a scenario analysis.

• The assumption of residual benefit for belzutifan after treatment discontinuation is associated with uncertainty. In the sponsor’s submitted economic model, the ToT for patients receiving belzutifan was modelled based on Weibull distributions fitted to ToT discontinuation data from the LITESPARK-004 trial. The sponsor assumed that, at the time of discontinuing belzutifan, a patient’s largest tumour would be smaller than if the patient had not been treated with belzutifan, based on changes in RCC tumour size from the LITESPARK-004 trial. In the submitted base-case analysis, the sponsor assumed that, upon discontinuation of belzutifan treatment after the maximum trial follow-up period (46.1 months), patients’ probabilities of requiring a surgery, experiencing metastatic disease, or dying (i.e., their transition probabilities from the presurgery and event-free-after-surgery health states) would linearly converge with that of active surveillance over 2.71 years, based on the time it would take for the average size of the largest RCC tumour at belzutifan discontinuation (i.e., 15.36 mm) to grow back to the baseline size (i.e., 24.9 mm), assuming a linear growth rate equal to that of the pretreatment period of the LITESPARK-004 trial (i.e., 3.52 mm per year).

  Clinical expert feedback obtained by CADTH indicated that, although there was no robust evidence to support an attenuation of benefit beyond treatment discontinuation, patients who received belzutifan may achieve some residual benefit after treatment discontinuation. However, feedback from the clinical experts considered the sponsor’s assumption of 2.71 years to be an overestimate. Clinical expert feedback noted that that tumour growth rates after treatment can vary significantly compared to the pretreatment growth rate (i.e., in some disease areas a tumour may have no growth after treatment discontinuation whereas other tumours may grow more quickly and greater in size than previously). Without post-discontinuation efficacy data there is therefore significant uncertainty in how tumours may grow after belzutifan discontinuation.

  The sponsor presented a scenario analysis exploring the effects of no residual benefit. However, this scenario was inappropriate, as it reduced the ToT while maintaining the clinical efficacy of belzutifan throughout the modelled time horizon. As noted previously, there is significant uncertainty in the long term efficacy of belzutifan and therefore this method overestimates the treatment effect of belzutifan.

  • In the CADTH base-case analysis, patients were modelled to converge to active surveillance immediately after belzutifan treatment discontinuation after the maximum trial follow-up period.

  • To explore the impact of some residual benefit post-discontinuation, a scenario analysis was conducted assuming a shorter time period (1 year) to convergence.

• The sponsor underestimated belzutifan drug costs. In the sponsor’s base-case analysis, the mean RDI observed in the LITESPARK-004 trial was used to derive the drug acquisition cost for belzutifan. The inclusion of RDI may underestimate the total cost of belzutifan in clinical practice as the dose received by a patient may be different from the planned dose for several reasons (e.g., missed doses or dose reductions).

  Additionally, the sponsor assumed that the price of belzutifan would experience a 30% price reduction after 10 years due to the introduction of generics. Although generics for belzutifan may become available in the future, it is unknown in what capacity and at what price, and the inclusion of such a price reduction for generics in the cost-effectiveness model may therefore underestimate the total treatment cost of belzutifan to the Canadian health care system.

  • In the CADTH base-case reanalysis, the RDI was set to 100% (following the sponsor’s provided drop down) and price reductions associated with generics were excluded.

• The inclusion of caregiver disutility in the base case was inappropriate. The sponsor noted that a patient’s health status and well-being would have a significant impact on caregivers, and it therefore included caregiver disutility in its base case. The sponsor assumed that each patient would have 1 caregiver, and disutility values were informed by published reports involving family members and caregivers of cancer patients. Although caregivers are often affected by a patient’s health status, the evidence to support the degree of impact for VHL-specific caregivers is weak and the sponsor’s incorporation of caregiver disutility is therefore uncertain. The inclusion of caregiver disutilities overestimates of the benefits associated with belzutifan in the sponsor’s base case as it increased the QALY gains received by patients in the form of caregiver benefits.

  • In the CADTH base-case reanalysis, caregiver disutilities were excluded. Caregiver disutilities were included in a scenario analysis with a societal perspective.

• Poor modelling practices were employed. The sponsor’s submitted model included numerous IFERROR statements, which can lead to situations in which the parameter value is over-written with an alternative value without alerting the user to the automatized overwriting. The systematic use of IFERROR statements makes thorough auditing of the sponsor’s model impractical, and it remains unclear whether the model is running inappropriately by overriding errors.

