Reimbursement Review

Amivantamab (Rybrevant)

Sponsor: Janssen Inc.

Therapeutic area: Locally advanced or metastatic non-small cell lung cancer

This multi-part report includes:

Clinical Review

Pharmacoeconomic Review

Clinical Review

Abbreviations

Executive Summary

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

Table 1: Background Information of Application Submitted for Review

NOC = Notice of Compliance.

Introduction

Lung cancer is the most commonly diagnosed cancer type in Canada, as well as the leading cause of cancer-related death in the country.¹ Canadian Cancer Statistics (2023) estimate that 1 in 14 Canadians will be diagnosed with lung cancer in their lifetime and that 1 in 4 cancer-related deaths in Canada will be attributed to the disease.¹ The overwhelming majority of newly diagnosed lung cancer cases in Canada are attributed to non–small cell lung cancer (NSCLC) (88%), and the 5-year net survival rates for patients in Canada with advanced (stage IV) NSCLC is only 3%.¹ Goals of treatment for advanced NSCLC include delaying progression, prolonging survival, palliation of symptoms, and improving quality of life.² EGFR exon 20 insertion (20ins) is a rare mutation that is associated with aggressive, highly symptomatic disease and significant clinical burden. Multiple studies have found that patients with EGFR exon 20ins-positive NSCLC are typically female, nonsmokers, and diagnosed with metastatic disease at approximately 60 years of age.^3-6 In Canada, it has been estimated that EGFR exon 20ins account for approximately 5% of EGFR mutations and between 0.4% to 1.2% of all NSCLC cases, with provincial variation likely being driven by differences in population demographics (exon 20ins mutations are more prevalent in those with East Asian ethnicity).^7-20 According to the clinical experts consulted by the review team, testing for EGFR exon 20ins mutations is currently performed as part of the standard of care for locally advanced or metastatic NSCLC in Canada.

Currently, there are no approved targeted therapies for patients with EGFR exon 20ins in the first-line setting. The current recommended standard of care remains chemotherapy (cisplatin or carboplatin, generally in combination with pemetrexed, followed by pemetrexed maintenance therapy).^21,22 Treatment with chemotherapy alone does not provide a durable treatment benefit for patients with EGFR exon 20ins and is associated with poor survival outcomes (median progression-free survival [PFS] ranging from 4.2 to 6.9 months and median overall survival [OS] from 16.1 to 22.4 months).^8,23-28

The objective of this report is to review and critically appraise the evidence submitted by the sponsor on the beneficial and harmful effects of amivantamab 1,400 mg IV infusion for the first-line treatment of adult patients with locally advanced (not amenable to curative therapy) or metastatic NSCLC with activating EGFR exon 20ins mutations.

Perspectives of Patients, Clinicians, and Drug Programs

The information in this section is a summary of input provided by the patient and clinician groups who responded to the Canada’s Drug Agency (CDA-AMC) call for input and from clinical expert(s) consulted by CDA-AMC for the purpose of this review.

Patient Input

This section was prepared by the CDA-AMC review team based on the input provided by patient groups. The full patient and clinician group submissions received by CDA-AMC are available in the consolidated patient and clinician group input document for this review on the project website.

The patient input was submitted jointly by Lung Cancer Canada (LCC), the Canadian Cancer Survivor Network (CCSN), and the Lung Health Foundation (LHF). Nine patients with NSCLC provided input through virtual interviews by LCC and the CCSN. Among them, 7 patients were from Canada, 1 from the US, and 1 from the UK. All data were collected between May and July in 2024. Based on the input, patients with exon 20ins mutations face a unique challenge, as the mutation is insensitive to conventional tyrosine kinase inhibitors (TKIs). Thus, patients face a poorer prognosis, necessitating different treatment options. Patient groups indicated that there is an urgent unmet need for novel treatment options for patients with EGFR-positive exon 20ins NSCLC. According to the patient group input, improved management of disease symptoms, quality of life, and survival, as well as delayed disease progression and manageable side effects, are considered important outcomes by patients with NSCLC. All participants (n = 9) had experience with amivantamab. Generally, patients indicated that amivantamab was effective in stabilizing disease and maintaining quality of life with manageable side effects. The most common side effects reported included facial and scalp rashes, cuts on fingers and toes, paronychia, eye dryness, sensitivity to the sun, fatigue, skin sensitivity, and nausea.

