CADTH Reimbursement Review

Nivolumab (Opdivo)

Sponsor: Bristol Myers Squibb

Therapeutic area: Non–small cell lung cancer

This multi-part report includes:

Clinical Review

Pharmacoeconomic 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: Submitted for Review

NSCLC = non–small cell lung cancer; NOC = Notice of Compliance.

Introduction

Lung cancer is the most frequently diagnosed cancer in Canada, and the leading cause of cancer-related death.¹ About 30,000 new diagnoses (50% in males and 50% in females)² and 20,700 cancer-related deaths were projected in Canada in 2022,^1,3 with about 98% of all cases anticipated in people 50 years and older.⁴ The adjusted 5-year net survival (based on 2015 to 2017 estimates) for all forms of lung cancer is only 22% (19% in males and 26% in females in Canada).^1,2 Non–small cell lung cancers (NSCLCs) are the most common forms of lung cancer, accounting for more than 80% of all lung cancers in Canada.^5,6 About 47.1% of all new cases of NSCLC are diagnosed at stage IV, 19.0% are diagnosed at stage III, 9.1% are diagnosed at stage II, and only 23.1% are diagnosed at stage I.⁷

The primary goal of treating patients with resectable NSCLC is to cure the disease, improve 5-year overall survival (OS), and prevent disease recurrence. Surgery with curative intent is the current gold standard for clinical stage I to stage IIIA NSCLC that is amenable to resection.^8,9 The standard of care, according to joint guidelines from the American Society of Clinical Oncology and Ontario Health (OH) Cancer Care Ontario (CCO) for completely resectable stage IIA to IIB and stage IIIA NSCLC (American Joint Committee on Cancer [AJCC] Cancer Staging Manual, 7th edition [AJCC 7th edition]), is surgical resection followed by adjuvant cisplatin-based chemotherapy. Cisplatin-based chemotherapy is not recommended for patients with stage IA disease and not routinely recommended for patients with stage IB disease; however, postoperative evaluations are recommended.¹⁰ Stereotactic ablative radiation with curative intent is available to some patients with early-stage disease who are ineligible for surgery (e.g., because of significant comorbidities that make them a high risk for general anesthetic) or who refuse surgery, whereas patients with resectable stage III cancer may be offered chemotherapy and/or radiation in current practice before surgery.¹¹

Neoadjuvant chemotherapy is seldom used in Canada, as it has not been shown to provide survival benefit over adjuvant therapy, and in the process of pursuing neoadjuvant chemotherapy, some patients may become ineligible for surgery (due to disease progression or treatment-related toxicity), according to the clinical experts consulted. However, neoadjuvant therapy has several advantages; for instance, it can reduce tumour size, increase resectability, and remove micrometastasis and tumour cells in more distant lymph nodes, which reduces the risk of recurrence caused by tumour cells that are not removed by surgery.^12,13

Nivolumab (Opdivo) is a human immunoglobulin G4 monoclonal antibody that binds to the programmed cell death 1 protein (PD-1) receptor and blocks its interaction with programmed cell death 1 ligand 1 (PD-L1) and ligand 2 (PD-L2), releasing PD-1 pathway-mediated inhibition of the immune response, including antitumour immune response.¹⁴ Nivolumab underwent review through the Project Orbis route at Health Canada and received a notice of compliance (NOC) on August 18, 2022, for the treatment of adults with resectable NSCLC (tumours ≥ 4 cm or node-positive disease) when used in combination with platinum-doublet chemotherapy.¹² The Health Canada–recommended dose of nivolumab is 360 mg administered intravenously over 30 minutes as neoadjuvant treatment in combination with platinum-doublet chemotherapy every 3 weeks for 3 cycles.¹⁴ The sponsor’s reimbursement request and patient population aligns with the Health Canada indication.

The objective of this CADTH report is to perform a systematic review of the beneficial and harmful effects of nivolumab (360 mg) in combination with platinum-doublet chemotherapy every 3 weeks for up to 3 cycles, for the neoadjuvant treatment of adults with resectable NSCLC (tumours ≥ 4 cm or node-positive disease).

Stakeholder Perspectives

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

Patient Input

Lung Cancer Canada (LCC), a member of Global Lung Cancer Coalition and the only national organization in Canada focused exclusively on lung cancer, engages in patient support, education, research, and advocacy. LCC submitted patient group input based on interviews conducted in Canada in September and October 2022 (1 patient with stage I to II NSCLC, 1 patient with stage III NSCLC, and 2 patients with stage IV NSCLC, and 1 caregiver for a patient with large-cell neuroendocrine carcinoma) and 1 interview with a patient who had stage IV NSCLC and had undergone an environmental scan (1 patient with stage IV NSCLC). All participants had experience with nivolumab. The LCC submitted information collected from 4 patients with stage I to IV NSCLC and from 1 patient with large-cell neuroendocrine carcinoma. The CADTH patient input summary will focus on the 2 patients with stage I to stage III NSCLC to align with the requested indication.

One patient said she used to be extremely active and was an avid runner for 10 years before her lung cancer diagnosis. The patient explained that lung cancer made exercise harder and made her feel more tired than ever, which impacted her independence. She also said that she experienced cough and some mild chest pain before her diagnosis of lung cancer. Another patient who had been diagnosed with early-stage NSCLC had two-thirds of his lung surgically removed, leaving him with 50% of his initial lung capacity. As a result, he explained, he became unable to do any vigorous exercise or activity because he tires quickly. As for improved outcomes, the input indicated that patients would value new treatment options that maintain or improve quality of life, delay the onset of symptoms, improve survivorship, improve or maintain functionality and/or mobility, and, ultimately, provide a cure. In addition, respondents said they would prefer a treatment that can be administered at a hospital located near home, or in a community clinic for those in rural settings, to minimize travel time and burden on caregivers. The input also noted that because the CheckMate 816 trial excluded patients with EGFR or ALK alterations, owing to a lack of evidence supporting neoadjuvant immunotherapy in this population, a biomarker screening (a routine practice) must be performed before neoadjuvant treatment with nivolumab.

Clinician Input

Input From Clinical Experts Consulted by CADTH

The clinical experts consulted for this review highlighted improved OS (prolonged life) and delay of disease relapse as important treatment goals in NSCLC curative settings. The clinical experts highlighted the infrequent use of neoadjuvant chemotherapy across jurisdictions in Canada because it has not been shown to provide a survival benefit over adjuvant therapy and noted that in the process of receiving neoadjuvant chemotherapy, some patients may become ineligible for surgery (due to disease progression or drug-related toxicity that may affect a patient’s performance status). The clinical experts highlighted an unmet need for patients with resectable NSCLC tumours, as some patients who undergo surgical resection and receive adjuvant chemotherapy may experience disease relapse. The clinical experts added that patients who experience relapse after surgery and/or adjuvant chemotherapy are generally incurable. Both clinical experts agreed that neoadjuvant nivolumab in combination with chemotherapy will cause a shift in the current Canadian treatment paradigm in the curative-intent setting.

The clinical experts indicated that patients with NSCLC who have tumours 4 cm or larger and/or node-positive disease, who have EGFR-negative or ALK-negative tumours, and who are eligible for upfront surgical resection (including patients with locoregional spread of the disease to lymph nodes who remain eligible for upfront surgical resection) would be eligible for treatment, provided they have no contraindications (severe and uncontrolled autoimmune disease, frailty, or poor baseline organ function). The experts explained that neoadjuvant nivolumab in combination with chemotherapy would not be appropriate for patients who are eligible for upfront surgery for borderline resectable NSCLC when the goal of neoadjuvant therapy is to downsize the tumour so the patient can become eligible for surgery. According to the clinical experts, patients are identified by a surgical and medical oncologist after appropriate review. The clinical experts added that the identification of driver mutations, like EGFR and ALK, is important but may not be routinely performed for early-stage disease at all centres.

The clinical experts added that patient response to treatment would be assessed in clinical practice using preoperative CT scans completed after neoadjuvant systemic chemotherapy, pathologic response, disease recurrence, and OS. The clinical experts emphasized that the schedule of follow-up assessments after completion of curative-intent surgery is not standardized across jurisdictions in Canada because of a lack of definitive literature suggesting the most appropriate timing for serial radiography. The clinical experts noted that nivolumab, platinum-doublet chemotherapy, or both could be discontinued at a patient’s request, in the event of disease progression during the 3 cycles of neoadjuvant therapy, or in the case of medically dangerous side effects or intolerable toxicity.

Clinician Group Input

Two clinician groups, the OH-CCO Lung Cancer Drug Advisory Committee (DAC) and LCC, each submitted inputs. The OH-CCO’s DAC provides guidance on drug-related issues, in support of CCO’s mandate, and collected information from 3 clinicians during a DAC meeting. LCC, a national charity and the only organization in Canada solely focused on lung cancer (education, advocacy, research) gathered information from published clinical data and 12 lung cancer medical oncologists across Canada.

Unmet Needs

According to the OH-CCO input, despite current treatments, a number of patients develop a recurrence quickly and do not survive. Therefore, neoadjuvant nivolumab would be an additional option for patients with resectable NSCLC. In its input, LCC cited several advantages of the neoadjuvant approach: it limits the risk of systemic dissemination of the cancer; downsizes tumours (leading to decreased postoperative complications [e.g., pain, infection], decreased performance status, improved surgical outcomes, and improved recovery times); makes the possibility of surgery easier, safer, and more efficacious; improves patient capacity to receive postoperative therapies; provides the opportunity for smoking cessation, physical therapy, and medical evaluations for surgery; helps to manage the surgical wait list; and improves the ability to provide prognosis and risk-stratification after surgery.

Place in Therapy

According to the OH-CCO clinicians, neoadjuvant nivolumab would be an additional option for patients with resectable NSCLC and could potentially replace adjuvant chemotherapy in some patients. Similarly, the LCC clinicians said that neoadjuvant nivolumab therapy would eliminate the need of postoperative, prolonged, and more expensive therapies (chemotherapy, radiation, immunotherapy). LCC clinicians noted that eligible patients might not choose neoadjuvant nivolumab therapy, such as patients with stage II, node-negative disease considered eligible for upfront surgery with optional adjuvant therapies, those at high risk of chemotherapy-associated or immunotherapy-associated complications, and those preferring upfront surgery.

Patient Population

The OH-CCO clinicians said that patients who meet the clinical trial inclusion criteria (i.e., resectable stage IIA to stage IIIB NSCLC [per AJCC 8th edition]) and/or those eligible for chemotherapy would be best suited for treatment with nivolumab. (Of note, the inclusion criterion for the CheckMate 816 trial was stage IB [tumour ≥ 4 cm] to stage IIIA NSCLC per AJCC 7th edition, which corresponds to stage IB to stage IIIB, non N3, non N2T4 per AJCC 8th edition.) LCC clinicians said that neoadjuvant therapy may be favoured for patients with stage IIIA and/or PD-L1-positive NSCLC, based on the favourable results in these strata in the CheckMate 816 study. However, the LCC group noted that a discussion with every eligible patient would be warranted and the treatment of all eligible patients with neoadjuvant therapy would be favoured to decrease the treatment burden after surgery (e.g., postoperative chemotherapy and/or atezolizumab). According to LCC clinicians, nivolumab would be least suited to patients with a contraindication to chemotherapy or immunotherapy treatments, such as those with renal failure, heart failure, severe hearing loss, severe neuropathy, an organ transplant, active and symptomatic autoimmune disease (e.g., Crohn disease being treated with immunosuppressive therapy or multiple sclerosis), and an Eastern Cooperative Oncology Group Performance Status [ECOG PS] of 2 to 4. The LCC clinicians added that a history of autoimmune disease or autoimmune disease that is clinically silent (e.g., immune thyroiditis) or well controlled without active immunosuppression is not a major contraindication. According to LCC input, the use of neoadjuvant immunotherapy in patients with EGFR, ALK, ROS1, RET, or NTRK alterations needs further investigation and/or will have to be addressed on a case-by-case basis, as these groups have not been specifically addressed in clinical trials.

Assessing Response to Treatment

The OH-CCO clinician group stated that response to nivolumab can be determined with clinical assessment of progression, the need for surgery, and pathologic assessment of a tumour. The LCC clinician group said that a CT scan would be needed after neoadjuvant therapy evaluate a patient’s eligibility for surgery. According to LCC clinicians, currently there is no clinical, biologic, or imaging tool that can be used to identify patients who will have a pathologic complete response (pCR) after neoadjuvant treatment and can be excluded from surgical treatment. At present, patients are followed with standard postoperative care, LCC clinicians reported. However, they anticipate that in the future, as more experience is gained with neoadjuvant approaches, risk-adapted follow-up strategies will be possible, based on, for example, circulating tumour DNA postoperative monitoring in combination with other clinical and pathologic features of the cancer.

Discontinuing Treatment

The OH-CCO clinicians said that intolerable toxicity and clinically obvious disease progression are factors to consider when deciding whether to discontinue nivolumab treatment. The LCC clinicians stated that clinical and biologic evaluations are performed at every cycle of therapy, as they are for patients undergoing chemotherapy and/or immunotherapy in the advanced disease setting, per standard practice in oncology.

Prescribing Conditions

The OH-CCO clinicians noted that a specialist, ideally part of a multidisciplinary team, is required in a hospital outpatient clinic to diagnose, treat, and monitor patients receiving nivolumab. The LCC clinicians also stated that, ideally, a multidisciplinary cancer tumour board, consisting of (nonexclusively) respirologists, radiologists, pathologists, thoracic surgeons, medical oncologists, and radiation oncologists, should discuss a multimodal treatment approach for patients who are usually referred to oncologic thoracic surgeons affiliated with major cancer centres. The LCC clinicians added that the delivery of care should be planned according to local structures; ideally, systemic therapies will be administered as close to a patient’s home as possible, while patients are continuously monitored by the cancer centre to coordinate neoadjuvant therapies with posttreatment imaging, preoperative evaluations, and the surgical admission itself.

Drug Program Input

Input was obtained from the drug programs that participate in the CADTH reimbursement review process. Key factors identified that could potentially impact the implementation of the CADTH recommendation for neoadjuvant nivolumab plus chemotherapy included considerations for the initiation of therapy, considerations for the discontinuation of therapy, considerations for the prescribing of therapy, and generalizability.

Clinical Evidence

Pivotal Study

Description of Study

CheckMate 816 is an ongoing, open-label, randomized, phase III trial comparing the efficacy and safety of 3 treatment regimens in patients 18 years and older with resectable (stage IB [tumours ≥ 4 cm], stage II, or stage IIIA) NSCLC: neoadjuvant nivolumab (3 mg/kg every 2 weeks for up to 3 cycles) in combination with ipilimumab (a single 1 mg/kg dose); neoadjuvant nivolumab (a 360 mg flat dose) in combination with platinum-based chemotherapy every 3 weeks for 3 cycles; and platinum-based chemotherapy alone.¹⁵ This CADTH review did not include findings from the ipilimumab plus nivolumab arm, as the indication under review is for nivolumab monotherapy in combination with platinum-based chemotherapy. Disease staging at screening was based on the AJCC/Union for International Cancer Control (UICC) TNM Classification of Malignant Tumours, 7th edition.¹⁵ After the completion of neoadjuvant treatment, all patients who remained operative candidates underwent definitive surgery for NSCLC within 6 weeks. Patients were also allowed to receive adjuvant chemotherapy with or without radiation after definitive surgery, per institutional standard at the discretion of the investigator.

pCR assessed by blinded independent pathologic review (BIPR) and event-free survival (EFS) assessed by blinded independent central review (BICR) were coprimary end points in the trial. Secondary end points included OS, time to death or distant metastases (TTDM), and major pathologic response (MPR). Safety, tolerability, and health-related quality of life (HRQoL) were exploratory outcomes.¹⁵ Radiologic tumour assessments were reviewed by a third-party vendor for BICR and BIPR. All investigator-assessed radiographic progressions and all disease recurrences were confirmed by BICR, per Response Evaluation Criteria in Solid Tumours Version (RECIST) 1.1 guidelines.

The CheckMate 816 trial was initially designed as a 2-arm trial in which patients were randomized in a 1:1 ratio to 1 of 2 treatment regimens: nivolumab plus ipilimumab (arm A); or platinum-doublet chemotherapy (arm B). An update to the protocol (protocol revision 2) introduced a third arm, nivolumab plus platinum-doublet chemotherapy (arm C) and allowed patients to be subsequently randomized in a 1:1:1 scheme to any treatment arm. A third update to the protocol (protocol revision 3) discontinued the randomization of patients to the nivolumab plus ipilimumab arm. Subsequently, patients enrolled in the study were randomized to the nivolumab plus chemotherapy arm or the chemotherapy arm in a 1:1 ratio and were stratified by 3 factors: PD-L1 expression level (1% or more versus less than 1%, not evaluable, indeterminate); disease stage (IB or II versus stage IIIA); and sex.

The primary data cut-off date for the prespecified final pCR analysis was September 16, 2020, and the data cut-off date for the first EFS interim analysis (IA1 EFS) was October 20, 2021. By the October 20, 2021, data cut-off date (IA1 EFS), ||| ||||||| |||||||| || ||| ||||| ||| ||||||||| ||||||||||| ||||| |||||||||| ||| |||| |||| || ||||||. Most patients in the nivolumab plus chemotherapy and the chemotherapy arms were male (71.5% and 70.9%, respectively), and about half were either ||||| |||||| ||| |||||| ||||||||||||| ||| ||||| |||||| ||| |||||| |||||||||||||. In total, 48.6% and 53.1% of patients in the nivolumab plus chemotherapy and chemotherapy arms, respectively, had squamous tumour histology; 63.1% and 64.2%, respectively, had stage IIIA disease, and 89.4% and 88.3%, respectively, were current or former smokers.¹⁵

Efficacy Results

Table 2 presents key efficacy end points and Table 3 presents key safety results from the CheckMate 816 trial.

Overall Survival

OS was formally tested at the IA1 EFS data cut-off date (October 20, 2021), as EFS was significant at the IA1 cut-off date. The median OS was not reached in either the nivolumab plus chemotherapy arm or the chemotherapy arm. The hazard ratio [HR] for death was 0.57 (99.67% confidence interval [CI] 0.30 to 1.07). The P value for OS (P = 0.008) did not cross the significance boundary (0.0033). A second OS testing is planned after ||| || |||||| |||| |||||||| ||| | |||| ||||| ||| ||| |||| |||||||||¹⁵

Event-Free Survival

By the October 20, 2021 cut-off date, ||| || ||| ||||||| |||||||| || ||| ||||||||| |||| |||||||||||| ||| ||| || || ||| ||||||| || ||| |||||||||||| ||| |||| |||||||| ||| ||| || ||||.¹⁵ In total, ||||| ||| || |||| EFS events had occurred in the nivolumab plus chemotherapy arm, and the median EFS was 31.57 months (95% CI, 30.16 to not available months). In the chemotherapy arm, ||||| ||| || |||| EFS events had occurred and the median EFS was 20.80 months (95% CI, 14.03 to 26.71 months). The estimated HR between nivolumab plus chemotherapy and chemotherapy was 0.63 (97.38% CI, 0.43 to 0.91), with a P value of 0.0052 from a stratified log-rank test.¹⁵ Sensitivity analyses for EFS were consistent with the primary analysis.¹⁵

Pathologic Complete Response and Major Pathologic Response

pCR and MPR analyses were conducted at the September 16, 2020, data cut-off date. The pCR rate per BIPR was 24.0% (43 of 179 patients; 95% CI, 18.0% to 31.0%) in the nivolumab plus chemotherapy arm and 2.2% (4 of 179 patients; 95% CI, 0.6% to 5.6%) in the chemotherapy arm. The stratified odds ratio between the nivolumab plus chemotherapy arm and the chemotherapy arm was 13.94 (99% CI, 3.49 to 55.75), with a P value of less than 0.0001, and the strata-adjusted difference based on the Cochran-Mantel-Haenszel method was 21.6% |||| ||| ||||| || ||||||.¹⁵ pCR sensitivity analyses were consistent with the primary analyses. In total, 36.9% of patients (95% CI, 29.8% to 44.4%) in the nivolumab plus chemotherapy arm and 8.9% (95% CI, 5.2% to 14.1%) in the chemotherapy arm were MPR responders.¹⁵

Health-Related Quality of Life

Patients completed the 3-Level EQ-5D (EQ-5D-3L) questionnaire at baseline, before on-treatment clinic visits, at postneoadjuvant visits 1 and 2, and at designated time points during the survival follow-up phase. Completion rates for the nivolumab plus chemotherapy and chemotherapy study arms were similar at baseline (EQ-5D-3L descriptive system: 94.4% versus 95.5%; EQ-5D-3L visual analogue scale [VAS]: 94.4% versus 95.0%) and did not change significantly at postneoadjuvant visit 1 (EQ-5D-3L descriptive system: 88.6% versus 83.9%; EQ-5D-3L VAS: 88.6% versus 83.9%). The EQ-5D index (based on the UK time trade-off value set) was collected for both study arms. The mean change from baseline to different time points (week 4, week 7, and postneoadjuvant visits 1 and 2) were minimal for the EQ-5D VAS and EQ-5D utility index scores of the questionnaire in the 2 treatment arms.¹⁵ Table 2 presents the change from baseline to different time points (week 4, week 7, and postneoadjuvant visits 1 and 2) in the EQ-5D VAS and EQ-5D utility index score portion of the EQ-5D-3L questionnaire.