  • CADTH was unable to address this limitation or conduct a thorough validation of the sponsor’s model.

Table 4: Key Assumptions of the Submitted Economic Evaluation (Not Noted as Limitations to the Submission)

CNS = central nervous system; HR = hazard ratio; ORR = objective response rate; pNET = pancreatic neuroendocrine tumour; RCC = renal cell carcinoma; VHL = von Hippel-Lindau.

CADTH Reanalyses of the Economic Evaluation

Base-Case Results

The CADTH base case was derived by making changes in model parameter values and assumptions, in consultation with clinical experts. These changes, summarized in Table 5, involved removal of belzutifan residual benefit after treatment discontinuation after the trial duration, removal of caregiver disutility, and correcting belzutifan costing inputs.

Table 5: CADTH Revisions to the Submitted Economic Evaluation

The CADTH base-case analysis focuses on the RCC population due to duplicated patients in the CNS and/or pNET cohort. The results of the CADTH base-case analyses demonstrated that belzutifan was associated with 2.71 additional QALYs at an additional cost of $975,998 versus active surveillance for the VHL-RCC population. Therefore, the ICER of belzutifan compared to active surveillance was $360,193 per QALY gained. The probability of cost-effectiveness at a $50,000 per QALY WTP threshold was 0%. Based on the deterministic results of the CADTH base case, approximately 90% of the incremental QALYs for belzutifan were accrued during the extrapolated period (i.e., after the 37.8-month observed median follow-up period from the LITESPARK-004 trial).

CADTH conducted further subgroup analyses for the CNS hemangioblastoma and pNET populations. The results of these analyses indicated that, compared to active surveillance, belzutifan was associated with 2.61 and 3.08 more QALYs at additional costs of $961,796 and $968,488, respectively. The ICERs for VHL-CNS hemangioblastoma and VHL-pNET subgroups relative to active surveillance were therefore $368,886 per QALY gained and $314,661 per QALY gained, respectively. A summary of the CADTH base-case and subgroup-reanalysis results can be found in Table 6. Details on the stepped analysis are available in Appendix 4.

Table 6: Summary of the Stepped Analysis of the CADTH Reanalysis Results

CNS = central nervous system; ICER = incremental cost-effectiveness ratio; pNET = pancreatic neuroendocrine tumour; QALY = quality-adjusted life-year; RCC = renal cell carcinoma; VHL = von Hippel-Lindau.

Note: Results were derived probabilistically.

Scenario Analysis Results

Three scenarios were conducted on the CADTH base-case analysis for the RCC population and key subgroups to investigate the impact of using the Gompertz distribution to inform belzutifan ToT, assuming convergence to active surveillance efficacy would occur gradually over 1 year after the LITESPARK-004 trial duration and including genetic testing costs. Results of scenario analyses are presented in Appendix 4.

The analysis examining the impact of using the Gompertz distribution to inform belzutifan ToT resulted in ICERs of $349,738 per QALY gained, $361,534 per QALY gained, and $304,739 per QALY gained for belzutifan compared with active surveillance for the VHL-RCC, VHL-CNS hemangioblastoma, and VHL-pNET cohorts, respectively. The analyses assuming discontinued patients would gradually converge to active surveillance efficacy over a 1-year period after the LITESPARK-004 trial duration resulted in ICERs of $344,302 per QALY gained, $352,723 per QALY gained, and $300,492 per QALY gained for belzutifan compared with active surveillance for the VHL-RCC, VHL-CNS hemangioblastoma, and VHL-pNET cohorts, respectively. Last, inclusion of genetic testing costs resulted in ICERs of $360,260 per QALY gained, $368,956 per QALY gained, and $314,720 per QALY gained for belzutifan versus active surveillance for the VHL-RCC, VHL-CNS hemangioblastoma, and VHL-pNET cohorts, respectively.

CADTH also undertook price-reduction analyses based on the sponsor’s and CADTH’s base cases. The CADTH base case suggested a price reduction of approximately 83% would be required to achieve cost-effectiveness of belzutifan at a WTP threshold of $50,000 per QALY gained in the full VHL-RCC cohort. The price reduction in the VHL-CNS hemangioblastoma, and VHL-pNET cohorts were similar (82% and 81%, respectively).

Table 7: CADTH Price-Reduction Analyses

CNS = central nervous system; ICER = incremental cost-effectiveness ratio; pNET = pancreatic neuroendocrine tumour; QALY = quality-adjusted life-year; RCC = renal cell carcinoma; VHL = von Hippel-Lindau.