Clinician Input

Input From Clinical Experts Consulted by CDA-AMC

• The clinical experts consulted by CDA-AMC on this review noted the toxicities associated with cytotoxic chemotherapy, its limited efficacy, and the need for targeted therapies with better efficacy.

• The clinical experts noted that amivantamab would be added to current standard doublet therapy in the first-line setting. The clinical experts believed that the patients most likely to respond to amivantamab are those with the EGFR exon 20ins mutation.

• The clinical experts noted that response would be determined by serial physical and symptom assessments and radiography, typically with CT scans every 8 to 12 weeks during therapy or if there were new symptoms suggesting progression. The clinical experts believed that treatment should be discontinued when patients experience unacceptable toxicity or significant progression or choose to stop treatment.

Clinician Group Input

Two clinician groups submitted input — the LCC Medical Advisory Committee (MAC) and the Ontario Health — Cancer Care Ontario (OH-CCO) Lung Cancer Drug Advisory Committee. In total, 32 clinicians from the LCC MAC and 7 clinicians from OH-CCO Lung Cancer Drug Advisory Committee provided input to the submissions.

The clinician groups agreed with the clinical experts consulted by CDA-AMC that, because of poor outcomes with current available treatments, there is a significant unmet need for novel targeted therapies with better efficacy. The clinician groups agreed with the clinical experts consulted by CDA-AMC that amivantamab plus carboplatin-pemetrexed should be used in the first-line setting, and clinicians from the OH-CCO Lung Cancer Drug Advisory Committee indicated that amivantamab plus carboplatin-pemetrexed can replace pembrolizumab or ipilimumab and/or nivolumab. All clinicians agreed that patients with EGFR exon 20ins mutations are best suited for treatment with amivantamab plus carboplatin-pemetrexed. Clinicians from the LCC MAC suggested radiological response assessments every 6 to 9 weeks, and clinicians from the OH-CCO Lung Cancer Drug Advisory Committee suggested response assessments every 9 to 12 weeks. All clinicians agreed that disease progression and unacceptable toxicities should be considered when deciding to discontinue treatment. All clinicians agreed that specialists with experience in using systemic therapy in cancer care are required for the treatment with amivantamab; outpatient cancer centres, satellite facilities, or hospitals would be appropriate settings.

Drug Program Input

Input was obtained from the drug programs that participate in the CDA-AMC reimbursement review process. The clinical experts consulted by CDA-AMC provided advice on the potential implementation issues raised by the drug programs (Table 5).

Clinical Evidence

Systematic Review

Description of Study

The PAPILLON study is an ongoing, phase III, multicentre, open-label trial conducted in 131 centres across 25 countries, including 3 study sites in Canada. Patients were ≥ 18 years of age with treatment-naive EGFR exon 20ins mutated locally advanced or metastatic NSCLC. The primary objective of the study was to assess the efficacy of amivantamab plus carboplatin-pemetrexed compared with carboplatin-pemetrexed in the first-line treatment of patients with EGFR exon 20ins mutated NSCLC. A total of 308 patients were randomized 1:1 to either the amivantamab plus carboplatin-pemetrexed or carboplatin-pemetrexed arm from December 2020 to November 2022. The clinical cut-off date for data inclusion was May 2023. The treatment phase for each participant started at cycle 1 day 1 and continued in 21-day cycles until the end of treatment visit (approximately 30 days after discontinuation of study treatment), and patients continued treatment until documented radiographic disease progression. Eligible patients in the carboplatin-pemetrexed arm who experienced disease progression were given the option to enter the crossover phase and receive amivantamab monotherapy in 21-day cycles. The primary outcome was PFS, assessed by blinded independent committee review (BICR), and secondary outcomes included OS and objective response rate (ORR).