Time to Death or Distance Metastasis

By the October 20, 2021, data cut-off date, || |||||| ||||||| ||| |||||||| || ||| ||||||||| |||| |||||||||||| ||| |||||| |||||| ||||||| || ||| |||||||||||| |||. The median time to death or distant metastasis was not reached in either study arm at the October 20, 2021, cut-off date, and the HR was 0.53 (95% CI, 0.36 to 0.77).¹⁵

EFS on Next Line of Therapy

By the October 20, 2021, data cut-off date, the median event-free survival on the next line of therapy (EFS2), per investigator assessment, was not reached in either the nivolumab plus chemotherapy or chemotherapy arm. || |||||| || |||||| ||| |||||||| || ||| ||||||||| |||| |||||||||||| ||| ||| || |||||| || ||| |||||||||||| |||. The estimated HR was 0.54 (95% CI, 0.37 to 0.80).¹⁵

Harms Results

Overall, 92.6% (n = 163) of patients in the nivolumab plus chemotherapy arm and 97.2% (n = 171) in the chemotherapy arm reported at least 1 adverse event (AE) in the CheckMate 816 trial. The most frequently reported AEs in the nivolumab plus chemotherapy arm were nausea (38.1%), constipation (33.5%), anemia (29.0%), decreased appetite (20.5%), fatigue (16.5%), and neutropenia (16.5%), and in the chemotherapy arm were nausea (44.9%), constipation (32.4%), anemia (26.7%), decreased appetite (23.3%), and neutrophil count decrease (21.0%) of any grade. AEs of grade 3 to 4 were reported in 40.9% (n = 72) and 43.8% (n = 77) of patients in the nivolumab plus chemotherapy and chemotherapy arms, respectively.¹⁵

Serious adverse events (SAEs) of any grade were reported in 30 (17.0%) patients in the nivolumab plus chemotherapy arm and 24 (13.6%) in the chemotherapy arm. SAEs of grades 3 or 4 were reported in 19 (10.8%) patients in the nivolumab plus chemotherapy arm and 17 (9.7%) patients in the chemotherapy arm.¹⁵

AEs leading to a dose delay or reduction were reported in || ||||||| ||| || ||||||| treated patients in the nivolumab plus chemotherapy and chemotherapy arms, respectively. AEs leading to discontinuation of study treatments were reported in 18 (10.2%) patients in the nivolumab plus chemotherapy arm and 20 (11.4%) in the chemotherapy arms.¹⁵

By the October 20, 2021, data cut-off date, ||||| ||| ||| || |||||||| ||| |||| || ||| ||||||||| |||| |||||||||||| ||| ||| ||||| ||| ||| patients in the chemotherapy arm.¹⁵ Notable harms reported in the nivolumab plus chemotherapy and chemotherapy arms of the CheckMate 816 trial are presented in Table 3.¹⁵

Table 2: Summary of Key Efficacy Results From the CheckMate 816 Trial

BICR = blinded independent central review; BIPR = blinded independent pathologic review; CI = confidence interval; CMH = Cochran-Mantel-Haenszel; EFS = event-free survival; EFS2 = event-free survival on the next line of therapy; HR = hazard ratio; HRQoL = health-related quality of life; IRT = Interactive Response Technology; MPR = major pathologic response; NA = not available; nivo = nivolumab; NR = not reported; pCR = pathologic complete response; PD-L1 = programmed cell death 1 ligand 1; SD = standard deviation; TTDM = time to death or distant metastases; VAS = visual analogue scale.

Note: The data cut-off date was October 20, 2021 (minimum follow-up was 21.0 months), except pCR and MPR were assessed at the September 2020 data cut-off date.

^aBased on Kaplan-Meier estimates.

^bHR of arm C to concurrent arm B from a Cox model stratified by PD-L1 expression level (≥ 1% vs. < 1%, not evaluable, or indeterminate), disease stage (IB or II vs. stage IIIA), and sex (male vs. female), as entered into the Interactive Response Technology.

^cLog-rank test stratified by the same factors as in the Cox proportional hazards model. The P value threshold for statistical significance was 0.0262.

^dPatients without samples for evaluation were counted as nonresponders.

^eCI based on the Clopper-Pearson method.

^fStrata-adjusted difference (arm C – concurrent arm B), based on the Cochran-Mantel-Haenszel method of weighting.

^gStratified by PD-L1 expression level (≥ 1% vs. < 1%, not evaluable, indeterminate), disease stage (IB or II vs. stage IIIA), and sex (male vs. female), as entered into the Interactive Response Technology.

^hStrata-adjusted odds ratio (arm C over concurrent arm B), the Mantel-Haenszel method.

ⁱP < 0.0001 (2-sided P value for pCR from the stratified Cochran-Mantel-Haenszel test).

Source: Clinical Study Report.¹⁵

Table 3: Summary of Key Safety Results From the CheckMate 816 Trial

AE = adverse event; SAE = serious adverse event.

Note: The data cut-off date was October 20, 2021.

Source: Clinical Study Report.¹⁵

Critical Appraisal

CheckMate 816 is an ongoing, randomized, open-label, phase III trial. Randomization was conducted using an Interactive Response Technology, and treatment allocation was concealed. Patients were stratified by 3 factors: PD-L1 expression level, disease stage, and gender or sex. The methods of randomization and the stratification factors were considered appropriate by the clinical experts consulted. Baseline characteristics in the 2 arms of interest were balanced, suggesting that randomization was successful. The methods of randomization and treatment allocation were considered appropriate.

The treatment effect for EFS and pCR was estimated during a prespecified interim analysis, adjusted using the Lan-DeMets alpha spending function with O’Brien-Fleming boundaries, which accounted for the actual number of events at an overall alpha of 4% or 5%. Overall, the treatment effect of the coprimary end points was estimated during the interim analysis. There is uncertainty about the magnitude of the treatment effect, given that interim analyses have the tendency to overestimate treatment effect.

OS was statistically nonsignificant at the preplanned interim analysis stopping rule (P = 0.008 against a prespecified level of significance at the interim analysis of 0.0033). Although the results showed a promising trend toward a significant treatment effect on OS, the final analysis may be needed to confirm the findings, particularly the exact estimate of the difference in median survival, which were not estimable at the data cut-off date. A follow-up OS analysis is planned after 128 OS events have occurred in both study arms.

Performance and assessment biases due to the open-label design of the trial were considered unlikely, given that radiologic assessments of CT scans for EFS and pathologic review of tumour sections were completed by a blinded independent review team, based on prespecified, and validated (RECIST 1.1) guidelines.

Further, the proportion of patients exposed to 3 doses of chemotherapy drugs was slightly higher in the nivolumab plus chemotherapy arm than in the chemotherapy arm. There were also slight differences reported in the cumulative dose intensity, which could bias the findings in favour of nivolumab plus chemotherapy.

More patients received subsequent anticancer therapy in the chemotherapy arm than in the nivolumab plus chemotherapy arm, and a higher proportion of patients received post-surgery adjuvant therapy in the chemotherapy arm, which could bias EFS and OS. However, the potential bias from the use of subsequent anticancer and adjuvant therapies was considered low.

The sponsor’s reimbursement request aligns with the Health Canada indication. The CheckMate 816 trial enrolled only patients with an ECOG PS of 0 or 1. The magnitude of benefit of nivolumab plus chemotherapy in patients with an ECOG PS of 2 or higher is uncertain.

The baseline and demographic characteristics in the trial were considered by the clinical experts to be generalizable to the population of patients with NSCLC in Canada. The experts highlighted notable differences between the patients enrolled in the CheckMate 816 trial and the patient population in Canada (younger, more stage IIIA disease, and more expression of PD-L1); however, the impact of these differences on the generalizability of the findings was considered low.

The dosing of nivolumab in the reimbursement request aligns with the Health Canada indication. Dose adjustments were allowed for chemotherapy drugs in the trial, but not for nivolumab, which aligns with the Health Canada indication. The experts indicated that a flat-dose approach to nivolumab, as implemented in the CheckMate 816 trial, would be used in practice.

Concomitant medications administered in the trial were considered appropriate by the clinical experts, and no major discrepancies in concomitant medications administered were identified that could have had a negative impact on the findings.

Indirect Comparisons

Description of Study

The sponsor provided a network meta-analysis (NMA)¹⁶ that compared the efficacy and safety of neoadjuvant nivolumab in combination with chemotherapy with other relevant treatments, including neoadjuvant chemotherapy, neoadjuvant chemoradiotherapy, adjuvant chemotherapy, and surgery alone, in patients diagnosed with resectable nonmetastatic NSCLC. The primary efficacy end points used for NMA estimates were EFS and OS. The secondary outcomes included time to locoregional recurrence (TTLRR), TTDM, and pCR. For each outcome, base-case and sensitivity analyses were carried out whenever data were available. The base-case analysis involved patients who were deemed candidates for surgery and who had undergone third-generation platinum-based doublet chemotherapy. The sensitivity analyses expanded to include second-generation platinum-based chemotherapies, resected patient populations, data stratified by PD-L1 expression level (i.e., ≥ 1% versus < 1%), and data from the intention-to-treat (ITT) population of the CheckMate 816 trial, instead of the subpopulations of the CheckMate 816 trial (i.e., stage IB to II, stage IIIA, and stage IIIA N2) used in the base-case stage-specific networks. Furthermore, for each outcome, in addition to the stage-agnostic network, which included studies regardless of the staging of the patient population, the network was stratified by tumour staging (i.e., stage IB to II, stage IIIA, and stage IIIA N2). Eight randomized controlled trials (RCTs) were eventually included in the base-case analyses, 5 additional RCTs were included in the sensitivity analyses that involved second-generation chemotherapies, and 4 additional RCTs were included in the sensitivity analyses expanded to include resected patients.

Efficacy Results

Although both EFS and OS were selected as primary end points in the sponsor-submitted NMA, OS data from the CheckMate 816 trial were based on an immature data cut-off date. According to the base analysis of EFS, patients with stage IIIA NSCLC who received neoadjuvant nivolumab in combination with chemotherapy had a significantly lower risk of an event (i.e., EFS HR < 1, credible interval [CrI] excluding 1) than those who received neoadjuvant chemotherapy or underwent surgery alone. The risk of an event in patients with stage IIIA N2 NSCLC who received neoadjuvant nivolumab in combination with chemotherapy was also significantly lower than in those who received neoadjuvant chemoradiotherapy or neoadjuvant chemotherapy. However, in patients with stage IB to II NSCLC, EFS improvement was not significantly different between neoadjuvant nivolumab plus chemotherapy and neoadjuvant chemotherapy, adjuvant chemotherapy, or surgery alone. For stage-agnostic patients (i.e., stage IB to stage IIIA), although the EFS results showed that neoadjuvant nivolumab plus chemotherapy led to a significantly lower risk of an event than neoadjuvant chemotherapy, adjuvant chemotherapy, or surgery alone, they were considered significantly biased because of the high heterogeneity in tumour staging, supported by significant differential treatment effects observed between patients with stage IB to II NSCLC and those with stage IIIA or stage IIIA N2 NSCLC.

Harms Results

The sponsor-submitted NMA did not quantitively synthesize evidence on safety outcomes due to the sparseness of the data and the differences in treatment regimens across the base-case studies. As a result, the sponsor provided a narrative description only.

Critical Appraisal

The systematic literature review (SLR) conducted by the sponsor to identify potentially eligible studies for the NMA was methodologically sound. The sponsor used a comprehensive literature search strategy, performed study selection and data extraction in duplicate, assessed the risk of bias appropriately, and described the characteristics of the included studies and narratively summarized the results in adequate detail. However, it was unclear whether the risk-of-bias assessment was carried out by a single assessor or multiple assessors. The reporting of the sponsor-submitted NMA generally followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) checklist.¹⁷

According to the clinical experts consulted by CADTH, the heterogeneity in tumour staging for patients with stage-agnostic NSCLC in the sponsor-submitted NMA was significant. And indeed, the effects of neoadjuvant nivolumab plus chemotherapy on EFS by stage, for example, were inconsistent between patients with stage IB to II NSCLC and those with stage IIIA or IIIA N2 NSCLC (i.e., statistically significant versus nonsignificant) compared to neoadjuvant chemotherapy in the base-case analysis. Because of the significant differential treatment effects of the 2 stage categories (IB or II versus IIIA or IIIA N2), the pooled indirect treatment comparison (ITC) results from the patients with stage-agnostic disease were deemed to be significantly biased, and the pooling to be inappropriate, methodologically, although neoadjuvant nivolumab plus chemotherapy markedly improved EFS relative to neoadjuvant chemotherapy, adjuvant chemotherapy, and surgery alone in patients with stage-agnostic NSCLC.

For each outcome examined in the NMA, both random-effects and fixed-effects models were run. The random-effects model was considered by the sponsor as the default model. However, the fixed-effects model was selected by the sponsor for all analyses in the NMA due to the sparseness of network not being able to estimate the between-study standard deviation with enough precision. Although it was considered appropriate to use the fixed-effects model instead of the random-effects model when the network is sparce, it is important to note that the fixed-effects model is not capable of capturing heterogeneity.

The misclassification of tumour stage with different versions of tumour classification criteria could have also had an impact on the ITC estimation. In the sponsor-submitted NMA, of the 8 RCTs in the base-case analysis, the CheckMate 816 trial was the only study to use TNM 7th edition to classify tumours; 2 other trials used the TNM 5th edition, 3 used the TNM 6th edition, and 2 used the 1997 International Staging System. Different staging criteria could lead to different classifications of NSCLC and would consequently result in differences in prognosis estimation and treatment selection.

Safety outcomes were only narratively described in the NMA. Without a quantitative synthesis, a balanced judgment of comparative benefits relative to comparative harms could not be made. In addition, outcomes that are important to patients, such as HRQoL, were not reported in the NMA. Furthermore, analyses comparing neoadjuvant nivolumab plus chemotherapy with adjuvant atezolizumab was determined by the sponsor to not be feasible and to be inappropriate because of significant methodological challenges. Nonetheless, feedback from the clinical experts consulted by CADTH emphasized that adjuvant atezolizumab is an appropriate treatment option for patients with resected stage IB to stage IIIA NSCLC (the 7th lung cancer TNM classification) with a PD-L1 expression level of at least 50%. Therefore, the lack of relevant analyses might have introduced uncertainty into the sponsor’s submitted analysis.

Other Relevant Evidence

Description of Study

As part of the submission for nivolumab, the sponsor submitted a systematic review and a meta-analysis that described clinical evidence in patients with nonmetastatic resectable NSCLC. The sponsor-conducted meta-analysis assessed the potential use of pathologic response (pCR or MPR) as a surrogate end point for long-term outcomes (EFS, OS) in patients with resectable NSCLC. The systematic review was informed by patient-level data from 32 studies that presented evidence of an association between OS and/or EFS and pCR and/or MPR, measured as an HR, or reported Kaplan-Meier curves for OS and/or EFS by pCR and/or MPR status, which allowed for the reconstruction of HRs.¹²

Efficacy Results

||||||| |||||||| || |||||||||||| |||||||| |||||||| |||||| |||| |||||| || |||||||||||||||| ||||||||||| ||||||| || ||||||||| ||| || ||| |||| |||| ||| |||| || ||||| |||||| ||| || ||||||| ||||||||| |||||||||| |||| ||||||||| ||| ||| |||||||||| |||| |||||| |||||||||||| |||||||||||| ||||||||| |||| ||||||| || ||| || || ||| |||||||| |||| |||| |||||||| ||| |||| |||| ||| |||| || ||||||

Critical Appraisal

The sponsor-submitted meta-analysis was informed by studies selected from an adequately conducted systematic review, with a clearly prespecified patient, intervention, comparison, outcome (PICO) model, in accordance with the PRISMA guidelines. The study-selection and data-extraction methods were considered appropriate. In the absence of detailed information on the baseline and study characteristics of patients enrolled in the studies included in the meta-analysis, the degree of heterogeneity between the included studies could not be assessed. Studies included had considerable variation in terms of study design (observational versus RCT) and sample size. Heterogeneity was not reported for the Bayesian analysis and, owing to the lack of baseline data on the trials included an assessment of the level of heterogeneity could not be made. The meta-analysis suggests that achievement of a pCR was associated with improved OS, based on the Bayesian and frequentist methods implemented in the analyses.

Description of Study

The sponsor submitted 1 real-world study conducted with data from electronic health records (EHRs) supplemented with chart reviews. The purpose of this retrospective, observational study was to generate real-world evidence to characterize the relationship between pathologic responses and survival and to describe patient profiles and neoadjuvant treatment patterns for patients with surgically resectable NSCLC (stage IB [tumour ≥ 4 cm] to stage IIIA) treated in the community oncology setting in the US. Neoadjuvant treatment regimens were characterized as chemotherapy or chemoradiotherapy in the study.

Efficacy Results

||| |||||| ||||||||| ||| |||| ||||||| ||| || |||||| ||||| ||||| |||||||| || ||||||| |||||||||| || |||||| |||||| || || ||||||| |||||||| |||||||| |||| ||| |||||| || || ||||||| |||||||| |||||||| || |||| ||||||||| |||||||| |||| ||| || ||||| ||||||| |||| ||| |||||||||| || || ||||||||||| ||| |||||||| ||| || ||| |||| |||| ||| |||| || |||||| |||| |||||||| ||| |||| ||||||| ||||| |||||| ||| ||||||||||| |||||||| ||| ||||||| |||||||||| |||| |||||||| || ||||||||| ||| |||| |||| ||| |||| || ||||||| |||| |||||| || ||||| |||||||| |||| |||| ||| |||| |||| ||| |||| || |||||| |||||| |||||||| || |||| |||| || |||| || ||||| |||||| ||||| |||||||| ||||||| ||||| |||| |||||||| |||| ||| |||| |||||||| || ||||| ||||||| |||| ||| |||||||||| || || ||||||||||| ||| |||||||| ||| || ||| |||| |||| ||| |||| || |||||| |||| |||||||| ||| |||| ||||||| ||||| |||||| ||||||| ||||| |||||| ||| ||||||||||| |||||||| ||| ||||||| |||||||||| |||| |||||||| || ||||||||| ||| |||| |||| ||| |||| || ||||||| ||| ||||||| ||| |||||||||||| ||||||| ||||||| ||| ||| |||||| ||||||| ||||||||| |||| |||||| || ||||| |||||||| |||| || ||| ||| |||| |||| || |||| || |||||| |||||| |||||||| || |||| |||| ||| |||| || ||||| |||||| ||||| |||||||| ||||||| || |||| ||| ||||| ||| ||||||||| |||| ||||| || || |||||| ||||||| |||||| |||| |||||||||| || ||| ||||||||| ||| |||| ||| |||| |||||||||| ||| ||||||||| ||||||||| || |||||||| ||||||||||||| |||||| ||||||||| |||||||| ||||||| || ||| || |||||||| ||||||||

Critical Appraisal

Several limitations of the study were identified: neoadjuvant treatment consisted only of chemotherapy and chemoradiotherapy, several variables had incomplete or missing data in the analyses, and sample size was reduced (and had reduced power) to assess the relationship between pCR and/or MPR and survival end points. Thus, it is difficult to conclude that this retrospective, observational study (i.e., real-world evidence), which characterizes the relationship between pathologic responses and survival, as well as patient profiles and neoadjuvant treatment patterns in patients with surgically resectable NSCLC (stage IB [tumours ≥ 4 cm] to stage IIIA) treated in the community oncology setting in the US, provides all the information required to address gaps in the evidence in support of the validity of end points in the pivotal trial.