Note: All price-reduction analyses were conducted deterministically.

Issues for Consideration

• Clinical expert feedback obtained by CADTH noted that the availability of belzutifan for adult patients with VHL disease who require therapy for associated nonmetastatic RCC, CNS hemangioblastoma, or nonmetastatic pNET not requiring immediate surgery is unlikely to notably increase the utilization of health care resources such as germline genetic testing as patients are expected to have already undergone genetic testing due to a family history or earlier presentation with VHL-associated morbidities, and it is not anticipated that testing will be required to confirm diagnosis for eligible patients. However, should testing frequency increase with belzutifan reimbursement, the true impact on health care resource use remains unknown. Based on a CADTH scenario analysis, the inclusion of genetic tests had minimal impact on the ICERs in all 3 of the cohorts.

• Despite VHL being a multiorgan disease, only a small number of patient groups (i.e., RCC only, RCC with CNS hemangioblastoma, RCC with pNET, and RCC with both CNS hemangioblastoma and pNET) were assessed in the LITESPARK-004 trial. The trial population is therefore narrower than the Health Canada–indicated population. While the clinical expert feedback received by CADTH noted that results in the explored patient groups are likely reflective of other patient groups, the cost-effectiveness of belzutifan for patients outside of those included in the LITESPARK-004 trial is unknown.

• Drug plan input obtained by CADTH considered the potential use of belzutifan in a pediatric population. At present, Health Canada has not authorized an indication for pediatric use due to the lack of data from this population. However, if a pediatric indication is authorized in the future, the cost-effectiveness and budgetary impact of using belzutifan in this population is unknown.

Overall Conclusions

Evidence from the LITESPARK-004 trial and feedback from clinical experts, suggests that, in adult patients with a VHL disease–associated nonmetastatic RCC, VHL disease–associated CNS hemangioblastoma, and VHL disease–associated nonmetastatic pNET, none of whom required immediate surgery, belzutifan demonstrated a promising and durable tumour response as measured by ORR and duration of response, which would likely not be otherwise attributable to active surveillance. However, uncertainty is associated with the LITESPARK-004 trial findings for other key supportive outcomes (e.g., TTS and PFS) due to major limitations, such as the lack of an internal comparator and no inferential statistical analyses. The sponsor’s ITC suggested that treatment with belzutifan may have a beneficial effect on time to RCC-related surgery compared to active surveillance in the RCC cohort; however, due to several key methodological limitations that bias the findings in favour of belzutifan, the comparative effectiveness estimates are uncertain. Given the lack of ITC evidence in the CNS hemangioblastoma and pNET cohorts, the comparative effects of belzutifan in these populations are unknown. Due to the absence of robust clinical evidence, it is unclear whether the modelled quality-of-life benefits associated with belzutifan would be observed in clinical practice.

CADTH identified several limitations in the sponsor’s pharmacoeconomic analysis that have notable implications on the cost-effectiveness of belzutifan. As part of the base-case analysis, CADTH revised the assumptions regarding residual benefit with belzutifan upon treatment discontinuation, removed caregiver disutilities, and removed belzutifan trial RDI and generic costing assumptions. The CADTH base-case reanalysis resulted in an ICER of $360,193 per QALY gained (incremental costs = $975,998; incremental QALYs = 2.71) for belzutifan versus active surveillance for the VHL-RCC cohort. In the VHL-CNS hemangioblastoma and VHL-pNET subgroup cohorts, the ICERs for belzutifan versus active surveillance were $368,886 per QALY gained (incremental costs = $961,796; incremental QALYs = 2.61) and $314,661 per QALY gained (incremental costs = $968,488; incremental QALYs = 3.08), respectively. The CADTH base-case results were aligned with the sponsor’s results, with belzutifan associated with a higher ICER and not considered cost-effective relative to active surveillance at a $50,000 WTP threshold. A price reduction of approximately 83% is therefore required for belzutifan to be considered cost-effectiveness at a WTP threshold of $50,000 per QALY for the broader RCC population. Based on a price reduction of this magnitude, the cost per 28 days of treatment would be approximately $3,050 per patient. Required price reductions for the CNS hemangioblastoma and pNET subgroups were similar.