Patients in the PAPILLON trial were 60 years of age, on average, and the majority were female and had no smoking history (58% each). Of total patients, 65% had an Eastern Cooperative Oncology Group Performance Status (ECOG PS) of 1 and the remainder had an ECOG score of 0, and the majority of patients (77%) had no history of brain metastases. Almost all patients had adenocarcinoma and had not used prior EGFR inhibitors (99% each). Almost all patients had either stage IVA (45%) or IVB (54%) NSCLC at screening. All patients had prior surgery for lung cancer, and 35% had prior radiotherapy. Although there were some differences in specific baseline characteristics between the amivantamab plus carboplatin-pemetrexed and carboplatin-pemetrexed groups, the clinical experts consulted by CDA-AMC on this review did not believe them to be clinically relevant.

Efficacy Results

As of the data cut-off in May 2023, there was a median follow-up of 14.92 months.

Progression-Free Survival

Analyses were performed after 216 events had been observed (amivantamab plus carboplatin-pemetrexed arm: 84, carboplatin-pemetrexed arm: 132). A treatment effect favouring amivantamab plus carboplatin-pemetrexed was observed, with an HR of 0.395 (95% confidence interval [CI], 0.296 to 0.528; P < 0.0001). The median PFS by BICR in the amivantamab plus carboplatin-pemetrexed arm was 11.37 months (95% CI, 9.79 to 13.70) compared with the carboplatin-pemetrexed arm (6.70 months; 95% CI, 5.59 to 7.33). Preplanned sensitivity analyses evaluating PFS, as assessed by the treating investigator, as well as a nonstratified analysis were consistent with the primary analysis. At the 6-month time point, the absolute difference in event-free probability between amivantamab plus carboplatin-pemetrexed and carboplatin-pemetrexed groups was █████ ████ ███ ████ ██ ███████ and at 12 months was █████ ████ ███ ████ ██ ██████ and at 18 months was █████ ████ ███ ████ ██ ███████. Note that by 18 months there were only 17 patients (15 in the amivantamab plus carboplatin-pemetrexed group and 2 in the carboplatin-pemetrexed group) remaining at risk.

Overall Survival

As of the data cut-off in May 2023, there were a total of 70 deaths reported across both groups (amivantamab plus carboplatin-pemetrexed arm: 28 deaths, carboplatin-pemetrexed arm: 42 deaths). At this point, 65 patients from the carboplatin-pemetrexed group had crossed over to amivantamab monotherapy, and the HR was 0.675 (95% CI, 0.418 to 1.090; P = 0.106). Median OS in the amivantamab plus carboplatin-pemetrexed arm was not estimable (NE), and the median OS in the carboplatin-pemetrexed arm was 24.38 (95% CI, 22.08 to NE) months. The final planned OS analysis will be conducted on more mature OS data, approximately 48 months after the first participant was randomized, when approximately 210 deaths overall are anticipated. Results of the stratified sensitivity analysis were consistent with unstratified analysis. At the 12-month time point, the absolute difference in event-free probability between amivantamab plus carboplatin-pemetrexed and carboplatin-pemetrexed groups was ████ ████ ███ ████ ██ ███████ and at 18 months was ████ ████ ███ ████ ██ ██████ and at 24 months was █████ ████ ███ ████ ██ ███████. Note that by 18 months there were only 11 patients (5 in the amivantamab plus carboplatin-pemetrexed group and 6 in the carboplatin-pemetrexed group) who remained at risk. The sponsor also conducted sensitivity analyses to adjust for patients who crossed over to amivantamab monotherapy from carboplatin-pemetrexed: inverse probability of censoring weighting (IPCW) (adjusted HR of ████ ████ ███ █████ ████), two-stage estimation (TSE) (adjusted HR of ████ ████ ███ █████ ████) and rank-preserving structural failure time (RPSFT) (adjusted HR of ███ ████ ███ █████ ████) analyses.