Conclusions

Although statistically significant EFS and favourable OS outcomes were observed in the CheckMate 816 trial for nivolumab plus chemotherapy compared to chemotherapy alone for the neoadjuvant treatment of adults with resectable NSCLC (tumours ≥ 4 cm or node-positive disease), there remains uncertainty about the magnitude of the treatment effect, given that the results are based on interim analyses, OS data were immature at the October 20, 2021, cut-off date, and the HR for death was not statistically significant. Although EFS is a validated surrogate for OS, final analyses may be warranted to increase confidence in the trial findings. Prespecified EFS subgroup analyses suggested that the majority of the benefit derived from nivolumab plus chemotherapy was in subgroups of patients with stage IIIA disease and a PD-L1 expression level above 50%. However, efficacy results from these subgroup analyses should be interpreted with caution, as the study was not statistically powered to assess subgroups individually. The clinical experts indicated that patients with the highest risk of recurrence (i.e., patients with stage IIIA disease) and with a PD-L1 expression level above 50% may derive more benefit from the treatment and clinicians may be more apt to treat them with this regimen. However, regardless of the magnitude of response observed, a benefit was observed in all subgroups and in the overall study population. No detriment to quality of life was observed for patients in the trial, and the mean differences from baseline to predefined time points did not exceed minimally important differences (MIDs) for the VAS and utility score portion of the EQ-5D-3L questionnaire. Notable differences between the trial population and patients treated in Canada were related to the younger population in the trial than in the patient population in Canada, the higher proportion of patients with stage IIIA NSCLC in the trial than is seen in clinical practice, and the lack of a comparator arm with the current standard of care. However, the clinical experts noted that these differences were unlikely to affect the generalizability of the efficacy results to the patient population in Canada. The clinical experts noted that this treatment will typically be administered to patients with surgically resectable NSCLC (tumours ≥ 4 cm and or node-positive disease) and a good ECOG PS (0 and 1). The safety profile of nivolumab was consistent with the known safety profile of immunotherapies, and no new safety signals were identified. The AEs reported were considered manageable in practice by the clinical exerts consulted.

It is worth noting that the comparator in the CheckMate 816 trial (neoadjuvant chemotherapy) is not the current standard of care in Canada for patients with resectable NSCLC (which is upfront surgery with curative intent, followed by adjuvant chemotherapy), and no direct evidence was available to assess the efficacy of nivolumab plus chemotherapy relative to surgery followed by adjuvant chemotherapy. Findings from the sponsor-submitted NMA suggest an improvement in EFS in the nivolumab plus chemotherapy arm for patients with stage IIIA resectable NSCLC, compared with surgery and adjuvant chemotherapy. However, there were sources of uncertainty identified, such as the significant heterogeneity in the RCTs included, sparse networks, and immature data for OS. In addition, there were no quantitative safety data presented, so no conclusions could be drawn. The sponsor-submitted meta-analysis and real-world evidence showed favourable OS outcomes with neoadjuvant nivolumab for those who achieved a pCR compared to those who did not. However, both studies have study design limitations that preclude definitive conclusions.

Introduction

Disease Background

Lung cancer is the most frequently diagnosed cancer in Canada, and the leading cause of cancer-related death.¹ About 30,000 new diagnoses (50% in males and 50% in females)² and 20,700 cancer-related deaths were projected in Canada in 2022,^1,3 with about 98% of all cases anticipated in people 50 years and older.⁴ The adjusted 5-year net survival (based on 2015 to 2017 estimates) for all forms of lung cancer is only 22% (19% in males and 26% in females in Canada).^1,2 NSCLCs are the most common forms of lung cancer, accounting for more than 80% of all lung cancers in Canada,^5,6 and adenocarcinomas are the most commonly diagnosed forms, accounting for 48% of new cases in Canada.⁴ The 5-year survival rate for patients with NSCLC is poor and highly dependent on disease stage at diagnosis; it is higher in patients with early-stage (stage IB) disease (66% to 68%) than in patients with stage IIA disease (52% to 60%), stage IIB disease (47% to 53%), and stage IIIA disease (36%).¹⁸

Diagnosis is based on symptoms and histology at initial presentation.^19,20 Staging is key in determining disease prognosis and facilitates treatment selection.^20,21 The current standard staging system recommended in Canada is the 8th edition of the AJCC/UICC staging guidelines, which involves a tumour, node, metastasis (TNM) classification of the disease based on the size and spread of the primary tumour (T), lymph node involvement (N), and occurrence of metastasis (M).²² About 47.1% of all new cases of NSCLC are diagnosed at stage IV, 19.0% at stage III, 9.1% at stage II, and only 23.1% at stage I.⁷ Patients with early-stage NSCLC have a higher risk of disease recurrence than patients with later-stage disease, despite curative surgery.²³

General symptoms of NSCLC include worsening cough, chest pain, hemoptysis, malaise, weight loss, dyspnea, and hoarseness at clinical presentation or upon chest imaging.^4,20 Patients with advanced or metastatic disease may experience additional symptoms such as trouble breathing, chronic cough and chest pain, pain in bones or the spine, yellowing of the skin or eyes, weakness or numbness of arms or legs, fatigue and unexplained weight loss, depression, insomnia, and pain.^21,24 Diagnostic procedures currently recommended include imaging with CT, PET, and/or MRI scans, bronchoscopy with or without endobronchial ultrasound, and tissue biopsy.²⁵

Available treatment options for NSCLC depend on disease stage at diagnosis, the potential for tumour resection with surgery, and eligibility of the patient for surgery (e.g., performance status, comorbidities). Treatment options include surgery, radiotherapy, and systemic therapy (with chemotherapy or immunotherapy). Patients may receive treatments before surgery (neoadjuvant setting), after surgery (adjuvant setting), before and after surgery (perioperative therapy), or after they have achieved disease control (consolidation therapy with radiation with or without chemotherapy).¹² For patients with early-stage disease, eligibility for surgical resection depends on various clinical factors that determine the safety of a lung resection (e.g., overall respiratory function, tumour location, and the presence of ipsilateral vocal cord paralysis, Horner syndrome, phrenic nerve palsy, and neurologic deficits associated with Pancoast superior sulcus tumours) and on general health factors that determine eligibility for major surgery (e.g., comorbidities such as active cardiac disease).²⁵

Standards of Therapy

The treatment goal for patients with resectable NSCLC is to cure, improve 5-year OS, and prevent disease recurrence.

Surgery with curative intent is the current gold standard for patients with clinical stage I to stage IIIA NSCLC amenable to resection.^8,9 The standard of care for completely resectable stage IIA or IIB and stage IIIA NSCLC (AJCC 7th edition), according to joint guidelines from the American Society of Clinical Oncology and CCO, is surgical resection followed by adjuvant cisplatin-based chemotherapy.¹⁰ Cisplatin-based chemotherapy is not recommended for patients with stage IA disease and not routinely recommended for patients with stage IB disease; however, a postoperative evaluation is recommended.^10,26 The British Colombia guidelines recommend adjuvant platinum-based therapy (cisplatin-based treatment is preferred or, if cisplatin is contraindicated, carboplatin-based treatment) as the standard of care for patients with completely resected stage II or stage IIIA (AJCC 7th edition) disease.²⁷

Neoadjuvant chemotherapy is seldom used in Canada for resectable NSCLC, as it has not been shown to provide a survival benefit over adjuvant therapy, and in the process of pursuing neoadjuvant chemotherapy, some patients may become ineligible for surgery (owing to disease progression or treatment-related toxicity), according to the clinical experts consulted. Still, neoadjuvant therapy has several advantages: it can reduce tumour size, increase resectability, and remove micrometastasis and tumour cells in distant lymph nodes, thereby reducing the risk of recurrence from tumour cells that are not removed by surgery.^12,13

Stereotactic ablative radiation with curative intent is available to some patients with early-stage disease who are ineligible for surgery (e.g., because of significant comorbidities that make them a high risk for general anesthetic) or who refuse surgery, according to the clinical experts. Patients with borderline resectable stage III cancer may be offered chemotherapy and/or radiation in current practice before surgery, whereas those with unresectable stage III NSCLC will be typically managed with combined chemoradiation followed by a year of consolidation with durvalumab if the patient has stable or responsive disease at postchemoradiation reassessment, the clinical experts explained.¹¹ Adjuvant chemotherapy options consist of cisplatin plus vinorelbine, carboplatin plus paclitaxel for patients with a contraindication to cisplatin, gemcitabine plus carboplatin, paclitaxel plus carboplatin, pemetrexed plus cisplatin, and immune-checkpoint inhibitors (atezolizumab and durvalumab). The experts indicated that patients with node-positive NSCLC or who have tumours 4 cm or larger may receive additional postoperative treatment, such as adjuvant chemotherapy (4 cycles of cisplatin-based doublet chemotherapy with vinorelbine or pemetrexed) or adjuvant osimertinib (for 3 years, but only for patients with the common exon 19 deletion or an exon 21 insertion L858R EGFR mutation).

For patients with PD-L1 Tumour Proportion Scores above 50%, a recent Health Canada NOC for atezolizumab has recommended funding in the adjuvant setting after adjuvant chemotherapy; however, at present, it is not available outside private insurance or patient self-pay options. Postoperative adjuvant radiation is limited to use in patients with R1 resections (i.e., positive surgical margins). The experts noted that evidence from the IMPOWER010 trial supports a year of adjuvant therapy with atezolizumab after adjuvant chemotherapy. However, the Health Canada NOC limits this treatment to patients whose tumours exhibit a PD-L1 Tumour Proportion Score of 50% or more. Although this therapy has received an initial recommendation for funding by CADTH, it is not yet funded in any health jurisdiction in Canada, nor is it available through patient support programs. Adjuvant chemotherapy leads to an absolute improvement in 5-year OS of approximately 5% for all outcomes (although this is likely an underestimate in patients with higher-stage disease,²⁸ and patients with stage IIIA disease [according to the AJCC 7th edition] have a high risk of recurrence [approximate 5-year survival in these patients is 30%]).

The clinician group input generally highlighted treatment options for resectable NSCLC similar to those described by the clinical experts consulted during this review. The clinician group input noted that adjuvant chemotherapy is not indicated for patients with stage I cancers, based on the 8th edition of the UICC staging system (tumours < 4 cm, node-negative disease).

Drug

Nivolumab is a human immunoglobulin G4 monoclonal antibody that binds to the PD-1 receptor and blocks its interaction with PD-L1 and PD-L2, releasing PD-1 pathway-mediated inhibition of the immune response, including antitumour immune response. In syngeneic mouse tumour models, blocking PD-1 activity resulted in decreased tumour growth.¹⁴ The Health Canada–recommended dose of nivolumab is 360 mg administered intravenously over 30 minutes as neoadjuvant treatment in combination with platinum-doublet chemotherapy every 3 weeks for 3 cycles.¹⁴ The key characteristics of nivolumab, carboplatin, paclitaxel, and pemetrexed are presented in Table 4.

Nivolumab underwent a review at Health Canada under Project Orbis and received a NOC on August 18, 2022, for the neoadjuvant treatment of adult patients with resectable NSCLC (tumours ≥ 4 cm or node-positive) in combination with platinum-doublet chemotherapy.¹² The sponsor’s reimbursement request and population aligns with the Health Canada indication. This submission is a post-NOC request.

Stakeholder Perspectives

Patient Group Input

This section was prepared by CADTH staff based on the input provided by patient groups.

LCC, a member of Global Lung Cancer Coalition and the only national organization in Canada focused exclusively on lung cancer, engages in patient support, education, research, and advocacy. LCC submitted patient group input based on interviews conducted in Canada in September and October 2022 ([1 patient with stage I/II NSCLC, 1 patient with stage III NSCLC, 2 patients with stage IV NSCLC, and 1 caregiver for a patient with large-cell neuroendocrine carcinoma] and [1 interview with a patient who had stage IV NSCLC and had undergone an environmental scan]). All participants had experience with nivolumab. The patient input summary will focus on the 2 patients with stage I to stage III NSCLC to align with the requested indication.

One patient said she used to be extremely active and was an avid runner for 10 years before her lung cancer diagnosis. Lung cancer made exercise harder and made her feel more tired than before, which had an impact on her independence, she explained. She experienced cough and some mild chest pain before her diagnosis of lung cancer. Another patient who had been diagnosed with early-stage NSCLC had two-thirds of his lung surgically removed, leaving him with 50% of his initial lung capacity. As a result, he became unable to do any vigorous exercise or activity and tires quickly. The input indicated that patients value treatment options that maintain or improve quality of life, delay the onset of symptoms, improve survivorship, improve, or maintain functionality and/or mobility, and ultimately provide a cure. Also, patients prefer treatments that can be administered at a hospital located near home, or in a community clinic for those in rural settings, to minimize travel time and burden on caregivers. The input emphasized that because the CheckMate 816 trial excluded patients with EGFR or ALK alterations, owing to a lack of evidence of supporting neoadjuvant immunotherapy in this population, a biomarker screening (a routine practice) must be performed before neoadjuvant treatment with nivolumab.

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., providing guidance on the development of the review protocol, assisting in the critical appraisal of clinical evidence, interpreting the clinical relevance of the results, and providing guidance on the potential place in therapy). The following input was provided by 2 clinical specialists with expertise in the diagnosis and management of NSCLC.

Unmet Needs

The clinical experts consulted noted that the treatment goal in this setting is to cure patients (improve survival by eradicating micrometastasis) and prolong life. Secondary treatment goals include the delay of disease relapse, which will allow patients to live longer in a disease-free state.

The clinical experts noted that many patients receiving the current standard of care (surgery followed by adjuvant chemotherapy) still experience disease relapse, after which the disease generally becomes incurable. The clinical experts added that although adjuvant chemotherapy adds a survival benefit to patients in current practice, the benefit observed is modest and adjuvant treatment is associated with a toxicity burden. The clinical experts explained that many patients treated with the current standard of care decline or are ineligible for adjuvant chemotherapy because of the potential for AEs related to cisplatin doublet chemotherapy, some of which may be irreversible (e.g., kidney or nerve damage, impaired hearing). The clinical experts also mentioned that many patients are not clinically fit to undergo 4 cycles of cisplatin-based adjuvant treatment, and therefore do not receive any adjuvant therapy. Thus, there is an unmet need for new treatments for patients with resectable NSCLC.

Place in Therapy

The clinical experts explained that nivolumab would change the treatment paradigm for early-stage resectable NSCLC, as neoadjuvant therapy is not common practice in Canada. The clinical experts indicated that there might be a shift in the treatment paradigm for patients who experience disease recurrence while receiving neoadjuvant nivolumab and chemotherapy, or shortly after completion (usually within 6 months), given that the standard first-line therapy for patients diagnosed with incurable disease is based on an immunotherapy backbone with single-drug pembrolizumab or cemiplimab, or on platinum-doublet chemotherapy combined with pembrolizumab or nivolumab and ipilimumab. However, the clinical experts noted that backbone immunotherapy may not be appropriate if clinicians believe the disease is resistant to PD-1 inhibition. The clinical experts also noted that the order of therapies in the current treatment paradigm may shift, given that patients were allowed to receive additional postoperative adjuvant chemotherapy and/or radiation at the discretion of the treating physician in the CheckMate 816 trial. For the majority of patients treated with neoadjuvant chemotherapy and nivolumab in practice, this would be considered a replacement for adjuvant chemotherapy, and in most cases no further adjuvant systemic therapy would be offered.

According to the clinical experts, nivolumab plus chemotherapy would only be used in the curative-intent setting before surgery to increase the time before disease recurrence and, presumably, reduce the number of patients who will experience disease recurrence, thereby increasing cure rates and OS. The clinical experts noted that this treatment may increase the number of patients who have a complete (R0) surgical resection, and may reduce the extent of the lung removed during surgery (e.g., a patient who otherwise would have had to have a full lung removed [pneumonectomy] may be able to have a smaller surgery limited to the removal of a single lobe of the lung [lobectomy] after the downsizing effect of preoperative chemotherapy plus nivolumab). The clinical experts noted that nivolumab will be administered in practice, in addition to 3 cycles of chemotherapy, to patients with tumours 4 cm or larger or with node-positive disease.

Patient Population

The clinical experts consulted indicated that eligible patients will include those who have tumours larger than 4 cm and/or node-positive disease, who have no EGFR or ALK mutations, and who are deemed fit for upfront surgical resection, including patients with locoregional spread of disease to the lymph nodes but who remain fit for upfront resection, provided there were no contraindications (e.g., severe or uncontrolled autoimmune diseases, frailty, poor baseline organ function). The experts noted that nivolumab would not be appropriate for patients considered to have borderline resectable disease fit for upfront surgery and for whom the goal of neoadjuvant therapy would be to downsize the tumour enough to make surgical resection feasible.

The clinical experts indicated that patients with the highest risk of recurrence (i.e., patients with stage IIIA disease) and a PD-L1 expression level above 50% would likely derive benefit from the treatment, and clinicians may be apt to treat them with this regimen.

According to the clinical experts, patients will be identified by a thoracic surgeon after appropriate review. The experts noted that surgical resectability is usually determined by a thoracic surgeon, and eligibility depends on multiple factors, including a patient’s lung function (measured with pulmonary lung function tests). The clinical experts indicated that the process of determining tumour resectability for patients with early-stage NSCLC is, to a certain extent, subjective. They noted that, ideally, tumour scans are reviewed by a thoracic tumour board, but added that the process may not always be feasible. Patients should undergo baseline staging with a PET scan, MRI brain scan, and mediastinal lymph node sampling. Driver mutations (EGFR or ALK) are important but may not be routinely assessed in all centres in patients with early-stage disease; however, the experts noted that patients with unknown EGFR and ALK mutational status should not be excluded from receiving the treatment. The clinical experts added that underdiagnosis and misdiagnosis are not issues in this setting. Further, there are projections that current screening strategies in Canada may increase the number of patients with early-stage disease who are asymptomatic and are candidates for curative-intent surgery.

Assessing Response to Treatment

The clinical experts noted that patient response to treatment is assessed with preoperative CT scans done after the completion of neoadjuvant systemic therapy, pathologic response, and OS. According to the clinical experts, the interpretation of restaging results can be difficult, given that some patients who receive neoadjuvant checkpoint inhibitors tend to exhibit pseudoprogression, especially in the lymph nodes. The clinical experts noted that any areas of possible progressive disease would be verified with tissue sampling, when feasible, before it is determined that a patient is no longer eligible for surgery because of disease progression.