The cost-effectiveness of belzutifan was sensitive to assumptions surrounding ToT and residual benefit after treatment discontinuation. Specifically, the CADTH base-case analysis considered no residual benefit after belzutifan treatment discontinuation after the maximum follow-up period of the LITESPARK-004 trial that resulted in a smaller difference in QALYs between belzutifan and active surveillance (gains in QALYs with belzutifan were 3.19 as reported by the sponsor versus 2.71 as estimated by CADTH in the RCC cohort). Although clinical expert feedback noted that belzutifan reimbursement is not expected to affect the current genetic testing landscape, CADTH undertook a scenario exploring an increase in genetic testing due to belzutifan reimbursement; the scenario analysis suggested a minimal impact on the ICER in all 3 cohorts.

The long-term treatment effect of belzutifan for the treatment of adult patients with a VHL disease–associated nonmetastatic RCC, VHL disease–associated CNS hemangioblastoma, and VHL disease–associated nonmetastatic pNET, none of whom required immediate surgery, as well as the comparative efficacy of belzutifan to active surveillance, remain uncertain. Based on the deterministic results of the CADTH base case, approximately 90% of the incremental QALYs for belzutifan accrued after the 37.8-month observed median follow-up period from the LITESPARK-004 trial. This benefit should be interpreted with caution given the lack of overall survival data to confirm this modelled benefit. Additionally, as the LITESPARK-004 trial population was narrower than the Health Canada–indicated population, there is uncertainty in the treatment effect of belzutifan for patients excluded from the analysis (i.e., patients with a VHL-associated CNS hemangioblastoma or pNET who do not have RCC). Should RCC have no effect on other tumour types, the subgroup analyses conducted may be generalized to patients with only a VHL-associated CNS or pNET. As CADTH was unable to fully assess the uncertainty around long-term treatment efficacy and comparative efficacy to active surveillance due to lack of data, a higher price reduction may be warranted.

References

1.Merck Canada response to February 17, 2023 CADTH request for additional information regarding Welireg (belzutifan) CADTH review [internal additional sponsor's information]. Kirkland (QC) Merck Canada Inc; 2023 Feb 24.

2.Pharmacoeconomic evaluation [internal sponsor's report]. In: Drug Reimbursement Review sponsor submission: Welireg (belzutifan). Kirkland (QC): Merck Canada Inc; 2023 May 1.

3.Welireg® belzutifan tablets, 40 mg, oral [product monograph]. Kirkland (QC) Merck Canada Inc; 2022 Dec 13.

4.Lonser RR, Weil RJ, Wanebo JE, DeVroom HL, Oldfield EH. Surgical management of spinal cord hemangioblastomas in patients with von Hippel-Lindau disease. J Neurosurg. 2003;98(1):106-116. PubMed

5.Adelphi Real World. VHL-Related Disease Impact on Quality of Life and Work Productivity in Patients with Renal Cell Carcinoma, Central Nervous System Hemangioblastoma, or Pancreatic Neuroendocrine Tumor Manifestation. 2022.

6.Krauss T, Ferrara AM, Links TP, et al. Preventive medicine of von Hippel-Lindau disease-associated pancreatic neuroendocrine tumors. Endocr Relat Cancer. 2018;25(9):783-793. PubMed

7.Johnson A, Sudarshan S, Liu J, Linehan WM, Pinto PA, Bratslavsky G. Feasibility and outcomes of repeat partial nephrectomy. J Urol. 2008;180(1):89-93; discussion 93. PubMed

8.Motzer R, Alekseev B, Rha SY, et al. Lenvatinib plus Pembrolizumab or Everolimus for Advanced Renal Cell Carcinoma. N Engl J Med. 2021;384(14):1289-1300. PubMed

9.Riaz IB, He H, Ryu AJ, et al. A Living, Interactive Systematic Review and Network Meta-analysis of First-line Treatment of Metastatic Renal Cell Carcinoma. Eur Urol. 2021;80(6):712-723. PubMed

10.Rini BI PE, Stus V, et al. Pembrolizumab plus axitinib versus sunitinib as first-line therapy for advanced clear cell renal cell carcinoma: Results from 42-month follow-up of KEYNOTE-426. Poster presented at the 2021 American Society of Clinical Oncology (ASCO) Annual Meeting.

11.Table: 17-10-0057-01 Projected population, by projection scenario, age and sex, as of July 1 (x 1,000). Ottawa (ON: Statistics Canada; 2022: https://www150.statcan.gc.ca/t1/tbl1/en/tv.action?pid=1710005701. Accessed 1800 Jan 1.