Objective Response Rates

Based on BICR assessment in patients with measurable disease at baseline (n = 304), there was a higher percentage of responders in the amivantamab plus carboplatin-pemetrexed arm (111 patients [72.5%]), compared with the carboplatin-pemetrexed arm (72 patients [46.4%]) at the data cut-off for an odds ratio (OR) of █████ ████ ███ ██████ ██████ ████████), favouring amivantamab plus carboplatin-pemetrexed. The absolute difference between groups ███ ███ ████ ███ ███ ██ ███). The percentage of patients with progressive disease as the best response in the amivantamab plus carboplatin-pemetrexed arm was ████ compared with █████ in the carboplatin-pemetrexed arm. Results of the sensitivity analysis evaluating ORR, as assessed by the treating investigator, were consistent with the assessment by BICR.

Health-Related Quality of Life

For the European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30 (EORTC QLQ-C30) Global Health Status (GHS) (range, 0 [worst] to 100 [best] points), the mean change from baseline to 12 months was ███ ████ ███ █████ ███) in the amivantamab plus carboplatin-pemetrexed group and ████ ████ ███ █████ ███) in the carboplatin-pemetrexed group. The mean difference between amivantamab plus carboplatin-pemetrexed and carboplatin-pemetrexed groups for change from baseline to 12 months was ███ ████ ███ █████ ███).

Progression-Free Survival After First Subsequent Therapy

At the data cut-off, there were ██ of 153 patients ███████ in the amivantamab plus carboplatin-pemetrexed group and ██ of 155 patients in the carboplatin-pemetrexed group (██████ who had PFS after first subsequent therapy (PFS2) event. The HR was 0.493 (95% CI, 0.32 to 0.759; P = 0.001). This analysis was not adjusted for multiplicity. The median PFS2 was NE (95% CI, 22.77 to NE) in the amivantamab plus carboplatin-pemetrexed arm and 17.25 months (95% CI, 13.96 to 21.52) in the carboplatin-pemetrexed arm.

Harms Results

Adverse Events

Overall, all 151 patients (100%) in the amivantamab plus carboplatin-pemetrexed group and 152 patients (98.1%) in the carboplatin-pemetrexed group experienced at least 1 adverse event (AE). The most common AE was rash, which occurred in 81 patients (53.6%) in the amivantamab plus carboplatin-pemetrexed group and 12 patients (7.7%) in the carboplatin-pemetrexed group. Other common AEs with large differences between groups included dermatitis acneiform, in 47 patients (31.1%) in the amivantamab plus carboplatin-pemetrexed group and 5 patients (3.2%) in the carboplatin-pemetrexed group; hypoalbuminemia, in 62 (41.1%) patients in the amivantamab plus carboplatin-pemetrexed group and 15 (9.7%) patients in the carboplatin-pemetrexed group; peripheral edema, in 45 (29.8%) of patients in the amivantamab plus carboplatin-pemetrexed group and 16 (10.3%) patients in the carboplatin-pemetrexed group; and infusion-related reaction, in 63 (41.7%) patients in the amivantamab plus carboplatin-pemetrexed group and 2 (1.3%) patients in the carboplatin-pemetrexed group.

Serious Adverse Events

Treatment-emergent serious AEs (SAEs) were reported in 56 patients (37.1%) in the amivantamab plus carboplatin-pemetrexed group and 48 patients (31.0%) in the carboplatin-pemetrexed group. The most common SAEs in the amivantamab plus carboplatin-pemetrexed group were pneumonia (6 patients [4.0%] versus 4 patients [2.6%] in the carboplatin-pemetrexed group), pneumonitis (4 patients [2.6%] and no patients in carboplatin-pemetrexed), and pulmonary embolism (4 patients [2.6%] in each group). All other SAEs in the amivantamab plus carboplatin-pemetrexed group had an incidence of less than 2%. The most common SAE in the carboplatin-pemetrexed arm was anemia (6 patients [3.9%] and 1 patient [0.7%] in the amivantamab plus carboplatin-pemetrexed group).