According to the clinical experts, follow-up after completion of curative-intent surgery is not standardized in current practice, given the lack of definitive literature suggesting the most appropriate timing interval for serial radiography. The clinical experts noted that, in practice, response assessments for patients include surveillance CT scans (typically every 6 months for 2 years, then yearly to 5 years after therapy), and that this would not change because of the use of neoadjuvant nivolumab. In addition, follow-up procedures currently in place would not be affected in patients treated with neoadjuvant chemotherapy and nivolumab; they are already in place for patients who have undergone curative-intent surgery. A clinically meaningful response would be an improvement in OS.

Discontinuing Treatment

According to the clinical experts, chemotherapy and/or nivolumab could be discontinued at a patient’s request in the event of disease progression during the 3 cycles of neoadjuvant therapy, autoimmune or medically dangerous side effects, or intolerable toxicity.

Prescribing Conditions

According to the clinical experts, nivolumab administration should be monitored by a qualified specialist medical doctor (a medical oncologist or, in very rare cases, a respirologist) with experience in the treatment of lung cancer and the use of immunotherapy checkpoint inhibitors. Hospital outpatient settings in which personnel have experience with immunotherapy checkpoint inhibitors and chemotherapy would be appropriate treatment settings.

Additional Considerations

The clinical experts highlighted the current lack of clinical evidence to support the use of adjuvant nivolumab. The experts added that there is no evidence to support the use of the Health Canada–approved atezolizumab in the adjuvant setting for patients who have received neoadjuvant chemotherapy without nivolumab.

The clinical experts advised against defining eligibility for treatment in terms of staging, as there will be further evolution of these manuals in the near future. The experts noted that in the AJCC 7th edition, T2 tumours are defined as being larger than 3 cm and up to 7 cm. On further subdivision, T2a tumours are defined as being larger than 3 cm and up to 5 cm, and T2b tumours are defined as being larger than 5 cm and up to 7 cm. If a tumour is T2aN0 (node negative), it is classified as stage IB. If a tumour is T2bN0, it is classified as stage IIA. The indication being examined for neoadjuvant nivolumab plus chemotherapy includes patients with node-negative disease and tumours 4 cm or larger, so some stage IB tumours, according to the AJCC 7th edition, would qualify for neoadjuvant nivolumab plus chemotherapy and others would not. In the AJCC 8th edition, T2a tumours have been redefined as being larger than 3 cm and up to 4 cm, T2b tumours are now larger than 4 cm and up to 5 cm, and T3 tumours are now larger than 5 cm and up to 7 cm. The overall staging for these groups has also shifted; in the AJCC 8th edition, T2aN0 tumours remain classified as stage IB, but T2bN0 tumours are now classified as stage IIA and T3N0 tumours are now classified as stage IIB. The key difference is that the cancers included in the CheckMate 816 trial, which used the AJCC 7th edition definition of stage IB (node-negative tumours from 4 cm to 5 cm), would be classified as stage IIA in the AJCC 8th edition.

Patients with N2 nodal disease limited to a single nodal station are generally considered surgical candidates, as long as there is no local invasion that would render a complete surgical resection unfeasible, according to the clinical experts. In the AJCC 7th edition, patients with T2b (> 5 cm to 7 cm), N2, or T3N2 disease were considered stage IIIA, and would have been enrolled in the CheckMate 816 trial if they were candidates for upfront resection. In the AJCC 8th edition, the clinical experts noted that patients with primary tumours larger than 5 cm and up to 7 cm are considered T3, and T3N2 tumours have been upstaged from stage IIIA to stage IIIB. Further, tumours that were classified as T3 in the AJCC 7th edition (> 7 cm or with invasion of the diaphragm) are now classified as T4 in the AJCC 8th edition, and T4N2 tumours have been upstaged from stage IIIA to stage IIIB. According to the clinical experts, this means that there are patients with stage IIIB disease who are resectable according to the AJCC 8th edition who would have been considered stage IIIA in the AJCC 7th edition and thus eligible for neoadjuvant chemotherapy plus nivolumab.

Clinician Group Input

This section was prepared by CADTH staff based on the input provided by patient groups.

Two clinician groups, the OH-CCO Lung Cancer DAC and LCC submitted 2 separate inputs. The OH-CCO’s DAC provides guidance on drug-related issues in support of CCO’s mandate and collected information from 3 clinicians during a DAC meeting. LCC, a national charity and the only organization in Canada solely focused on lung cancer (education, advocacy, research), gathered information from published clinical data and 12 lung cancer medical oncologists from across Canada.

Unmet Needs

According to the OH-CCO group, despite current treatments, a number of patients experience a recurrence quickly and do not survive. Therefore, neoadjuvant nivolumab would be an additional option for patients with resectable NSCLC. LCC cited several advantages of neoadjuvant approaches: limiting the risk of systemic dissemination of the cancer; tumour downsizing (leading to decreased postoperative complications [e.g., pain, infection], decreased performance status, improved surgical outcomes, and improved recovery times); the possibility of easier, safer, and more efficacious surgeries; improving patient capacity to receive postoperative therapies; providing the opportunity for smoking cessation, physical therapy, and medical evaluations for surgery, helping to manage the surgical wait list; and improving the ability to provide prognosis and risk-stratification after surgery.

Place in Therapy

According to the OH-CCO clinicians, neoadjuvant nivolumab would be an additional option for patients with resectable NSCLC and could potentially replace adjuvant chemotherapy in some patients. Similarly, LCC clinicians said that neoadjuvant nivolumab therapy would eliminate the need for postoperative, prolonged, and more expensive therapies (chemotherapy, radiation, immunotherapy). LCC clinicians said it is not clear whether all eligible patients would choose neoadjuvant nivolumab therapy, such as patients with stage II, node-negative disease considered for upfront surgery with optional adjuvant therapies, those at high risk of chemotherapy-associated or immunotherapy-associated complications, and those preferring upfront surgery.

Patient Population

The OH-CCO clinicians said that patients who meet the clinical trial inclusion criteria (i.e., those with resectable stage IIA to stage IIIB NSCLC [AJCC 8th edition] and/or those eligible for chemotherapy) would be best suited for treatment with nivolumab. The LCC clinicians said that neoadjuvant therapy may be favoured for patients with stage IIIA and/or PD-L1-positive disease, based on the favourable results in these strata shown in CheckMate 816 study. However, the LCC group said that a discussion with every eligible patient appears warranted, and that treating all eligible patients with a neoadjuvant would be favoured to decrease the treatment burden after surgery (e.g., postoperative chemotherapy and/or atezolizumab). According to the LCC clinicians, the least suitable patients would be those with a contraindication to chemotherapy and/or immunotherapy, such as those with renal failure, heart failure, severe hearing loss, severe neuropathy, an organ transplant, active and symptomatic autoimmune disease (e.g., patients with Crohn disease on immunosuppressive therapy or those with multiple sclerosis), and an ECOG PS of 2 to 4. The LCC group added that a history of autoimmune disease or autoimmune disease that is clinically silent (e.g., immune thyroiditis) or well controlled without active immunosuppression is not a major contraindication. Last, according to the LCC input, the use of neoadjuvant immunotherapy in patients with EGFR, ALK, ROS1, RET, or NTRK alterations needs further investigation and/or will have to be addressed on a case-by-case basis, as these groups have not been specifically addressed in clinical trials.

Assessing Response to Treatment

The OH-CCO clinician group stated that outcomes such as clinical assessment (to ensure no progression), surgery, and pathological assessment can be used to determine response to nivolumab. The LCC clinician group said that an additional CT scan would be needed after neoadjuvant therapy to evaluate a patient’s eligibility for surgery. According to LCC clinicians, currently there is no clinical, biologic, or imaging tool that can be used to identify patients who will have a pCR on pathological assessment after neoadjuvant treatment and can be excluded from surgical treatment. At present, patients are followed with standard postoperative care, the LCC clinicians reported. However, they anticipate that in the future, as more experience is gained with neoadjuvant approaches, risk-adapted follow-up strategies will be possible, based on, for example, postoperative monitoring of circulating tumour DNA in combination with other clinical and pathologic features of the cancer.

Discontinuing Treatment

The OH-CCO clinicians said that intolerable toxicity and clinically obvious disease progression are factors to consider when deciding whether to discontinue nivolumab treatment. The LCC clinicians stated that clinical and biologic evaluations are performed at every cycle of therapy, as they are for patients undergoing chemotherapy and/or immunotherapy in the advanced disease setting, per standard practice in oncology.

Prescribing Conditions

The OH-CCO clinicians noted that a specialist, ideally as part of a multidisciplinary team, is required in a hospital outpatient clinic to diagnose, treat, and monitor patients receiving nivolumab. The LCC clinicians also stated that, ideally, a multidisciplinary cancer tumour board, consisting of (nonexclusively) respirologists, radiologists, pathologists, thoracic surgeons, medical oncologists, and radiation oncologists, should discuss a multimodal treatment approach for patients who are usually referred to oncologic thoracic surgeons affiliated with major cancer centres. The LCC clinicians added that the delivery of care should be planned according to local structures; ideally, systemic therapies will be administered as close to a patient’s home as possible while patients are continuously monitored by the cancer centre to coordinate neoadjuvant therapies with posttreatment imaging, preoperative evaluations, and the surgical admission itself.

Table 5: Summary of Drug Plan Input and Clinical Expert Response

ALK = anaplastic lymphoma kinase; ECOG PS = Eastern Cooperative Oncology Group Performance Status; EGFR = epidermal growth factor receptor; NSCLC = non–small cell lung cancer PD-1 = programmed cell death 1 protein; PD-L1 = programmed cell death 1 ligand 1; pERC = CADTH pan-Canadian Oncology Drug Review Expert Review Committee.

Drug Program Input

The drug programs provide input on each drug being reviewed through CADTH’s reimbursement review processes by identifying issues that may affect their ability to implement a recommendation. The implementation questions and corresponding responses from the clinical experts consulted by CADTH for nivolumab are summarized in Table 5.

Clinical Evidence

The clinical evidence included in the review of nivolumab is presented in 3 sections. The first section, the Systematic Review, includes pivotal studies provided in the sponsor’s submission to CADTH and Health Canada, as well as studies that were selected according to an a priori protocol. The second section includes indirect evidence from the sponsor and indirect evidence selected from the literature that met the selection criteria specified in the review. The third section includes sponsor-submitted long-term extension studies and additional relevant studies that were considered to address important gaps in the evidence included in the Systematic Review.

Systematic Review (Pivotal and Protocol-Selected Studies)

Objectives

To perform a systematic review of the beneficial and harmful effects of nivolumab (360 mg administered intravenously), in combination with platinum-doublet chemotherapy every 3 weeks for up to 3 cycles, for the neoadjuvant treatment of adults with resectable NSCLC (tumours ≥ 4 cm or node-positive disease).

Methods

Studies selected for inclusion in the Systematic Review included pivotal studies provided in the sponsor’s submission to CADTH and Health Canada, as well as those meeting the selection criteria presented in Table 6. Outcomes included in the CADTH review protocol reflect outcomes considered to be important to patients, clinicians, and drug plans.

Two CADTH clinical reviewers independently selected studies for inclusion in the review based on titles and abstracts, according to the predetermined protocol. Full-text articles of all citations considered potentially relevant by at least 1 reviewer were acquired. Reviewers independently made the final selection of studies to be included in the review, and differences were resolved through discussion.

Table 6: Inclusion Criteria for the Systematic Review

E = adverse event; ALK = anaplastic lymphoma kinase; ECOG PS = Eastern Cooperative Group Performance Status; EFS = event-free survival; EGFR = epidermal growth factor receptor; HRQoL = health-related quality of life; MPR = major pathologic response; NSCLC = non–small cell lung cancer; OS = overall survival; pCR = pathologic complete response; PD-L1 = programmed cell death 1 ligand 1; RCT = randomized controlled trial; SAE = serious adverse event; WDAE = withdrawal due to adverse event; vs. = versus.

^aIdentified by clinical experts as a relevant comparator, but not currently publicly funded in Canada

The literature search for clinical studies was performed by an information specialist using a peer-reviewed search strategy according to the PRESS Peer Review of Electronic Search Strategies checklist.³²

Published literature was identified by searching the following bibliographic databases: MEDLINE All (1946–) via Ovid and Embase (1974– via Ovid. All Ovid searches were run simultaneously as a multifile search. Duplicates were removed using Ovid deduplication for multifile searches, followed by manual deduplication in Endnote. The search strategy comprised both controlled vocabulary, such as the National Library of Medicine’s MeSH (Medical Patient Headings) and keywords. The main search concepts were Opdivo (nivolumab) and non–small cell lung cancer. The following clinical trials registries were searched: the US National Institutes of Health’s clinicaltrials.gov, WHO’s International Clinical Trials Registry Platform (ICTRP) search portal, Health Canada’s Clinical Trials Database, and the European Union Clinical Trials Register.

CADTH-developed search filters were applied to limit retrieval to RCTS or controlled clinical trials. Retrieval was not limited by publication date or by language. Conference abstracts were excluded from the search results. Refer to Appendix 1 for the detailed search strategies.

The initial search was completed on September 29, 2022. Regular alerts updated the search until the meeting of the CADTH pan-Canadian Oncology Drug Review Expert Committee (pERC) on February 8, 2023.

Grey literature (literature that is not commercially published) was identified by searching relevant websites from the Grey Matters: A Practical Tool For Searching Health-Related Grey Literature checklist.³³ Included in this search were the websites of regulatory agencies (FDA and European Medicines Agency). Google was used to search for additional internet-based materials. Refer to Appendix 1 for more information on the grey literature search strategy.

These searches were supplemented by reviewing bibliographies of key papers and through contacts with appropriate experts. In addition, the manufacturer of the drug was contacted for information regarding unpublished studies.

Findings From the Literature

A total of 640 studies were identified from the literature for inclusion in the Systematic Review (Figure 1). The included studies are summarized in Table 7. A list of excluded studies is presented in Appendix 2.

Description of Study

CheckMate 816 is an ongoing, open-label, randomized, phase III trial assessing the efficacy and safety of nivolumab (3 mg/kg every 2 weeks for up to 3 cycles) in combination with ipilimumab (single 1 mg/kg dose), nivolumab (360 mg flat dose) in combination with platinum-based chemotherapy every 3 weeks for 3 cycles, and platinum-chemotherapy alone as a neoadjuvant treatment in adults 18 years and older with resectable (stage IB [≥ 4 cm], stage II, and resectable stage IIIA) NSCLC.¹⁵ Disease staging at screening was based on the AJCC/UICC TNM 7th edition.¹⁵ After completion of neoadjuvant treatment, all patients who remained operative candidates underwent definitive surgery for NSCLC within 6 weeks. Patients were also allowed to receive adjuvant chemotherapy with or without radiation after definitive surgery, per institutional standard, at the discretion of the investigator.

Figure 1: Flow Diagram for Inclusion and Exclusion of Studies

Table 7: Details of Included Study

AE = adverse event; ALK = anaplastic lymphoma kinase; AUC = area under the plasma drug concentration-time curve; BICR = blinded independent central review; BIPR = blinded independent pathologic review; CR = complete response; CTLA-4 = cytotoxic T-lymphocyte-associated protein-4; DLco = diffusing capacity; ECOG PS = Eastern Cooperative Group Performance Status; EFS = event-free survival; EFS2 = event-free survival on the next line of therapy; EGFR = epidermal growth factor receptor; EQ-5D-3L = 3-Level EQ-5D; FEV₁ = forced expiratory volume in 1 second; FRC = functional residual capacity; FVC = forced vital capacity; HRQoL = health-related quality of life; IA1 = first interim analysis; IHC = immunohistochemistry; MPR = major pathologic response; N = node; NSCLC = non–small cell lung cancer; OS = overall survival; pCR = pathologic complete response; PD-1 = programmed cell death 1 protein; PD-L1 = programmed cell death 1 ligand 1; RECIST = Response Evaluation Criteria in Solid Tumours Version 1.1; SAE = serious adverse event; TTDM = time to death or distant metastases; VAS = visual analogue scale.

Source: Clinical Study Report.¹⁵

Coprimary end points investigated in the trial were pCR assessed by BIPR and EFS assessed by BICR. Secondary end points included OS, TTDM, and MPR. Safety, tolerability, and HRQoL were exploratory outcomes.¹⁵ All BICR and BIPR reviews of radiology scans and tumour and/or lymph node tissue for progression and disease recurrence were completed by a third-party vendor using prespecified guidelines (RECIST 1.1). Any investigator-assessed radiographic progression per RECIST 1.1 guidelines to confirm progression or recurrence was confirmed by BICR, per RECIST 1.1 guidelines.

The CheckMate 816 trial was initially designed as a 2-arm trial in which patients were randomized in a 1:1 ratio to 1 of 2 trial arms: nivolumab plus ipilimumab or chemotherapy. An update to the protocol (protocol revision 2) introduced a third arm, nivolumab plus chemotherapy, thereby allowing patients to be subsequently randomized in a 1:1:1 scheme to any treatment arm. A third update to the protocol (protocol revision 3) discontinued the randomization of patients to the nivolumab plus ipilimumab arm. Subsequently, patients enrolled in the study were randomized to either nivolumab plus chemotherapy or to chemotherapy in a 1:1 ratio and stratified by 3 factors:

• PD-L1 expression level (≥ 1% versus < 1%, not evaluable, indeterminate)

• disease stage (stage IB/ II versus stage IIIA)

• gender or sex.

Patients who were already randomized to the nivolumab plus ipilimumab arm stayed in the study and continued scheduled study procedures. Study objectives in the CheckMate 816 trial were updated to focus on the comparison between nivolumab plus chemotherapy and chemotherapy for patients who were concurrently randomized. This CADTH review focuses on findings reported in the nivolumab plus chemotherapy and chemotherapy arms.

Patient enrolment began in March 2017 and the last patient was randomized on December 11, 2019. A total of 773 patients were enrolled at 111 sites in 14 countries (North America [including Canada], Europe, Asia, and the rest of the world). In total, 505 patients were randomized, 358 of whom were placed in the nivolumab plus chemotherapy arm (n = 179) or the chemotherapy arm (n = 179).¹⁵

The primary data cut-off date for the prespecified final pCR analysis was September 16, 2020, and the data cut-off date for IA1 EFS was October 20, 2021. By the October 20, 2021, data cut-off date, all treated patients had been off neoadjuvant study treatment for more than 18 months; 93.8% and 84.7% of treated patients in the nivolumab plus chemotherapy and chemotherapy arms, respectively, had completed the course of neoadjuvant therapy.¹⁵ Figure 2 presents the design flow of the CheckMate 816 trial.

At Screening

Baseline screening assessments were performed in the 28 days before randomization. PET-CT with contrast was assessed at baseline and a separate CT with contrast of the chest, abdomen, and other suspected areas (as well as the PET-CT) was required if the CT component of a PET-CT was not of sufficient diagnostic quality for RECIST 1.1 measurements. Patients with suspected brain metastases, including those with stage II disease or higher were evaluated with MRI/CT of the brain before and after contrast procedures. Tumour assessments were performed using the RECIST 1.1 guidelines.¹²

Figure 2: Study Design of the CheckMate 816 Trial

cRR = clinical response rate; EFS = event-free survival; EFS2 = event-free survival on the next line of therapy; HCP = health care provider; NSCLC = non–small cell lung cancer; OS = overall survival; pCR = pathologic complete response; PD = disease progression; PD-L1 = programmed cell death 1 ligand 1; ; PK = pharmacokinetic; PRO = patient-reported outcome; Q12W = every 12 weeks; Q2W = every 2 weeks; Q3W = every 3 weeks; Q6 mos = every 6 months; RT = radiotherapy; RT = radiotherapy; SOC = standard of care; TTDM = time to death or distant metastases; tx = treatment; WT = wild-type.