12.Lee AJ, Morgan CL, Conway P, Currie CJ. Characterisation and comparison of health-related quality of life for patients with renal failure. Curr Med Res Opin. 2005;21(11):1777-1783. PubMed

13.Simianu VV, Gaertner WB, Kuntz K, et al. Cost-effectiveness Evaluation of Laparoscopic Versus Robotic Minimally Invasive Colectomy. Ann Surg. 2020;272(2):334-341. PubMed

14.Gandhi SK, Jensen MM, Fox KM, Smolen L, Olsson AG, Paulsson T. Cost-effectiveness of rosuvastatin in comparison with generic atorvastatin and simvastatin in a Swedish population at high risk of cardiovascular events. Clinicoecon Outcomes Res. 2012;4:1-11. PubMed

15.Zargar M, McFarlane T, Chan KKW, Wong WWL. Cost-Effectiveness of Nivolumab in Recurrent Metastatic Head and Neck Squamous Cell Carcinoma. Oncologist. 2018;23(2):225-233. PubMed

16.Poole CD, Chambers C, Allsopp R, Currie CJ. Quality of life and health-related utility analysis of adults with moderate and severe atopic dermatitis treated with tacrolimus ointment vs. topical corticosteroids. J Eur Acad Dermatol Venereol. 2010;24(6):674-678. PubMed

17.Stevenson SM, Danzig MR, Ghandour RA, et al. Cost-effectiveness of neoadjuvant chemotherapy before radical cystectomy for muscle-invasive bladder cancer. Urol Oncol. 2014;32(8):1172-1177. PubMed

18.Wang CY, Pham PN, Kim S, et al. Predicting Cost-Effectiveness of Generic vs. Brand Dabigatran Using Pharmacometric Estimates Among Patients with Atrial Fibrillation in the United States. Clin Transl Sci. 2020;13(2):352-361. PubMed

19.Verkleij ML, Heijnsdijk EAM, Busse AML, et al. Cost-Effectiveness of Neonatal Hearing Screening Programs: A Micro-Simulation Modeling Analysis. Ear Hear. 2021;42(4):909-916. PubMed

20.National Institute for Health and Care Excellence. Pembrolizumab for adjuvant treatment of renal cell carcinoma [TA830]. 2022; https://www.nice.org.uk/guidance/ta830/documents/committee-papers. Accessed Feb 24, 2022.

21.Turner D, Adams E, Boulton M, et al. Partners and close family members of long-term cancer survivors: health status, psychosocial well-being and unmet supportive care needs. Psychooncology. 2013;22(1):12-19. PubMed

22.Sjolander C, Rolander B, Järhult J, Mårtensson J, Ahlstrom G. Health-related quality of life in family members of patients with an advanced cancer diagnosis: a one-year prospective study. Health Qual Life Outcomes. 2012;10:89. PubMed

23.Song JI, Shin DW, Choi JY, et al. Quality of life and mental health in the bereaved family members of patients with terminal cancer. Psychooncology. 2012;21(11):1158-1166. PubMed

24.Government of A. Interactive drug benefit list. 2022; https://idbl.ab.bluecross.ca/idbl/load.do. Accessed 1800 Jan 1.

25.Ontario Case Costing Initiative (OCCI). Toronto (ON): Ontario Health and Long-Term Care; 2017: https://data.ontario.ca/dataset/ontario-case-costing-initiative-occi. Accessed 1800 Jan 1.

26.Zelmer JL. The economic burden of end-stage renal disease in Canada. Kidney Int. 2007;72(9):1122-1129. PubMed

27.Manns B, Hemmelgarn B, Tonelli M, et al. The Cost of Care for People With Chronic Kidney Disease. Can J Kidney Health Dis. 2019;6:2054358119835521. PubMed

28.Spencer Netto F, Quereshy F, Camilotti BG, et al. Hospital costs associated with laparoscopic and open inguinal herniorrhaphy. JSLS. 2014;18(4). PubMed

29.Guerriere DN, Choiniere M, Dion D, et al. The Canadian STOP-PAIN project - Part 2: What is the cost of pain for patients on waitlists of multidisciplinary pain treatment facilities? Can J Anaesth. 2010;57(6):549-558. PubMed

30.Widjaja E, Guttmann A, Tomlinson G, Snead OC, 3rd, Sander B. Economic burden of epilepsy in children: A population-based matched cohort study in Canada. Epilepsia. 2021;62(1):152-162. PubMed

31.Rosella LC, Lebenbaum M, Fitzpatrick T, et al. Impact of diabetes on healthcare costs in a population-based cohort: a cost analysis. Diabet Med. 2016;33(3):395-403. PubMed

32.Bornstein SR, Allolio B, Arlt W, et al. Diagnosis and Treatment of Primary Adrenal Insufficiency: An Endocrine Society Clinical Practice Guideline. J Clin Endocrinol Metab. 2016;101(2):364-389. PubMed

33.Quebec Directory and Measurement System for Medical Biology Procedures. 2021-2022. https://publications.msss.gouv.qc.ca/msss/fichiers/2020/20-922-05W.pdf.