Withdrawals Due to Adverse Events

Overall, from study initiation through the data cut-off, 36 patients (23.8%) in the amivantamab plus carboplatin-pemetrexed arm and 16 patients (10.3%) in the carboplatin-pemetrexed arm had treatment-emergent AEs (TEAEs) leading to discontinuation of at least 1 study treatment. Of the 36 patients in the amivantamab plus carboplatin-pemetrexed arm who discontinued any study treatment, 17 discontinued amivantamab. Pneumonitis was the most common cause of discontinuation of amivantamab (4 patients [2.6%]), followed by dermatitis acneiform (████████ ████████ Thrombocytopenia (3 patients [1.9%]) and neutropenia (2 patients [1.3%]) were the most common reasons for treatment discontinuation of either carboplatin or pemetrexed in the carboplatin-pemetrexed arm. Most of the TEAEs leading to study treatment discontinuation occurred at a frequency of less than 2% in both treatment arms.

Mortality

There were 4 patients (2.6%) in the amivantamab plus carboplatin-pemetrexed group and 9 patients (5.8%) in the carboplatin-pemetrexed group who died due to an AE during the study, and 3 patients (2.0%) in the amivantamab plus carboplatin-pemetrexed group and 4 patients (2.6%) in the carboplatin-pemetrexed group who died due to an AE within 30 days of the last dose.

Notable Harms

The notable harms identified for this review were rash and infusion reactions. As noted previously, rash was the most common AE, occurring in 81 (53.6%) patients in the amivantamab plus carboplatin-pemetrexed group and 12 (7.7%) patients in the carboplatin-pemetrexed group. Under the grouping skin and subcutaneous disorders, there were ███ ████████ ███████ ████████ in the amivantamab plus carboplatin-pemetrexed group and ██ ███████ patients in the carboplatin-pemetrexed group with such an event. Infusion-related reactions were reported in 63 (41.7%) patients in the amivantamab plus carboplatin-pemetrexed group and 2 (1.3%) patients in the carboplatin-pemetrexed group.

Critical Appraisal

• The lack of blinding in the PAPILLON trial introduces significant potential for bias in the assessment of patient-reported outcomes such as health-related quality of life (HRQoL). There was a large number of patients (42%) who crossed over from the carboplatin-pemetrexed group to the amivantamab monotherapy after disease progression, which is a major confounder when assessing OS. Another limitation of assessment of OS is that the data are not yet mature.

• With respect to external validity, the clinical experts consulted on this review believed the population enrolled in the PAPILLON trial to be generalizable to the patients they expect to be treated with amivantamab in Canada, although they would likely consider expanding the population to an ECOG PS of 2, rather than limiting it to ECOG PS of 0 or 1, as was seen in the trial.

GRADE Summary of Findings and Certainty of the Evidence

The selection of outcomes for Grading of Recommendations Assessment, Development and Evaluation (GRADE) assessment was based on the sponsor’s Summary of Clinical Evidence, consultation with clinical experts, and input received from patient and clinician groups and public drug plans. The following list of outcomes was finalized in consultation with expert committee members:

• PFS

• OS

• ORR

• HRQoL (EORTC QLQ-C30 GHS)

• Notable harms: rash, infusion-related reactions.

The target of the certainty-of-evidence assessment was the presence or absence of a clinically important effect based on thresholds informed by the clinical experts consulted for this review for PFS and OS. The literature-based minimal important difference (MID) of 10 points was used for the EORTC QLQ-C30 GHS.^29,30 This MID has been estimated for within-group changes and was applied in the absence of an estimate of the MID for a between-group difference. The target of the certainty of evidence was the presence or absence of any (non-null) effect for the ORR because a threshold for a clinically important between-group difference could not be estimated.

Table 2: Summary of Findings for Amivantamab Plus Carboplatin-Pemetrexed Versus Carboplatin-Pemetrexed for Patients With NSCLC With Exon 20ins

BICR = blinded independent central review; CI = confidence interval; CP = carboplatin-pemetrexed; EORTC QLQ-C30 = European Organization for the Research and Treatment of Cancer Quality of Life Questionnaire Core 30; HRQoL = health-related quality of life; NA = not applicable; NR = not reported; NSCLC = non–small cell lung cancer; ORR = objective response rate; OS = overall survival; PFS = progression-free survival; RCT = randomized controlled trial.