Source: Clinical Study Report.¹⁵

Postoperative Assessments

Preoperative PET-CT with contrast was obtained in the 14 days before surgery. Postoperative assessments with CT with contrast of the chest, including the adrenal glands, and CT or MRI of other additional suspected or known sites of disease were performed. The first tumour assessment took place 12 weeks (± 7 days) after definitive surgery, per RECIST 1.1, and then every 12 weeks (± 7 days) for 2 years (104 weeks), every 6 months (24 weeks ± 7 days) for 3 years, and every year (52 weeks ± 7 days) for 5 years or until disease recurrence or progression confirmed by BICR.

The same imaging methods used at baseline were recommended for postoperative procedures.¹²

First tumour assessments were performed 12 weeks (± 7 days) following tumour restaging in patients who did not receive definitive surgery and who had no tumour progression, as confirmed by BICR. Subsequent tumour assessments were performed for these patients using methods similar to those implemented for patients who had received definitive surgery and at the same frequency.

In the event that a planned initiation of subsequent anticancer therapy was within 12 weeks of restaging, tumour assessments were repeated before the administration of subsequent anticancer therapy.¹² Postoperative radiation was administered per institutional standard of care.

The survival follow-up phase began 3 months after postneoadjuvant follow-up visit 2. Participants were followed every 3 months for survival. Survival follow-up visits could be performed with phone contact or an office visit.

Investigators at all participating sites were trained to perform radiology imaging in accordance with the study’s image acquisition guidelines before scanning the first patient.¹²

Loss to Follow-Up

Loss to follow-up was adequately defined in the protocol and measures were taken to account for missing data based on prespecified criteria.¹⁵

Populations

Inclusion and Exclusion Criteria

Patients 18 years and older with histologically confirmed stage IB (≥ 4 cm), stage II, or stage IIIA NSCLC (per TNM 7th edition) with resectable disease were included in the study. Patients underwent PET-CT, including contrast scans from the base of the skull to the upper thighs, to confirm disease stage. Patients were also required to undergo a CT scan with contrast of the chest, abdomen, and other suspected areas of disease (as well as the PET-CT) if the CT component of the PET-CT was not of sufficient diagnostic quality for RECIST 1.1 measurements. All suspicious mediastinal lymph nodes, including those that were pathologically enlarged or fluorodeoxyglucose avid on PET-CT required further pathologic sampling before randomization. Brain MRI or a CT of the head was required for patients with stage II or stage III disease or any patient with suspicion of brain metastases. Patients were expected to have an ECOG PS of 0 or 1 and measurable disease according to RECIST 1.1.¹⁵

Patients who had received prior chemotherapy or any other cancer therapy for early-stage NSCLC, patients with distant active brain metastases, and patients with an active, known, or suspected autoimmune disease were excluded. Patients with locally advanced, unresectable or metastatic disease and patients with known EGFR mutations or ALK translocation were also excluded.¹⁵

Baseline Characteristics

Most patients randomized to the nivolumab plus chemotherapy and chemotherapy arms were male (71.5% and 70.9%, respectively) and about half were white (49.7% and 44.7%, respectively) and about half were Asian (48.0% and 52.0%, respectively). In total, 48.6% and 53.1% of patients in the nivolumab plus chemotherapy and chemotherapy arms, respectively had squamous tumour histology; 63.1% and 64.2%, respectively, had stage IIIA disease, and 89.4% and 88.3%, respectively, were current or former smokers. PD-L1 expression in the nivolumab plus chemotherapy and chemotherapy arms was less than 1% in 43.6% and 43.0% of patients, respectively, 1% or more in 49.7% and 49.7% of patients, respectively, 1% to 49% in 28.5% and 26.3% of patients, respectively, 50% or more in 21.2% and 23.5% of patients, respectively, and not evaluable in 6.7% and 7.3% of patients, respectively. The sponsor noted that there was a small discrepancy in disease stage between the case report form and Interactive Response Technology system, which affected 8.4% of patients in the nivolumab plus chemotherapy arm and 7.8% of patients in the chemotherapy arm.¹⁵

Interventions

Nivolumab Plus Chemotherapy (Arm C)

Nivolumab: Nivolumab 360 mg (flat dose), intravenously, every 3 weeks for up to 3 cycles.

Chemotherapy: Investigator choice of platinum-based doublet chemotherapy (intravenously).

• cisplatin (75 mg/m² on day 1 of a 3-week cycle for up to 3 cycles) and 1 of the following:

  • gemcitabine (1,000 mg/m² or 1,250 mg/m² [per local prescribing information] on days 1 and 8 of a 3-week cycle for up to 3 cycles) (squamous histology)

  • pemetrexed (500 mg/m² on day 1 of a 3-week cycle for up to 3 cycles) (nonsquamous histology).

• carboplatin (area under the plasma drug concentration-time curve 5 or 6 on day 1 of a 3-week cycle for up to 3 cycles) and the following:

  • paclitaxel (175 mg/m² or 200 mg/m2 on day 1 of a 3-week cycle for up to 3 cycles) (any histology).

Patients who were unable to tolerate cisplatin were properly documented. The investigator obtained approval from the medical monitor before using a carboplatin-containing regimen, except when the regimen was carboplatin plus paclitaxel.¹⁵

Chemotherapy (Arm B)

Investigator choice of platinum-based doublet chemotherapy (intravenously):

• cisplatin (75 mg/m² on day 1 of a 3-week cycle for up to 3 cycles) and 1 of the following:

  • gemcitabine (1,000 mg/m² or 1,250 mg/m² [per local prescribing information] on days 1 and 8 of a 3-week cycle for up to 3 cycles) (squamous histology)

  • pemetrexed (500 mg/m² on day 1 of a 3-week cycle for up to 3 cycles) (nonsquamous histology)

  • vinorelbine (25 mg/m² or 30 mg/m² [per local prescribing information] on days 1 and 8 of a 3-week cycle for up to 3 cycles)

  • docetaxel (60 mg/m² or 75 mg/m² [per local prescribing information] on day 1 of a 3-week cycle for up to 3 cycles)

• carboplatin (area under the plasma drug concentration-time curve 5 or 6 on day 1 of a 3-week cycle for up to 3 cycles) and the following:

  • paclitaxel (175 mg/m² or 200 mg/m² on day 1 of a 3-week cycle for up to 3 cycles).

For patients who were unable to tolerate cisplatin, the reasons were documented. The investigator obtained approval from the medical monitor before using a carboplatin-containing regimen, except when the regimen was carboplatin plus paclitaxel.¹⁵

Note that 2 chemotherapy treatments allowed in the chemotherapy arm (cisplatin plus docetaxel and cisplatin plus vinorelbine) were not allowed in the nivolumab plus chemotherapy arm because, at the time nivolumab plus chemotherapy was added to the protocol, safety data were not available for nivolumab in combination with these chemotherapy backbones. The remaining chemotherapy options were the same for the chemotherapy and nivolumab plus chemotherapy arms.¹⁵

Dose Modification: Dose reductions were permitted for platinum-doublet chemotherapy, according to the trial’s protocol. Any dose reductions implemented were permanent and were not re-escalated in subsequent cycles. Dose reductions for each drug in the chemotherapy regimen was independently adjusted in accordance with a prespecified schedule. Patients experiencing toxicity that required a dose reduction after 2 previous dose reductions for any 1 drug in the regimen were discontinued from the drug. Dose modifications were allowed for hematological toxicities due to platinum-doublet therapy, based on nadir blood counts (assessed per local standards) from the proceeding drug administration. Growth factors were permitted after cycle 1 to assist hematologic recovery. Dose reductions were not permitted for nivolumab or ipilimumab.¹⁵

Dose Delay: Dose delays were permitted for all treatments in the trial; any dose delay of more than 7 days was skipped, and treatment resumed for the patient at the next scheduled dose (provided that criteria for re-treatment were met). Nivolumab and platinum-doublet chemotherapy were delayed in the event of AEs, and patients were re-evaluated weekly and treatment was resumed if they met re-treatment criteria.¹⁵

Treatment Discontinuation: Patients were allowed to discontinue treatment for the following reasons.

• At the request of the patient. Patients who chose to terminate treatments were allowed to remain in the study and receive protocol-specified follow-up procedures (unless the patient withdrew consent for further contact for all study procedures), including posttreatment study follow-up and any contact with persons previously authorized by the patient to provide follow-up information.

• In the event of any clinical AE, laboratory abnormality, or intercurrent illness experienced by the patient that, in the opinion of the investigator, indicated that continued participation in the study was not in their best interest.

• The study was terminated by the sponsor.

• Loss of the patient’s ability to freely provide consent because of imprisonment or involuntarily incarceration for the treatment of either a psychiatric or physical (e.g., infectious disease) illness.

Treatment was also discontinued in the event that a patient became pregnant. Patients who discontinued their assigned study treatment prematurely were not allowed to receive subsequent neoadjuvant therapy and proceeded to receive surgery within the study’s indicated timelines.¹⁵

After definitive surgery, patients in each arm could receive up to 4 cycles of adjuvant chemotherapy, per institutional standard, at the discretion of the investigator. Investigators chose from the following treatments:

Treatment 1:

• vinorelbine 30 mg/m² IV push over 10 minutes on days 1 and 8

• cisplatin 75 mg/m² IV over 120 minutes on day 1, immediately following vinorelbine

Treatment 2:

• docetaxel 75 mg/m² IV over 60 minutes on day 1

• cisplatin 75 mg/m² IV over 120 minutes on day 1, immediately following docetaxel

Treatment 3 (for squamous histology):

• gemcitabine 1,250 mg/m² IV over 30 minutes on days 1 and 8

• cisplatin 75 mg/m² IV over 120 minutes on day 1, immediately following gemcitabine

Treatment 4 (nonsquamous histology only):

• pemetrexed 500 mg/m² IV over 10 minutes on day 1

• cisplatin 75 mg/m² IV over 120 minutes on day 1, immediately following pemetrexed

Concomitant Medications

• Permitted Medications

  • Topical, ocular, intra-articular, intranasal, and inhalational corticosteroids (with minimal systemic absorption) in the absence of active autoimmune disease were permitted. Adrenal replacement steroid doses greater than 10 mg daily prednisone were also permitted. A brief (less than 3 weeks) course of corticosteroids for prophylaxis (e.g., contrast dye allergy) or for treatment of nonautoimmune conditions (e.g., delayed-type hypersensitivity reaction caused by a contact allergen) was also permitted.¹⁵

• Prohibited Medications

  • Immunosuppressive drugs

  • Immunosuppressive doses of systemic corticosteroids (exception those highlighted in the protocol)

  • Any concurrent antineoplastic therapy (i.e., chemotherapy; hormonal therapy; immunotherapy; extensive, nonpalliative radiation therapy; or standard or investigational drugs for the treatment of NSCLC)

  • Discontinuation of herbal medications was encouraged before study enrolment

  • Live vaccines (e.g., yellow fever; measles, mumps, and rubella; nasal flu; chicken pox [varicella])

  • Concomitant administration of strong CYP3A4 inhibitors or inducers and docetaxel, vinorelbine, or paclitaxel

  • Concomitant administration of strong CYP2C8 inhibitors or inducers and paclitaxel

  • Concomitant administration of drugs that are also tubularly secreted (e.g., probenecid)

Outcomes

A list of efficacy end points identified in the CADTH review protocol that were assessed in the clinical trial included in this review is provided in Table 9 and subsequently summarized. A detailed discussion and critical appraisal of the outcome measures are provided in Appendix 4.

Table 9: Summary of Outcomes of Interest Identified in the CADTH Review Protocol

AE = adverse event; BICR = blinded independent central review; BIPR = blinded independent pathologic review; EFS = event-free survival; EFS2 = event-free survival on the next line of therapy; EQ-5D-3L = 3-Level EQ-5D; HRQoL = health-related quality of life; MPR = major pathologic response; OS = overall survival; pCR = pathologic complete response; RECIST 1.1 = Response Evaluation Criteria in Solid Tumours Version 1.1; SAE = serious adverse event; TTDM = time to death or distant metastasis’ VAS = visual analogue scale.

Assessment of Outcomes

Radiologic tumour assessments were reviewed by a BICR from a third-party radiology vendor. Both the investigator and independent radiologists assessed whether each patient met RECIST 1.1 for progression or disease recurrence. All investigator-assessed radiographic progressions and all recurrences were confirmed by BICR, per RECIST 1.1 guidelines. Participants whose disease progression or recurrence was not confirmed by central review continued tumour assessments (when clinically feasible) in accordance with protocol-specified schedules. Subsequent tumour assessments were submitted to a third-party radiology vendor for subsequent review and, if applicable, patients were discontinued from the study when both the investigator and independent radiologists assessed that the participant met RECIST 1.1 for progression or recurrence.

A BIPR was implemented for the central review of tumour scans, for tumour and/or lymph node samples, and for the confirmation of end points. Personnel at participating sites were trained before enrolment of the first study participant. Sections of tumours were used for central pathology review to assess pCR and MPR. Any tumour, tumour bed, or lymph node specimens reviewed locally were submitted for central pathology review.

All participants (in either treatment arm) who remained operative candidates after completion of neoadjuvant treatment (up to 3 cycles) underwent definitive surgery for NSCLC within 6 weeks.

Lymph node levels were sampled at the time of definitive surgery:

• right-sided tumour: levels 4, 7, 9R, 10R, and 11R

• left-sided tumour: levels 5 and/or 6; 7, 9L, 10L, and 11L.

Tumour and lymph node samples were collected from definitive surgical resection, and sampling of fresh tumour tissue in RNAlater was mandatory for biomarker studies (as applicable, depending on the size of the residual tumour) on the day of surgery. The remainder of the specimens were processed for histopathologic analysis, performed within 72 hours of the procedure. Sections of the fresh tumour samples collected were used to assess pCR and MPR.¹² Disease recurrence, change in tumour measurements, and tumour response were assessed by BICR, per RECIST 1.1. After definitive surgery, patients in each arm could receive adjuvant chemotherapy with or without radiation, per institutional standard, at the discretion of the investigator.

Health-Related Quality of Life

EQ-5D-3L is a generic preference-based HRQoL instrument that has been applied to a wide range of health conditions and treatments.^36,37 One part of the EQ-5D-3L is a descriptive system that classifies respondents (12 years and older) into 1 of 243 distinct health states. The descriptive system consists of 5 dimensions: mobility, self-care, usual activities, pain/discomfort, and anxiety/depression. Each dimension has 3 possible levels, with 1 indicating no problems, 2 indicating some problems, and 3 indicating extreme problems. Respondents are asked to choose 1 level that reflects their own health state for each of the 5 dimensions. A scoring function can be used to assign a value (i.e., EQ- 5D-3L index score) to each self-reported health state, based on a set of population-based preference weights.^36,37 The other part of EQ-5D-3L is a vertical, calibrated EQ VAS that has end points labelled 0, indicating the worst imaginable health state, and 100, indicating the best imaginable health state. A detailed discussion and critical appraisal of the EQ-5D-3L questionnaire are available in Appendix 3.

Participants were asked to complete the EQ-5D-3L before any clinical activities were performed during on-treatment clinic visits, at postneoadjuvant visits 1 and 2, and at designated visits or during phone calls during the survival follow-up phase. A standardized script was used to facilitate telephone administration of the EQ-5D-3L.

• EQ-5D-3L questionnaire completion rate (i.e., the proportion of questionnaires actually received out of the expected number of patients on treatment or in follow-up), was calculated and summarized for each assessment time point by treatment arm.

• The proportion of patients reporting problems for any of the 5 EQ-5D-3L dimensions at each assessment time point was stratified by the level of problem and by treatment group. Percentages were based on the number of patients assessed at the assessment time point.

• For EQ-5D-3L utility index and VAS scores, mean score and mean change from baseline at each assessment time point (baseline, week 4, week 7, postadjuvant visit 1, and postadjuvant visit 2) were stratified by treatment group using descriptive statistics (n, mean with standard deviation and 95% CI, median, first and third quartiles, minimum, maximum), and a line graph summarizing the mean changes from baseline was presented.

Harms Outcome

Safety parameters evaluated included frequency of death, SAEs, AEs leading to discontinuation or dose modification, overall AEs, AEs of special clinical interest that are potentially associated with the use of nivolumab or ipilimumab (i.e., select AEs, immune-mediated AEs), clinical laboratory assessments (hematology, serum chemistry, and liver and thyroid function tests), and vital sign measurements.¹²

Statistical Analysis

Sample-Size and Power Calculation: The sample size of the CheckMate 816 trial was calculated based on EFS (primary end point) and accounted for multiple primary end point comparisons (pCR and EFS), with an initial alpha allocation of 0.01 and 0.04, respectively.¹²

The overall type I error (alpha) was controlled using the following procedure for nivolumab plus chemotherapy versus chemotherapy (alpha and P values are 2-sided):

The primary end point of pCR rate was tested at a 1% alpha level, using a fallback method:

• if the pCR rate was not significant, the primary end point EFS was tested at 4%

• if the pCR rate was significant, the 1% alpha was reallocated to the EFS primary end point, which was tested at a 5% alpha level

• if EFS was significant, OS was tested hierarchically at the same level as EFS.

EFS and OS were tested at a planned interim analysis, and both have a planned final analysis. Stopping boundaries were calculated for each end point individually, according to the observed number of events, using the Lan-DeMets alpha spending function with O’Brien-Fleming boundaries corresponding to an overall alpha of 4% or 5%. Given that the EFS and OS end points were tested using a group sequential approach, an overall hierarchical testing approach was used in which each end point would have its own specific Lan-DeMets alpha spending function with O’Brien-Fleming boundaries.¹² The sponsor estimated that about 500 patients would be randomized in the study.¹² The randomization of about 350 patients was planned for the 2 arms (175 patients each in the nivolumab plus chemotherapy and chemotherapy arms).

Primary Outcome Analyses

Pathologic Complete Response

A formal analysis for pCR took place after 350 patients who had been randomized into the nivolumab plus chemotherapy and chemotherapy arms from the start of the 1:1:1 randomization and had undergone surgery (approximately 30 months after the 1:1:1 randomization was initiated).¹²

The sponsor assumed that 350 patients would be needed to provide more than 90% power to detect an odds ratio of 3.857 with a 2-sided type I error rate of 1% if a pCR rate of 10% were to occur in the chemotherapy arm (arm B) and a pCR rate of 30% were to occur in the nivolumab plus chemotherapy arm (arm C).¹²

The pCR rate was computed for each treatment arm, along with the exact 95% CI, using the Clopper-Pearson method. The numerator was the number of randomized patients who achieved a pCR in both the tumour and the lymph nodes, as assessed by BIPR. The denominator was the number of all patients randomized to the nivolumab plus chemotherapy and chemotherapy arms. Patients who were no longer eligible for surgery, who were on an alternative anticancer therapy before surgery, who discontinued before surgery, or who did not have pCR results available were counted as nonresponders, ensuring a robust pCR definition. pCR was compared in the nivolumab plus chemotherapy and chemotherapy arms using the stratified Cochran Mantel-Haenszel test with a 2-sided 1% alpha level.¹²

Event-Free Survival

A formal EFS analysis took place after 148 events had occurred in the nivolumab plus chemotherapy arm and the chemotherapy arm after the start of the 1:1:1 randomization scheme (54 months after the 1:1:1 randomization scheme was initiated). The superiority hypothesis was tested in the EFS interim analysis.