34.Schedule of benefits for physician services under the Health Insurance Act: (January 25, 2022 (effective July 1, 2022)). Toronto (ON): Ontario Ministry of Health; 2022: https://www.health.gov.on.ca/en/pro/programs/ohip/sob/physserv/sob_master.pdf. Accessed 1800 Jan 1.

35.Stellato D, Thabane ME, Chandiwana D, Park J, Delea TE. Cost Effectiveness of Ribociclib Plus a Nonsteroidal Aromatase Inhibitor in Pre-/Perimenopausal, HR+ and HER2- Advanced Breast Cancer: A Canadian Healthcare Perspective. Pharmacoeconomics. 2021;39(7):853-867. PubMed

36.Thavorn K, Wang Z, Fergusson D, van Katwyk S, Arnaout A, Clemons M. Cost implications of unwarranted imaging for distant metastasis in women with early-stage breast cancer in Ontario. Curr Oncol. 2016;23(Suppl 1):S52-55. PubMed

37.Rocchi A, Verma S. Anastrozole is cost-effective vs tamoxifen as initial adjuvant therapy in early breast cancer: Canadian perspectives on the ATAC completed-treatment analysis. Support Care Cancer. 2006;14(9):917-927. PubMed

38.Budget Impact Analysis [internal sponsor's report]. In: Drug Reimbursement Review sponsor submission: Welireg (belzutifan). Kirkland (QC): Merck Canada Inc; 2023 Jan 23.

39.Orphanet. Von Hippel-Lindau Disease. https://www.orpha.net/consor/cgi-bin/Disease_Search.php?lng=EN&data_id=99&Disease_Disease_Search_diseaseGroup=von-hippel-lindau&Disease_Disease_Search_diseaseType=Pat&Maladie(s)/groupes%20de%20maladies=Maladie-de-von-Hippel-Lindau&title=Maladie%20de%20von%20Hippel-Lindau&search=Disease_Search_Simple. Accessed Feb 21 2023.

40.Qalcul. Qalcul Canadian drug coverage estimator. https://qalcul.com/en, 2023.

41.Understanding the gap: A pan-Canadian analysis of prescription drug insurance coverage. The Conference Board of Canada; 2018.

42.Alignment among public formularies in Canada, Part 2: Oncology medicines. Government of Canada; 2021: https://www.canada.ca/en/patented-medicine-prices-review/services/npduis/analytical-studies/formularies-part2-oncology-medicines.html

43.Phillips K. Catastrophic drug coverage in Canada. Ottawa: Library of Parliament; 2016: https://lop.parl.ca/staticfiles/PublicWebsite/Home/ResearchPublications/BackgroundPapers/PDF/2016-10-e.pdf. Accessed April 2023.

Appendix 1: Cost-Comparison Table

Note that this appendix has not been copy-edited.

The comparators presented in the following table have been deemed to be appropriate based on feedback from clinical experts and CADTH-participating public drug plans. Comparators may be recommended (appropriate) practice or actual practice. Existing Product Listing Agreements are not reflected in the table and as such, the table may not represent the actual costs to public drug plans.

Table 8: CADTH Cost-Comparison Table for the Treatment of Patients with von Hippel-Lindau Disease–Associated Nonmetastatic Renal Cell Carcinoma, Central Nervous System Hemangioblastoma, or Nonmetastatic Pancreatic Neuroendocrine Tumour Not Requiring Immediate Surgery

^aRecommended dosage is from the product monograph.³

^bSponsor-submitted price.³⁸

Appendix 2: Submission Quality

Note that this appendix has not been copy-edited.

Table 9: Submission Quality

Appendix 3: Additional Information on the Submitted Economic Evaluation

Note that this appendix has not been copy-edited.

Figure 1: Model Structure

Source: Sponsor’s Pharmacoeconomic Evaluation²

Detailed Results of the Sponsor’s Base Case

Table 10: Disaggregated Summary of Sponsor’s Base-Case Economic Evaluation Results