In the event that pCR was significant, the 1% alpha allocated to pCR was reallocated to EFS, which was then tested at an overall 5% alpha level. The interim boundary was derived from the Lan-DeMets alpha spending function with O’Brien-Fleming boundaries.¹²

A stratified log-rank test, with stratification factors per Interactive Response Technology (PD-L1 expression [≥ 1% versus < 1%, not evaluable, indeterminate], disease stage [stage IB or II versus stage IIIA], and gender or sex), and a 2-sided P value was used to assess EFS between the nivolumab plus chemotherapy arms and chemotherapy. A Lan-DeMets alpha spending function with an O’Brien and Fleming type of boundary was used to determine the nominal significance levels for the interim and final analyses. The HR and the corresponding (adjusted alpha) CIs were estimated for the nivolumab plus chemotherapy and chemotherapy arms using a stratified Cox proportional hazards model with the randomized arm as a single covariate.¹²

The EFS curves for each randomized arm were estimated using the KM product-limit method. The median and 2-sided 95% CI for EFS in each treatment arm was computed using the log-log transformation method. In addition, EFS rates at different time points were derived from KM estimates on the EFS curve for each randomized arm. Associated 2-sided 95% CIs were calculated using the Greenwood formula (using a log-log transformation).¹²

The sponsor assumed that the occurrence of 185 EFS events among the 358 randomized patients would ensure that an overall 2-sided 5% significance level sequential test procedure with 2 interim analyses — after 148 events (80% of events required for final analysis) and after 167 events (90% of events required for final analysis) — would have 82% power, assuming an HR of 0.65 between the 2 arms. IA1 EFS was event driven (at least 148 events expected by data cut-off date of October 20, 2021). The stopping boundaries at the interim and final EFS analyses were derived from the exact number of events using the Lan-DeMets alpha spending function with O’Brien-Fleming boundaries.¹² Figure 3 presents the schematic presentation of the planned analyses in the CheckMate 816 trial. Table 10 presents a summary of scheduled analyses in the CheckMate 816 trial.

Table 10: Scheduled Analyses, Criteria, and Projected Timelines for Formal Analyses of pCR and EFS

EFS = event-free survival; FPFV = first patient’s first visit; OS = overall survival; pCR = pathologic complete response.

^aIf EFS reaches significance at the interim analysis of EFS but not OS, only the interim analysis of OS will be conducted; EFS will not be formally re-tested.

^bIf EFS reaches significance at the interim analysis of EFS but not OS, interim OS analyses will be triggered by the number of OS events.

Source: Sponsor’s submission.¹²

Figure 3: Schematic Representation of the Planned Analyses Time Points

EFS = event-free survival; FA = final analysis; IA1 = first interim analysis; IA2 = second interim analysis; IA3 = third interim analysis; mo = months; OS = overall survival; pCR = pathologic complete response; t = time point.

Note: EFS and OS were tested using their own O’Brien-Fleming alpha spending functions.

* Analysis occurred at the September 16, 2020, data cut-off date.

** Analysis occurred at the October 20, 2021, data cut-off date.

Source: Sponsor submission.¹²

Censoring Rules for EFS (Primary Definition)

• Patients who did not report progression or recurrence of disease or death will be censored on the date of their last evaluable tumour assessment.

• Patients who did not have any on-study tumour assessment and did not die will be censored on the date they were randomized.

• Patients who received subsequent anticancer therapy, outside of the protocol-specified adjuvant therapy, before documented progression or recurrence or death will be censored at the date of the last evaluable tumour assessment conducted on or before the date of initiation of the subsequent anticancer therapy.

• Patients who did not have a documented progression or recurrence or death and who received subsequent anticancer therapy outside of the protocol-specified adjuvant therapy will be censored at the date of the last evaluable tumour assessment conducted on or before the initiation of the subsequent anticancer therapy.

• Patients without a baseline scan and without surgery will be censored on the date of randomization (regardless of death).

• In the case of a new primary cancer, if such lesions are present on tumour assessment at the BICR, they will be considered new lesions, because the BICR does not have access to biopsy results.

Censoring Rules for EFS (Secondary Definition)

• Patients who did not report progression or recurrence of disease or death will be censored on the date of their last evaluable tumour assessment.

• Patients who did not have any on-study tumour assessments and did not die will be censored on the date they were randomized.

• Patients without a baseline scan and without surgery will be censored on the date of randomization (regardless of death).

Sensitivity Analyses

Sensitivity analyses were conducted in the nivolumab plus chemotherapy and chemotherapy arms for the primary end point of EFS. The P values from these sensitivity analyses were presented for descriptive purposes only and were not adjusted for multiplicity.

The following sensitivity analyses were considered:

• EFS was compared between treatment groups with a 2-sided MaxCombo test.

• To examine the assumption of proportional hazards in the Cox regression model, in addition to treatment, a time-dependent variable defined by a treatment by time interaction was added in the model.

• A multivariate Cox regression model was used to estimate the treatment effect after adjustment for possible imbalances in known or potential prognostic factors. The factors used in the randomization were included in the model as stratification factors. All additional factors were incorporated as covariates. The additional factors, which are all measured at baseline, included:

  • histology (squamous, nonsquamous)

  • age categorization (< 65 years, ≥ 65 years)

  • ECOG PS (0, ≥ 1)

  • race (white, Black, Asian, other).

Subgroup Analyses: Subgroup analyses were conducted for pCR and EFS for the following baseline characteristics¹²:

• age category (in years): < 65, ≥ 65 and < 75, ≥ 75 and < 85, ≥ 85, ≥ 75, ≥ 65

• sex (male, female), per Interactive Response Technology and per case report form

• race (white, Black, Asian, other)

• region (North America, Europe, Asia, the rest of the world)

• baseline ECOG PS (0, 1, > 1)

• tobacco use (current/former, never smoked, unknown)

• disease stage (IB or II versus stage IIIA) per Interactive Response Technology and per case report form

• baseline histology (squamous, nonsquamous)

• PD-L1 expression subgroups (< 1%, not evaluable, indeterminate; ≥ 1%; 1% to 49%; ≥ 50)

• tumour mutational burden of tissue evaluable (≥ 12.3 mut/MB, < 12.3 mut/MB, overall)

• tumour mutational burden of tissue not evaluable

• type of platinum therapy (cisplatin, carboplatin, switch from cisplatin to carboplatin)

• type of chemotherapy regimen in arm B (available in arm C are gemcitabine + cisplatin, pemetrexed + cisplatin, and paclitaxel + carboplatin; not available in arm C are vinorelbine + cisplatin and docetaxel + cisplatin), based on first neoadjuvant cycle.

Handling Missing Data

The sponsor outlined several methods to account for missing data for safety outcomes, missing death dates, missing data for date of progression or recurrence, missing dates for definitive surgery start date and definitive surgery end date, as follows:¹²

• For missing and partial AE onset dates, imputation was performed using the AE domain requirements specification.

• Missing and partial nonstudy medication domain dates were imputed using a derivation algorithm.

For death dates, the following conventions were used for imputing partial dates:

• If only the day of the month is missing, the first of the month was used to replace the missing day. The imputed date was compared with the last known date alive, and the maximum was considered as the death date.

• For a missing year, the death date was imputed as the last date the patient was known to be alive. If the date was completely missing but the reason for death was present, the death date was imputed as the last known date alive.¹²

Secondary Outcome Analyses

Overall Survival

The sponsor assumed that 185 OS events among the 358 randomized patients would ensure that an overall 2-sided 5% significance level sequential test procedure with 3 interim analyses — after approximately 101, 128, and 161 events (55%, 69%, and 87% of events required for final analysis, respectively) — would have 82% power, assuming an exponential distribution, with a median OS time of 54 months in the chemotherapy arm and 83 months in the nivolumab plus chemotherapy arm (corresponding to a target HR of 0.65). The sponsor anticipated that OS analyses would take place after about 48 months (IA1 EFS), 58 months (EFS interim analysis 2), 73 months (EFS final analysis), and 86 months (OS, final analysis) after the start of 1:1:1 randomization. The trigger for the timing of these interim analyses was the number of EFS events.¹²

In the event that EFS became significant before OS, the planned formal remaining OS analyses would be triggered by the number of OS events (approximately 128 OS events [69% information fraction] or 1 year after IA1 EFS, whichever occurred first for interim analysis 2, and 161 OS events or 4 years after last patient’s randomization, whichever occurred first for OS interim analysis 3).¹²

To account for a potential slowdown in event rates in the longer term, which could prevent OS analysis from being performed in a reasonable time window, the sponsor planned that in the event that the 185th event failed to occur 5 years after randomization of the last participant, the final OS analysis would be conducted at that time. The final analysis boundary would then be recalculated using the actual updated final number of events.¹²

Time to Death or Distant Metastases

TTDM, based on BICR assessments, was compared between the nivolumab plus chemotherapy and chemotherapy arms using the same methods as described for EFS.¹²

Major Pathologic Response

MPR rates were computed in each treatment group, along with the exact 95% CI, using Clopper-Pearson method. An estimate of the difference and the odds ratio in MPR rates between concurrent arms B and C and corresponding 95% CIs were calculated the using Cochran-Mantel-Haenszel methodology and adjusted by stratification factors.¹²

EFS on Next Line of Therapy

EFS2 was estimated using KM techniques, and graphically presented. A 2-sided 95% CI for the median in each treatment group were computed using a log-log transformation method. Events rates at fixed time points (e.g., 6 months, 12 months, depending on the minimum follow-up) were presented along with their associated 95% CIs. Estimates were derived from the KM estimate and corresponding CIs were based on the Greenwood formula for variance derivation and on log-log transformation applied to the survivor function.

Analysis Populations

Efficacy analysis population: The ITT population consisted of all randomized participants in the nivolumab plus chemotherapy and chemotherapy arms (which included all patients concurrently randomized to nivolumab plus chemotherapy or chemotherapy as of the 1:1:1 randomization and after revised protocol 2).

Safety analyses population: The primary analysis population consisted of all participants concurrently randomized to the nivolumab plus chemotherapy arm or chemotherapy arm at the 1:1:1 randomization who received at least 1 dose of any study medication in the neoadjuvant setting and was used to assess drug exposure and safety in the nivolumab plus chemotherapy and chemotherapy arms.

Protocol Amendments

Protocol revision 1 (March 3, 2017): This revision incorporated changes from amendment 2 and administrative letters 1 and 2. It clarified the use of TNM 7th edition in the study; adjusted dosing details of the chemotherapy drugs to include the dose approved by the local prescribing information and the standard-of-care infusion time for each country; clarified lymph node samples at screening and definitive surgery; clarified the tissue sample process for calculation of the primary end point; and clarified requirements for PET-CT scans and broadened the window for scans before surgery.¹⁵

Protocol revision 2 (July 6, 2017): This revision added the third arm of nivolumab plus chemotherapy; increased the sample size to 642 to accommodate the new treatment arm; changed the primary objective to dual primary objectives of pCR and EFS and changed a secondary objective to MPR based on health authority feedback; and increased the prescreening tissue requirement to 15 slides, updated contrast requirements for brain MRI scans, expanded the window for pulmonary function tests to within 6 weeks of randomization, and included updates to synopsis, rationale or background information, and study personnel.¹⁵

Protocol revision 3 (September 21, 2018): This revision stopped enrolment in the nivolumab plus ipilimumab arm, so the primary population concurrently randomized patients to the nivolumab plus chemotherapy and chemotherapy arms based on external clinical data on PD-1 expression and chemotherapy; clarified the definition of EFS, excluded patients with large-cell neuroendocrine carcinoma tumour histology, and added a platinum-doublet chemotherapy regimen (paclitaxel plus carboplatin); updated dose modification for docetaxel; added TTDM as a secondary end point; clarified tumour assessments for patients who did not proceed to definitive surgery; updated the statistical analysis plan, rationale, background information, and trial schematic; and clarified pulmonary function parameters, the time relationship between adjuvant radiotherapy and tumour imaging assessments, and the time window of cycle 1 day 1 end-of-infusion pharmacokinetic sampling.¹⁵

Protocol revision 4 (June 25, 2019): This revision updated the postsurgery collection of serum and plasma-soluble factors; added the concomitant administration of substances that are tubularly secreted (e.g., probenecid) and could potentially result in the delayed clearance of pemetrexed; added hypothesis testing for OS; and clarified the analysis population for pCR, added an exploratory end point of EFS on the next line of therapy, and added instructions for BICR.¹⁵

Protocol revision 5 (September 18, 2019): This revision made modifications to the pCR analysis population and projected timelines; updated the surgical-approach end point; updated censoring rules for TTDM; removed optional biopsy at disease progression in China; and updated the management algorithms to include myocarditis.¹⁵

Protocol revision 6 (July 14, 2020): This revision clarified an EFS event to be any progression that precluded surgery and clarified that RECIST 1.1 progression and/or recurrence, per BICR, should be applied after surgery or for patients who did not undergo surgery; corrected the number of patients; removed the first of 2 IA of EFS (60% events) and updated alpha spending based on the remaining single interim and final analyses of EFS; and clarified that the actual timing of analyses may differ from projected timing.¹⁵

Protocol revision 7 (August 18, 2021): This revision added an interim analysis of EFS and added a calendar-based rule for the final analysis of EFS, with a corresponding OS interim analysis (if EFS was significant).¹⁵

Results

Patient Disposition

Table 11 presents patient disposition data from the CheckMate 816 trial at the October 20, 2021, data cut-off date (EFS IA1). In total, 773 patients were enrolled, and 505 patients were randomized: 358 to nivolumab plus chemotherapy (n = 179) or chemotherapy (n = 179). After neoadjuvant treatment, 149 (83.2%) patients in the nivolumab plus chemotherapy arm and 135 (75.4%) in the chemotherapy arm underwent definitive surgery; 28 (15.6%) patients in the nivolumab plus chemotherapy arm and 37 (20.7%) in the chemotherapy arm did not undergo surgery for the following reasons:¹⁵

• AEs, which affected 2 of 28 (7.1% of cancellations) patients in the nivolumab plus chemotherapy arm and 1 of 37 (2.7%) patient in the chemotherapy arm

• disease progression, which affected 12 of 28 (42.9%) patients in the nivolumab plus chemotherapy arm and 17 of 37 (45.9%) patients in the chemotherapy arm

• other (including any reason other than an AE and clinical or radiographic disease progression [e.g., unsuitable for surgery per investigator, patient refused surgery, poor lung function], which affected 14 of 28 (50.0%) patients in the nivolumab plus chemotherapy arm and 19 of 37 (51.4%) patients in the chemotherapy arm. In addition, 2 (1.1%) patients in the nivolumab plus chemotherapy arm and 7 (3.9%) in the chemotherapy arm did not complete surgery because they left the study before surgery (no longer met the study criteria, withdrawal of consent, or death).

The proportion of patients who did not complete the neoadjuvant treatment period because of study drug toxicity was similar in the nivolumab plus chemotherapy and chemotherapy arms (5.7% and 6.8%, respectively). The proportion of patients who did not complete the neoadjuvant treatment period because of disease progression was low in the nivolumab plus chemotherapy and chemotherapy arms (0.6% and 1.1%, respectively). The minimum follow-up was 21.0 months, and the median follow-up was 29.5 months.¹⁵ Table 12 presents the patients who received definitive surgery in the CheckMate 816 trial.

Table 11: Patient Disposition

AE = adverse event.

^aPercentages based on patients randomized.

Source: Clinical Study Report.¹⁵

Table 12: Patients Who Received Definitive Surgery — All Randomized Patients

AE = adverse event; CR = complete response; SD = standard deviation.

Notes: Patients in arm B randomized in the initial protocol are included in the total.

The data cut-off date was October 20, 2021.

^aPatients with clinical downstaging have a lower disease stage before surgery than at baseline.

^bDenominator based on the number of patients with cancelled surgery.

^cDenominator based on the number of patients with surgery.

^dDenominator based on the number of patients with delayed surgery.

^eTime from last neoadjuvant dose to surgery > 6 weeks.

^fIncludes patients, regardless of treatment status, who received subsequent anticancer therapy (outside of protocol-specified adjuvant therapy) without a prior reported EFS event. Those patients were censored at the last tumour assessment before or on the start date of subsequent anticancer therapy.

^gRadiotherapy other than protocol-defined adjuvant radiotherapy.

Source: Clinical Study Report.¹⁵

Protocol Deviations

Important protocol deviations that were considered to potentially affect the interpretability of study results were reported in 2% of patients in the study (1.1% in the nivolumab plus chemotherapy arm and 2.2% in the chemotherapy arm). In total, 3 patients (1 in the nivolumab plus chemotherapy arm and 2 in the chemotherapy arm) received concurrent cancer therapy. Important protocol deviations in the 2 arms are presented in Table 44 in Appendix 3.

Exposure to Study Treatments

The proportion of patients who received all 3 doses of nivolumab or chemotherapy in the 2 treatment arms in the CheckMate 816 trial are presented in the Table 45 and Table 46 in Appendix 3, and are as follows:

• in the nivolumab plus chemotherapy arm, the proportion was 93.2% for neoadjuvant nivolumab, 70.6% for carboplatin, 84.6% for cisplatin, 85.7% for paclitaxel, and 92.8% for pemetrexed

• in the chemotherapy arm, the proportion was 66.7% for carboplatin, 79.7% for cisplatin, 89.7% for docetaxel, 81.8% for paclitaxel, and 84.1% for pemetrexed.

The proportion of all treated patients in the 2 treatment arms who received 90% or more of the planned dose intensity during the neoadjuvant period was as follows:

• in the nivolumab plus chemotherapy arm, the proportion was 87.5% for nivolumab, 60.8% for carboplatin, 72.8% for cisplatin, 46.2% for gemcitabine, 60.7% for paclitaxel, and 85.5% for pemetrexed

• in the chemotherapy arm, the proportion was 59.5% for carboplatin, 69.9% for cisplatin, 65.5% for docetaxel, 63.3% for gemcitabine, 81.8% for paclitaxel, 79.4% for pemetrexed, and 23.1% for vinorelbine.

Interruptions, Delays, Reductions, and Omissions of Neoadjuvant Study Therapy

Dose delays were permitted in all treated patients in the nivolumab plus chemotherapy and chemotherapy arms. Dose reductions were permitted for chemotherapy drugs only, and interruptions of infusions or infusion rate reductions were infrequent. There was a higher proportion of dose omissions for gemcitabine and vinorelbine, which are dosed twice per cycle, than for other chemotherapy drugs.¹⁵

Table 13: Interruptions, Delays, Reductions, and Omissions of Neoadjuvant Study Therapy — All Randomized Patients

NA = not applicable.

Source: Clinical Study Report.¹⁵

Concomitant Medications

In total, all patients in the nivolumab plus chemotherapy arm and 99.4% in the chemotherapy arm received concomitant medications during the trial.¹² Table 14 presents the anatomic class of concomitant therapies permitted in the CheckMate 816 trial.

Subsequent Anticancer Therapy

Table 15 presents subsequent anticancer therapies administered to patients in the CheckMate 816 trial. In total, 38 (21.2%) patients in the nivolumab plus chemotherapy arm and 78 (43.6%) in the chemotherapy arm received subsequent cancer therapy (local and/or systemic therapy) started on or after the first study drug dose was administered (i.e., started on or after the date of randomization, if not treated) and not part of the on-protocol adjuvant study therapy (systemic and radiotherapy).¹²

Table 14: Concomitant Medications Permitted — All Treated Patients

Source: Sponsor submission.¹²

Table 15: Subsequent Cancer Therapy — All Randomized Patients

ALK = anaplastic lymphoma kinase; CTLA-4 = cytotoxic T-lymphocyte-associated protein-4; EGFR = epidermal growth factor receptor; PD-1 = programmed cell death 1 protein; PD-L1 = programmed cell death 1 ligand 1; VEGFR = vascular endothelial growth factor receptors.

Notes: Patient may have received more than 1 type of subsequent therapy. Subsequent therapy was defined as therapy started on or after the first dosing date (randomization date if patient never treated) outside of the protocol-specified adjuvant therapy. The reported medication Avastin was coded as simvastatin by error in the locked database. It should have been bevacizumab.

The data cut-off date was October 20, 2021.

Source: Sponsor submission.¹²

Adjuvant Therapy

Optional adjuvant chemotherapy or radiotherapy was allowed following surgery per protocol (per the investigator’s judgment). Table 16 presents the number of patients who received adjuvant therapy in the CheckMate 816 trial.

Table 16: Adjuvant Treatment — All Treated Patients

Source: Clinical Study Report.¹⁵

Efficacy

Only efficacy outcomes and analyses of subgroups identified in the CADTH review protocol are reported here.

Overall Survival

OS was formally tested at the October 20, 2021, data cut-off date (IA1 EFS), based on the hierarchical statistical testing rules for EFS and OS outlined in the statistical analysis plan.¹⁵

Median OS was not reached in either the nivolumab plus chemotherapy arm or the chemotherapy arm. The HR for death was 0.57 (99.67% CI, 0.30 to 1.07). The P value for OS (P = 0.008) did not cross the prespecified statistical significance boundary (0.0033). A second OS test is planned after 128 OS events have occurred (or 1 year after the IA1 EFS data cut-off date).¹⁵ Figure 4 presents the KM curves of OS at the data cut-off date.

Figure 4: OS — All Randomized Patients

CI = confidence interval; mo = months; No. = number; NR = not reported; OS = overall survival.

Note: The data cut-off date was October 20, 2021.

Source: Forde et al. (2022).³⁴

Event-Free Survival

Table 17 presents EFS findings at the IA1 data cut-off date (October 20, 2021) and Figure 5 presents the KM curves for the nivolumab plus chemotherapy and chemotherapy arms.

In total, 64 of 179 (35.8%) EFS events occurred in the nivolumab plus chemotherapy arm, and the median EFS was 31.57 months (95% CI, 30.16 to not available months). In the chemotherapy arm, 87 of 179 (48.6%) EFS events occurred, and the median EFS was 20.80 months (95% CI, 14.03 to 26.71 months). The estimated HR was 0.63 (97.38% CI, 0.43 to 0.91), with a P value of 0.0052 based on stratified log-rank test, which was statistically significant according to a prespecified threshold of 0.0262.¹⁵

The Cox model analyses based on baseline stratification factors (PD-L1 expression level [≥ 1% versus < 1%], disease stage [IB to II versus stage IIIA], and sex) had an HR of 0.64 (95% CI, 0.46 to 0.89); the estimated HR based on the unstratified Cox model was 0.63 (95% CI, 0.45 to 0.87). Additional sensitivity analyses that accounted for missing tumour assessments before EFS events had an estimated an HR of 0.66 (95% CI, 0.47 to 0.92).^15.

Table 17: EFS per BICR — All Randomized Patients

BICR = blinded independent central review; CI = confidence interval; EFS = event-free survival; HR = hazard ratio; NA = not available.

Note: The data cut-off date was October 20, 2021.

^aBased on KM estimates.

^bHR of arm C to concurrent arm B from a Cox model stratified by PD-L1 expression level (≥ 1% vs. < 1%, not evaluable, indeterminate), disease stage (IB or II vs. stage IIIA), and sex (male vs. female) as entered into the Interactive Response Technology.

^cLog-rank test stratified by the same factors as in the Cox proportional hazards model. The P value threshold for statistical significance was 0.0262.

^dProgression not necessarily meeting RECIST 1.1.

^eProgression and/or recurrence, per RECIST 1.1.

Source: Clinical Study Report.¹⁵

Figure 5: EFS per BICR — All Randomized Patients

BICR = blinded independent central review; Chemo = chemotherapy; CI = confidence interval; Conc. = concurrent; EFS = event-free survival; HR = hazard ratio; KM = Kaplan-Meier; N.A. = not available; Nivo = nivolumab.

Note: The data cut-off date was October 20, 2021.

Source: Clinical Study Report.¹⁵

Subgroup Analyses

Although there were inconsistencies across identified subgroups in terms of benefit, overall, a positive EFS benefit was observed in the nivolumab plus chemotherapy arm compared with chemotherapy. Increasing benefit was reported for nivolumab plus chemotherapy over chemotherapy across the PD-L1 subgroup, with a higher magnitude of benefit reported in patients with a PD-L1 expression level of at least 50% (HR = 0.24; 95% CI, 0.10 to 0.61) than in those with a PD-L1 expression level of 1% to 49% (HR = 0.58; 95% CI, 0.30 to 1.12), of at least 1% (HR = 0.41; 95% CI, 0.24 to 0.70), and of more than 1% (HR = 0.85; 95% CI, 0.54 to 1.32). Similarly, benefit was observed in the nivolumab plus chemotherapy arm in the disease-stage subgroup, with a higher magnitude of benefit reported for patients with stage IIIA disease (HR = 0.54; 95% CI, 0.37 to 0.80) than for those with stage IB or II disease (HR = 0.87; 95% CI, 0.48 to 1.53). A benefit was also observed in the histology subgroup — the HR for patients with nonsquamous disease was 0.50 (95% CI, 0.32 to 0.79) — and in the performance-status subgroup, in which the magnitude of benefit was higher in patients with an ECOG PS of 0 (HR = 0.61, 95% CI, 0.41 to 0.91) than in those with an ECOG PS of 1 (HR = 0.71; 95% CI, 0.41 to 1.21).

A subgroup analysis conducted to evaluate EFS by pCR status showed an EFS HR for pCR versus no pCR of 0.13 (95% CI, 0.05 to 0.37) in the nivolumab plus chemotherapy arm. No results were presented for the chemotherapy arm because of the small number of patients who achieved a pCR.

Of note, multiplicity adjustments to account for a type I error were not conducted during the subgroup analyses; as such, the subgroup findings are considered exploratory. Table 47 in Appendix 4 presents EFS findings in the subgroups identified in the CADTH review protocol.

Pathologic Complete Response and Major Pathologic Response

pCR analyses were conducted at the September 16, 2020, data cut-off date. The pCR rate, per BIPR, was 24.0% (43 of 179 patients; 95% CI, 18.0% to 31.0%) in the nivolumab plus chemotherapy arm compared to 2.2% (4 of 179 patients; 95% CI, 0.6% to 5.6%) in the chemotherapy arm. The difference in MPR between the 2 arms was 27.9% (95% CI, 19.6% to 36.1%). Table 18 presents the pCR and MPR rates obtained at the September 8, 2020, data cut-off date.

The pCR rate in the sensitivity analysis was 30.5% (43 of 141 patients; 95% CI, 23.0% to 38.8%) in the nivolumab plus chemotherapy arm and 3.2% (4 of 126 patients; 95% CI, 0.9% to 7.9%) in the chemotherapy arm.

Table 18: pCR — All Randomized Patients

CI = confidence interval; MPR = major pathologic response; pCR = pathologic complete response.

Note: The data cut-off date was September 16, 2020.

^aCI based on the Clopper-Pearson method.

^bStrata-adjusted difference (arm C – concurrent arm B) based on the Cochran-Mantel-Haenszel method of weighting.

^cStratified by PD-L1 expression level (≥ 1% vs. < 1%, unevaluable, indeterminate), disease stage (stage IB to II vs. stage IIIA), and sex (male vs. female) as entered into the Interactive Response Technology.

Source: Clinical Study Report.¹⁵

Subgroup Analyses

The pCR rates observed in the CADTH specified-protocol subgroups favoured nivolumab plus chemotherapy over chemotherapy. A numerically higher difference between treatments was observed in the subgroup of patients with a PD-L1 expression level of at least 50% (difference = 40.0%; 95% CI, 21.7% to 55.9%) than in the other PD-L1 subgroups. Differences were also observed in the disease-stage subgroup, with a difference between treatments of 20.0% for stage IIIA disease (95% CI, 12.8% to 28.4%) and of 24.4% (95% CI, 11.6% to 36.6%) for stage IB to II disease; in the histology subgroup, with a difference between treatments of 2.1% (95% CI, 11.0% to 31.4%) for squamous cell disease and of 22.8% (95% CI, 14.2% to 32.4%) for nonsquamous cell disease; and in the performance-status subgroup, with a difference between treatments of 24.9% (16.7% to 33.4%) for an ECOG PS of 0 and of 15% (95% CI, 3.8% to 33.4%) for an ECOG PS of 1. Table 48 of Appendix 4 presents the pCR obtained in subgroups at the September 8, 2020, data cut-off date.

Time to Death or Distance Metastasis

The median TTDM was not reached in the nivolumab plus chemotherapy arm at the October 28, 2021, data cut-off date. The estimated HR was 0.53 (95% CI, 0.36 to 0.77), as presented in Table 19.

Table 19: TTDM — All Randomized Patients

CI = confidence interval; HR = hazard ratio; NA = not available; TTDM = time to death or distance metastasis.

Note: The data cut-off date was October 20, 2021.

^aBased on KM estimates.

^bHR for arm C to concurrent arm B from a Cox model stratified by PD-L1 expression level (≥ 1% vs. < 1%, not evaluable, indeterminate), disease stage (IB or II vs. stage IIIA), and sex (male vs. female) as entered into the Interactive Response Technology.

Source: Clinical Study Report.¹⁵

Health-Related Quality of Life

Patients completed the EQ-5D-3L questionnaire at baseline, before on-treatment clinic visits, at postneoadjuvant visits 1 and 2, and at designated time points during the survival follow-up phase. Completion rates of questionnaires were reported for both study arms and are presented in Table 49 in Appendix 4.

EQ-5D index scores (based on the UK time trade-off value set) were collected for both study arms. Mean change from baseline in EQ-5D VAS and EQ-5D index during the neoadjuvant treatment phase were minimal and similar to baseline levels in the 2 treatment arms (Table 20 and Table 21). Figure 6 and Figure 7 presents mean changes in EQ-5D-3L utility index scores and VAS from baseline.¹²

Table 21: Summary of Overall Self-Rated Health Status on EQ VAS— All Randomized Patients

CI = confidence interval; NA = not available; Q1 = first quartile; Q3 = third quartile; SD = standard deviation.

Note: The data cut-off date was October 20, 2021.

Source: Sponsor submission.¹²

Table 22: EFS2 — All Randomized Patients

CI = confidence interval; EFS2 = event-free survival on the next- line of therapy; HR = hazard ratio; NA = not available.

^aBased on KM estimates.

^bHR of arm C to concurrent arm B from a Cox Model stratified by: PD-L1 expression level (≥ 1% vs. < 1%, not evaluable, indeterminate), disease stage (IB or II vs. stage IIIA), and sex (male vs. female) as entered into the Interactive Response Technology.

Source: Clinical Study Report.¹⁵

Figure 6: Mean Change in EQ-5D-3L Utility Index Score from Baseline — All Randomized Patients

EQ-5D-3L = 3-Level EQ-5D; SD = standard deviation.

Note: The data cut-off date was October 20, 2021.

Source: Clinical Study Report.¹⁵

Figure 7: Mean Changes in Overall Self-Rated Health Status on EQ VAS From Baseline — All Randomized Patients

EQ VAS = EQ visual analogue scale.

Note: The data cut-off date was October 20, 2021.

Source: Clinical Study Report.¹⁵

Figure 8: EFS2 — All Randomized Patients

Chemo = chemotherapy; CI = confidence interval; Conc. = concurrent; EFS = event-free survival; EFS2 = event-free survival on the next line of therapy; KM = Kaplan-Meier; N.A. = not available; Nivo = nivolumab.

Note: The data cut-off date was October 20, 2021.

Source: Clinical Study Report.¹⁵

Event-Free Survival on the Next Line of Therapy

By the October 20, 2021, data cut-off date, median EFS2, per investigator, was not reached in either the nivolumab plus chemotherapy or chemotherapy arm. The estimated HR was 0.54 (95% CI, 0.37 to 0.80). Table 22 presents EFS2 data and Figure 8 presents the KM curves for the nivolumab plus chemotherapy and chemotherapy arms.

Harms

Only harms identified in the CADTH review protocol are reported here. Safety data were reported for all treated patients randomized into the nivolumab plus chemotherapy and chemotherapy arms by the October 20, 2021, data cut-off date.

Adverse Events

Overall, 163 (92.6%) patients treated in the nivolumab plus chemotherapy arm and 171 (97.2%) in the chemotherapy arm reported at least 1 AE in the CheckMate 816 trial. The most frequently reported AEs in the nivolumab plus chemotherapy arm were nausea (38.1%), constipation (33.5%), anemia (29.0%), decreased appetite (20.5%), fatigue (16.5%), and neutropenia (16.5%), and in the chemotherapy arm were nausea (44.9%), constipation (32.4%), anemia (26.7%), decreased appetite (23.3%), and decreased neutrophil count (21.0%). AEs of grade 3 or 4 were reported in 40.9% (n = 72) and 43.8% (n = 77) of patients, respectively, in the nivolumab plus chemotherapy and chemotherapy arms.¹⁵ Table 23 presents details of the most frequently reported AEs of grade 3 or 4 in the 2 study arms of the CheckMate 816 trial.

Serious Adverse Events

SAEs of any grade were reported in 30 (17.0%) patients in the nivolumab plus chemotherapy arm and 24 (13.6%) patients in the chemotherapy arm. SAEs of grades 3 or 4 were reported in 19 (10.8%) patients in the nivolumab plus chemotherapy arm and 17 (9.7%) patients in the chemotherapy arm. The most commonly reported SAEs in the nivolumab plus chemotherapy arm were vomiting (2.3%), pneumonia (2.3%), embolism (1.1%), and febrile neutropenia (1.1%), and in the chemotherapy arm were febrile neutropenia (2.8%), pneumonia (1.7%), neutropenia (1.1%), and diarrhea (1.1%).¹⁵

Dose Delay or Reduction Due to AEs

AEs leading to a dose delay or reduction were reported in 53 (30.1%) and 66 (37.5%) treated patients, respectively, in the nivolumab plus chemotherapy and chemotherapy arms.¹⁵ The most common AEs leading to a dose delay or reduction are presented in Table 23.

Discontinuations Due to AEs

AEs leading to the discontinuation of study treatments were reported in 18 (10.2%) patients in the nivolumab plus chemotherapy arm and 20 (11.4%) in the chemotherapy arm. Grade 3 or 4 AEs leading to the discontinuation of at least 1 study drug were reported in 10 (5.7%) patients in the nivolumab plus chemotherapy arm and 7 (4.0%) patients in the chemotherapy arm. The most frequently reported AEs of any grade leading to discontinuation in the nivolumab plus chemotherapy arm were anaphylactic reaction (1.7%), decreased neutrophil count (1.1%), and fatigue (1.1%), and in the chemotherapy arm were neutropenia (2.3%), decreased neutrophil count (1.1%), increased blood creatinine (1.1%), and pneumonia (1.1%).¹⁵

Mortality

By the October 20, 2021, data cut-off date, 35 (19.9%) patients had died in the nivolumab plus chemotherapy arm and 59 (33.5%) patients had died in the chemotherapy arm (Table 24). There were no deaths due to study drug toxicity (per investigator) in the nivolumab plus chemotherapy arm, but 3 deaths due to study drug toxicity (per investigator) in the chemotherapy arm (caused by pancytopenia, diarrhea, and acute kidney injury [all 3 reported in 1 patient]; enterocolitis infection; and lung infection and/or pneumonia). Among the 149 patients who underwent surgery after treatment with nivolumab plus chemotherapy, 5 (3.4%) died within 90 days of surgery, most commonly from intraoperative hemorrhage, esophageal perforation, pneumonia, cardiopulmonary arrest due to pulmonary embolism, or aortic rupture. Among the 135 patients who underwent surgery after chemotherapy, 2 (1.5%) patients died within 90 days from an AE or SAE associated with pneumonia.¹⁵

Notable Harms

AEs of special interest, specifically immune-related AEs, are presented in Table 23.

Table 23: Summary of Safety With Neoadjuvant Treatment — All Treated Patients

AE = adverse event; GI = gastrointestinal; imAE = immune-mediated adverse event; PT = preferred term, SAE = serious adverse event; WBC = white blood cell.

Note: The data cut-off date was October 20, 2021.

Source: Clinical Study Report.¹⁵

Table 24: Summary of Death — All Treated Patients

Notes: Only AEs that led to death within 24 hours were documented as grade 5. Events leading to death > 24 hours after onset are reported as the worst grade before death.

Data cut-off of October 20, 2021.

^aThe causes of death, per investigator, in the chemotherapy arm were as follows: pancytopenia, diarrhea, and acute kidney injury (all 3 reported in 1 patient); enterocolitis infection; and lung infection and/or pneumonia.

^bThe verbatim terms reported for other reasons for death and were consistent with events expected in the study population. None were considered to be related to study drug (per the investigator).

Source: Clinical Study Report.¹⁵

Critical Appraisal

Internal Validity

CheckMate 816 is an ongoing, randomized, open-label, phase III trial. Initially, patients were randomized using an Interactive Response Technology in a 1:1 ratio into 1 of 2 treatment arms, and treatment allocation was concealed. Baseline characteristics in the 2 arms of interest were balanced, suggesting that randomization was successful. The methods of randomization, treatment allocation, and stratification factors were considered appropriate. Disease stage and PD-L1 expression level were considered prognostic factors for NSCLC, and sex was considered a significant factor in the assessment of outcomes for immunotherapy.

The treatment effect for the 2 coprimary end points (EFS and pCR) was estimated during a prespecified interim analysis, adjusted using the Lan-DeMets alpha spending function with O’Brien-Fleming boundaries that accounted for the actual number of events at an overall alpha of 4% or 5%. Overall, the estimated treatment effect of the coprimary end points was based on interim analyses. There is uncertainty about the magnitude of the treatment effect, given that interim analyses have a tendency to overestimate treatment effect.^38-40

OS estimated by the HR for death was statistically nonsignificant, based on the preplanned interim analysis stopping rule (P = 0.008 against a prespecified level of significance at the interim analysis of 0.0033). Although the results showed a promising trend toward a significant treatment effect for OS, the final analysis may be needed to confirm the findings, particularly the exact estimate of the difference in median survival, which was not estimable at the date cut-off date.

Performance and assessment biases due to the open-label design of the trial were considered unlikely, given that radiologic assessments of CT scans for EFS and pathologic review of tumour sections were completed by a blinded independent review team, based on prespecified and validated RECIST 1.1 guidelines. Further, an independent pathology review team assessed images and tumour or lymph node samples for pCR and MPR, which minimizes performance bias.

The proportion of patients exposed to 3 doses of chemotherapy drugs was slightly higher in the nivolumab plus chemotherapy arm than in the chemotherapy arm, which could bias the findings in favour of nivolumab plus chemotherapy (70.6% versus 66.7% for carboplatin; 84.6% versus 79.7% for cisplatin; 85.7% versus 81.8% for paclitaxel; 92.8% versus 84.1% for pemetrexed; and 72.3% versus 75.5% for gemcitabine). In addition, patients in the nivolumab plus chemotherapy arm did not receive docetaxel or vinorelbine because the safety profile of these drugs used in combination with nivolumab had not been established before the establishment of the study arm. The impact of docetaxel and vinorelbine in the estimated effect between treatment arms was considered likely to be negligible. There were also slight differences in the cumulative dose intensity (which also accounted for dose delays and reductions during the trial) in the nivolumab plus chemotherapy and chemotherapy arms for some chemotherapy drugs (46.2% versus 63.3% for gemcitabine [important]; 60.7% versus 81.8% for paclitaxel; and 85.5% versus 79.4% for pemetrexed), which may introduce bias in either direction in the nivolumab plus chemotherapy arm (it may have affected the proportion of patients that became eligible for surgery).

More patients in the chemotherapy arm than in the nivolumab plus chemotherapy arm received any subsequent therapy (n = 78 versus 38 [43.6% versus 21.2%]), received subsequent systemic anticancer therapy (n = 65 versus 31 [36.3% versus 17.3%]), and received subsequent immunotherapy (n = 42 versus 10 [23.5% versus 5.6%]). The use of subsequent anticancer therapy may bias EFS and OS outcomes in the 2 groups. Further, the proportion of patients who received subsequent adjuvant therapy (optional) after surgery was higher in the chemotherapy arm than in the nivolumab plus chemotherapy arm (n = 56 versus 35 [31.8% versus 19.9%]). The use of subsequent adjuvant therapy may bias EFS and OS findings in favour of the chemotherapy arm, which may have led to the improved OS and EFS reflected in the KM curves estimates. The sponsor-presented analyses that did not apply censoring at subsequent anticancer therapy use showed an HR of 0.63 (97.38% CI, 0.44 to 0.89), consistent with the primary analysis (HR = 0.63; 97.38% CI, 0.44 to 0.89; P = 0.0027 [descriptive]). The sponsor also performed additional analyses that adjusted for the use of adjuvant chemotherapy (as a time-dependent covariate), which favoured nivolumab plus chemotherapy over chemotherapy alone (adjusted HR = 0.65; 95% CI, 0.47 to 0.90). There is a potential risk of bias from the use of systemic anticancer therapy and adjuvant chemotherapy; however, the bias was in the outcomes was considered low.

The methods used to address missing data in the trial were considered conservative and may have had an impact on the assessment of efficacy outcomes. The extent of bias may favour of nivolumab plus chemotherapy over chemotherapy for OS and EFS outcomes, given that there were more patients in the chemotherapy arm who did not undergo definitive surgery.

All efficacy end points were adequately described in the sponsor-submitted statistical analysis plan, and censoring rules, hierarchical hypothesis testing, and stopping rules were provided for the primary outcomes, including OS, investigated in the CheckMate 816 trial. Crossing over from 1 study arm to the other was not permitted during the trial, which preserved the treatment differences observed in the randomized arms for EFS and OS. A Cox proportional hazards model, which relies on the assumption of proportional hazards in both treatment groups, was used to assess OS and EFS survival curves. Visual assessment of the KM curve for EFS did not suggest a violation of the proportional hazards assumption, as the curves did not cross; however, visual assessment of the OS curves did suggest a violation of the proportional hazards assumption. In addition, no assessment of the proportional hazards assumption was provided for EFS and OS; therefore, it is uncertain whether the sponsor addressed potential violations in their analyses.

The EQ-5D-3L questionnaire is a generic instrument that has been used to measure patient-reported outcomes and treatments across different health conditions, including cancers.^36,37 However, there is limited evidence available on its validity, reliability, and responsiveness to change in patients with NSCLC. In general, the estimated MID for EQ-5D-3L index scores for patients with lung cancer was 0.06 in the US and 0.08 in the UK, using an anchor-based method.⁴¹ The MID for EQ VAS scores identified in literature was 7 to 12, and was estimated using the ECOG PS method.⁴¹ Utility index and VAS scores obtained from the EQ-5D-3L questionnaire at different time points (baseline, week 4, week 7, postadjuvant visit 1, and postadjuvant visit 2) were used to assess the general health status of patients enrolled in the trial. Missing data for HRQoL were considered missing at specific time points in the analysis. Multiplicity adjustments were not conducted to account for the type I error rate, so the findings were considered exploratory, and no conclusions could be drawn.

Several changes were made to the study protocol, which were adequately reported. These frequent amendments occurred after the study was initiated to reflect the high degree of uncertainty in the objectives when the study was designed. Sample-size calculations and other efficacy outcomes were updated in the statistical analysis plan to reflect the amendments.

All protocol deviations in the 2 study arms of the trial were adequately reported. Relevant protocol deviations were slightly higher in the chemotherapy arm than in the nivolumab plus chemotherapy arm (1.1% versus 2.2%). The most common protocol deviations were incorrect disease staging at baseline and on-treatment deviations related to concurrent cancer therapy.

There was a low risk of selective reporting, as all the outcomes prespecified in the protocol were reported. Interim analysis for pCR and EFS on the first line of therapy were prespecified in the protocol, including the trigger factors and stopping rules. The approaches used to preserve the alpha and the power in the interim analyses were outlined adequately in the statistical analysis plan and were considered to be appropriate. Efficacy analyses were conducted using the ITT population (patients randomized in the 1:1:1 randomization scheme in revised protocol 2 and in the 1:1 randomization scheme in revised protocol 3 [i.e., patients concurrently randomized in the nivolumab plus chemotherapy and chemotherapy arm]), which maintained randomization and minimized the risk of bias that could be introduced when comparing groups that differ in prognostic factors.

External Validity

The sponsor’s reimbursement request aligns with the Health Canada indication: in combination with platinum-doublet chemotherapy for the neoadjuvant treatment of adults with resectable NSCLC (tumours ≥ 4 cm or node-positive disease). The CheckMate 816 study only included patients with an ECOG PS of 0 or 1. The magnitude of benefit of nivolumab plus chemotherapy in patients with an ECOG PS of 2 or higher is uncertain. The experts highlighted the need for patients to have a robust performance status before receiving nivolumab plus chemotherapy to mitigate the risk of treatment-related toxicity, which can occur during treatment and preclude patients from receiving surgery. Thus, patients with an ECOG PS of 0 or 1 are more likely to receive the treatment than those with an ECOG PS of 2 and higher. Further, they are less likely to have other comorbidities unrelated to the cancer that would place them at higher risk for AEs.

There were differences identified in baseline characteristics between the CheckMate 816 trial population and the patient population in Canada. The mean age of patients enrolled in the CheckMate 816 trial was younger (by 10 years) than in the NSCLC population in Canada. Further, more patients with stage IIIA disease than with stage IB or stage II disease were enrolled in the trial, which, according to the experts, is uncommon in Canadian practice. As highlighted by the experts, patients eligible for surgical resection in practice are at an earlier disease stage than those eligible in the trial (the majority of patients enrolled in the trial had stage IIIA disease). The study included only patients with an ECOG PS of 0 or 1 but, according to the experts, more patients with an ECOG PS of 2 or 3 are seen in current practice than in the trial. It was also noted by the experts that a higher proportion of patients with a PD-L1 expression level of at least 50% would be expected in practice than in the CheckMate 816 trial (in practice, the proportion of patients with a PD-L1 expression level of less than 1%, 1% to 49%, and 50% or more, according to the clinical experts, is about one-third for each category. These differences were considered unlikely to affect the generalizability of the CheckMate 816 trial findings to the Canadian population, according to the clinical experts.

The dosing of nivolumab in the reimbursement request aligns with the Health Canada indication.

Dose adjustments were allowed for chemotherapy drugs but not for nivolumab in the trial, which aligns with the Health Canada indication. The experts indicated that the flat-dose approach for nivolumab, as implemented in the CheckMate 816 trial, would be used in practice because of the convenience to patients and pharmacists, and to reduce drug wastage.

The clinical experts indicated that the current standard of care in Canada for resectable NSCLC is upfront surgery with curative intent, followed by adjuvant chemotherapy. Therefore, there is uncertainty in the magnitude of benefit derived from using nivolumab plus chemotherapy in the neoadjuvant setting for patients with resectable NSCLC in the absence of a comparator arm that included surgery followed by adjuvant chemotherapy or comparative studies that evaluate the relative effects of the treatment against surgery followed by adjuvant chemotherapy. Regardless, the clinical experts noted that the magnitude of benefit in EFS for nivolumab plus chemotherapy, as estimated by the postsurgery HR, was clinically meaningful compared to current outcomes observed in patients who receive upfront surgery in current practice.

The clinical experts agreed that the chemotherapy treatments and dosing schedules implemented in the CheckMate 816 trial align with treatments used in the adjuvant setting for NSCLC in Canada. Dose delays and dose discontinuations are anticipated for chemotherapy treatments, according to the experts; however, the experts mentioned that no dose adjustments would be performed for nivolumab, and that there is currently no evidence to support the use of nivolumab in the adjuvant setting for patients with resected NSCLC.

Of note, the specific chemotherapy drugs to be used in combination with nivolumab were not specified in the indication. It is uncertain whether the addition of nivolumab to chemotherapy drugs not included in the nivolumab plus chemotherapy arm of the trial (e.g., docetaxel and vinorelbine) or in the Canadian product monograph will result in benefits similar to those seen with the chemotherapy drugs used in the CheckMate 816 trial. There is therefore uncertainty about the generalizability of the findings to real-world practice with respect to the use of chemotherapies not included in the trial.

Efficacy outcomes investigated in the CheckMate 816 trial were considered to be appropriate by the clinical experts and were considered to be important to patients and clinicians and reflective of outcomes assessed in real-world practice. There is limited evidence on the correlation of EFS as a surrogate outcome for OS with its validation for use in the neoadjuvant setting for NSCLC by regulatory agencies. The clinical experts indicated that EFS was presumably used in the CheckMate 816 trial to incorporate progression that precluded surgery and surgical mortality following neoadjuvant treatment as meaningful end points. pCR has not been validated as a surrogate outcome for OS by Health Canada. The clinical experts cited studies that have reported a correlation between pCR and MPR with OS in neoadjuvant trials for NSCLC following the use of chemotherapy; however, there is limited literature on the correlation of pCR and MPR in trials with targeted therapy or immunotherapy.¹¹

There were more follow-up procedures implemented in the CheckMate 816 trial than would be seen in real-world practice. Patients underwent more frequent assessments in the CheckMate 816 trial than are likely reflective of real-world practice. The clinical experts noted that patients in Canada are followed according to thoracic guidelines (i.e., every 6 months in the first year and annually thereafter, which may differ by practice and may depend on how late patients are in their follow-up schedule). The diagnostic methods and efficacy measurements implemented in the CheckMate 816 trial were considered appropriate by the experts. The assessment methods implemented in the trial were considered to be more reflective of academic than practice settings by the clinical experts, which may not be reflective of current Canadian practice guidelines, as patients would not be excluded from treatment if some of the procedures applied in the trial were not conducted. The clinical experts did not identify any major deviations in the diagnostic methods or assessment procedures that might have had an impact on eligibility criteria or that might have had an influence on the generalizability of the trial findings to current patient population in Canada.

Concomitant medications administered to patients in the trial were considered appropriate by the clinical experts. No significant discrepancies were identified that would have an impact on the trial outcomes or that would deviate from medications administered in real-world practice. Subsequent anticancer therapies administered during the trial were also considered appropriate.

Indirect Evidence

Objectives and Methods for the Summary of Indirect Evidence

The objective of this section is to summarize and critically appraise available indirect evidence comparing nivolumab plus platinum-doublet chemotherapy with other relevant treatments in patients with resectable, nonmetastatic NSCLC.

A focused literature search for ITCs dealing with NSCLC was run in MEDLINE All (1946–) on September 29, 2022. No search limits were applied. Titles, abstracts, and full-text articles were screened for inclusion based on the population, intervention, comparator, and outcome criteria outlined in the protocol for the CADTH review. No eligible ITCs were identified from the CADTH literature search.

Description of Indirect Comparison

The sponsor provided an NMA¹⁶ that assessed the efficacy and safety of neoadjuvant nivolumab plus chemotherapy relative to other relevant treatments, including neoadjuvant chemotherapy, neoadjuvant chemoradiotherapy, adjuvant chemotherapy, and surgery alone in patients diagnosed with resectable, nonmetastatic NSCLC.

Methods of Sponsor-Submitted ITC

Objectives

The objective of the sponsor-submitted ITC was to evaluate the efficacy and safety of neoadjuvant nivolumab plus chemotherapy relative to other relevant treatments, including neoadjuvant chemotherapy, neoadjuvant chemoradiotherapy, adjuvant chemotherapy, and surgery alone in patients with resectable, nonmetastatic NSCLC.

Study-Selection Methods

Based on the prespecified eligibility criteria outlined in Table 25, the sponsor conducted an SLR⁴² to identify studies investigating the efficacy and safety of current treatments for patients diagnosed with resectable, nonmetastatic NSCLC. The literature searches, last updated on April 1, 2022, were conducted in MEDLINE, Embase, and the Cochrane Central Register of Controlled Trials. The study screening and selection process was conducted by 2 independent reviewers, and disagreement was resolved by consultation with a third reviewer. Data were extracted by 1 reviewer and verified by a second reviewer. Risk-of-bias assessment was carried out on the recommendations from the National Institute for Health and Care Excellence (NICE) Single Technology Appraisal and Highly Specialised Technologies Evaluation: User Guide for Company Evidence Submission Template.⁴³ From the evidence base created from the SLR, the sponsor identified studies that were eligible for an ITC based on a set of additional criteria, listed in Table 25.

ITC Analysis Methods

An NMA using the Bayesian approach was conducted, and reporting of the NMA generally followed guidance for the PRISMA (Table 26).

For each outcome examined in the NMA, random-effects and fixed-effects models were run. The random-effects model was considered by the sponsor to be the default model. However, the fixed-effects model was selected by the sponsor for all analyses in the NMA because of the sparseness of the network, which was not able to estimate the between-study SD with enough precision. Deviance information criterion was also used to determine model fit, with a lower deviance information criterion indicating a relative improvement in model fit. The sponsor was not able to assess consistency between the direct and indirect evidence because of a lack of closed loops in the network. Convergence was assessed by the monitoring of plots.

The primary outcomes of interest were EFS and OS. Secondary outcomes of interest included TTLRR, TTDM, pCR, and safety outcomes. Of note, the OS data from the sponsor-conducted CheckMate 816 trial was based on an immature data cut-off date.

In addition to the stage-agnostic network, which included studies regardless of the staging of the patient population, the network for each outcome was stratified by tumour staging (i.e., stage IB or II, stage IIIA, and stage IIIA N2). The base-case analysis for each outcome in the NMA synthesized evidence from patients who were deemed candidates for surgery and received third-generation platinum-based doublet chemotherapy. Sensitivity analyses were also carried out for each outcome whenever data were available. These sensitivity analyses involved second-generation platinum-based chemotherapies, the population of resected patients, data stratified by PD-L1 expression level (i.e., ≥ 1% versus < 1%), and data from the ITT population of the CheckMate 816 trial instead the subpopulations of the CheckMate 816 trial (i.e., stage IB or II, stage IIIA, and stage IIIA N2) used in the base-case stage-specific networks.

Results of Sponsor-Submitted NMA

Summary of Included Studies

Of the 59 RCTs identified from the SLR,⁴² 23 met the inclusion criteria (Table 25) for the sponsor-submitted NMA.¹⁶ Of those, 8 were included in the base-case analyses, 5 additional RCTs were included in the sensitivity analyses involving second-generation chemotherapies, and 4 additional RCTs were included in the sensitivity analyses expanded to capture patients who had undergone resection. With respect to the 6 RCTs that met the NMA inclusion criteria but were not analyzed in the NMA, the sponsor claimed that 2 RCTs (i.e., 1 evaluating postadjuvant atezolizumab and 1 evaluating postadjuvant pembrolizumab) would be included in exploratory analyses in a separate report. However, the sponsor confirmed later that, to date, the exploratory analyses were not available. The remaining 4 RCTs were not involved in any analysis as they included patients who had undergone complete resection and investigated a second-generation chemotherapy or the combination of tegafur plus uracil, which was relevant only to the population in Japan.

Table 27, Table 28, and Table 29 present study characteristics, patient characteristics, and tumour staging information, respectively, for the RCTs included in the base-case analyses in the sponsor-submitted NMA. Of note, the inclusion and exclusion criteria for the patient populations in the 8 included RCTs were not explicitly stated in the NMA report. In addition, according to the clinical experts consulted by CADTH, the proportion of male participants in the included RCTs was higher than would be seen in real-world clinical practice. Furthermore, no information on the modalities used to stage the tumour was provided.

Table 25: Study-Selection Criteria and Methods for SLR and ITCs

ALK = anaplastic lymphoma kinase; BSC = best supportive care; CR = complete response; DFS = disease-free survival; EFS = event-free survival; EGFR = epidermal growth factor receptor; HRQoL = health-related quality of life; ITC = indirect treatment comparison; MPR = major pathologic response; NICE = National Institute for Health and Care Excellence; nmNSCLC = nonmetastatic non–small cell lung cancer; OS = overall survival; pCR = pathologic complete response; PD = progressive disease; PFS = progression-free survival; PR = partial response; RCT = randomized controlled trial; RFS = recurrence-free survival; S-1 = tegafur, gimeracil, and oteracil; SLR = systematic literature review; UFT = tegafur and uracil.

^aBased on study designs in which patients were enrolled before surgery with the intention of surgical resection.

^bIn trials that include other stages (i.e., stage IA or stage IIIB), no more than 20% of the trial population could represent stages that were not of interest, otherwise such trials were ineligible. Of note, although the CheckMate 816 population was restricted to patients with stage IB disease and tumour sizes ≥ 4 cm, this criterion was not applied to the population, intervention, control, and outcomes study design (PICOS), as this was not consistently reported across trials.

^cTo address the objectives of treatment patterns, patients receiving BSC or no treatment at all are included.

^dTargeted therapies include oncogene-targeted therapies, such as tyrosine kinase inhibitors and drugs like bevacizumab.

^eFirst-generation-based chemotherapies were included if they were used in the neoadjuvant setting.

^fCisplatin or carboplatin in combination with gemcitabine, vinorelbine, paclitaxel, docetaxel, or pemetrexed. Note that if a trial arm included a mix of treatments (e.g., third- and second-generation platinum-based chemotherapies), at least 80% of the patients in the trial had to have received third-generation platinum-based chemotherapies for the study to be included in the base-case analysis).

^gIn studies involving neoadjuvant chemotherapy or neoadjuvant chemoradiotherapy (including, for example, neoadjuvant chemotherapy and surgery), the additional administration of postsurgical chemotherapy and/or radiotherapy was permitted. Postsurgical chemotherapy and/or radiotherapy could differ across treatment arms if administered to only a subset of patients (e.g., postsurgical chemotherapy given to patients who responded to neoadjuvant chemotherapy, or chemotherapy and/or radiotherapy given to patients with an R1 or R2 resection). However, if protocol-defined chemotherapy and/or radiotherapy was administered to all patients, irrespective of surgical outcome or response to initial therapy, the same postsurgical regimen must have been used in all relevant trial arms being compared, otherwise such trials were ineligible.

^hCisplatin or carboplatin in combination with the following second-generation therapies: ifosfamide, mitomycin, vindesine, vinblastine, and etoposide. Note that if a trial included a mix of treatments (e.g., second- and first-generation platinum-based chemotherapies), at least 80% of the patients in the trial had to have received second- or third-generation platinum-based chemotherapies for the study to be included in the sensitivity analysis). All other trials (e.g., trials enrolling > 20% of patients who received first-generation platinum-based chemotherapies) were excluded.

ⁱFirst-generation chemotherapies include methotrexate, cyclophosphamide, vincristine, and doxorubicin.

^jLocoregional recurrences and distant metastases were not included in the PICOS of the SLR that informed the NMA (although these data were extracted, when available); however, given their relevance to the cost-effectiveness analysis, these outcomes were included in the NMA PICOS.

Sources: Sponsor SLR and ITC report.^16,42

Table 26: ITC Analysis Methods

DFS = disease-free survival; DIC = deviance information criterion; EFS = event-free survival; ITC = indirect treatment comparison; NICE = National Institute for Health and Care Excellence; NMA = network meta-analysis; OS = overall survival; pCR = pathologic complete response; PD-L1 = programmed cell death 1 ligand 1; PFS = progression-free survival; PRISMA = Preferred Reporting Items for Systematic Reviews and Meta-Analyses; RFS = relapse-free survival; SD = standard deviation; TTDM = time to death or distant metastases; TTLRR = time to locoregional recurrence.

Source: Sponsor ITC Report.¹⁶

Table 27: Study Characteristics for Base-Case Analysis in the Sponsor-Submitted NMA

IQR = interquartile range; NMA = network meta-analysis; NR = not reported.

^aIFCT 0101 involved 2 different regimens of induction chemoradiotherapy, which included a third cytotoxic drug (i.e., cisplatin and vinorelbine, carboplatin and paclitaxel).

Source: Sponsor ITC Report.¹⁶

Table 28: Patient Characteristics for the Studies Included in the Base-Case Analysis in the Sponsor-Submitted NMA