Notes: Study limitations (which refer to internal validity or risk of bias), inconsistency across studies, indirectness, imprecision of effects, and publication bias were considered when assessing the certainty of the evidence. All serious concerns in these domains that led to the rating down of the level of certainty are documented in the table footnotes.

Cycle 17 day 1 and cycle 25 day 1 were selected to correspond 12- and 18-month time points, respectively. For EORTC QLQ-C30 GHS, means and 95% CIs are derived based on the mixed-effects model with repeated measures, in which the dependent variable is change from baseline in score, and independent variables are baseline, visit, treatment, and visit by treatment interaction as fixed effects and individual patient as random effect.

The between-group differences for PFS, OS, HRQoL, and notable harms were not part of the sponsor’s statistical analysis plan and were requested by the review team to facilitate interpretation.

^aThe clinical experts consulted by the review team considered that both the point estimate and lower bound of the CI constituted clinically meaningful benefit.

^bRated down 1 level for study limitations; results are from an interim analysis and there is a risk of bias due to confounding as a result of crossover of patients from the carboplatin-pemetrexed group to amivantamab monotherapy postprogression. Rated down 1 level for imprecision; the point estimate suggests benefit and CI includes little to no difference and in some cases, potential harm (threshold of 5% suggested by clinical experts).

^cNo threshold of clinical importance could be established; effects were appraised using the null. Rated down 1 level for indirectness; this is a surrogate end point without strong evidence that it predicts the treatment effect on OS.

^dRated down 2 levels for study limitations; there is risk of bias due to (a) lack of blinding and a subjective outcome, (b) substantial missing outcome data. Based on a MID of 10 points, the point estimate and both bounds of the CI suggest little to no difference. The 10-point MID has been estimated for within-group changes, and was applied in the absence of an estimate of a between-group MID. However, both within- and between- group differences were smaller than the MID.

^eLower-level rash was used instead of higher-level rash because it was thought to capture rash events more specifically.

^fRated down 1 level for study limitations; there is a risk of bias due to lack of blinding and potential subjectivity in the outcomes.

Sources: Details included in the table are from the sponsor’s Summary of Clinical Evidence.³¹ Data request of the sponsor, and the Clinical Study Report (CSR) for the PAPILLON trial.

Long-Term Extension Studies

No long-term extension studies were submitted.

Indirect Comparisons

The sponsor submitted an indirect treatment comparison (ITC) report that included analyses that used individual patient data from the PAPILLON trial and real-world databases and inverse probability of treatment weighting (IPTW) methods to reduce the risk of bias due to confounding. This was considered by the review team as a nonrandomized study and is reported in the following section.

Studies Addressing Gaps in the Evidence From the Systematic Review

This section includes the contents of the ITC report submitted by the sponsor, which were considered by the review team to be a nonrandomized study, as it utilized individual patient data for each arm in the comparison.

Description of the Nonrandomized Study

The sponsor performed a nonrandomized study using IPTW. These analyses used individual patient data from the PAPILLON trial for amivantamab plus carboplatin-pemetrexed and from real-world databases COTA and ConcertAI for EGFR TKI monotherapy and platinum chemotherapy plus immunotherapy (IO). The analyses examined the outcomes of OS, PFS, real-world PFS, and time to next treatment (TTNT).

Covariates were identified by the sponsor that were considered treatment effect modifiers or prognostic factors. The base-case results were adjusted for ECOG at index date, history of brain metastases, history of liver metastases, and age at index date. The full model adjusted for all the variables in the base case, plus Asian race, history of smoking, sex, and history of other metastases.

For the PAPILLON trial versus EGFR TKI comparison, before weighting, moderate differences (absolute standardized mean difference [SMD] ████ ███ ████) were observed for ECOG at index date and history of smoking, and substantial differences (absolute SMD ████) were seen for history of liver metastases, age, Asian race, sex, and history of other metastases. In the primary analysis, base-case average treatment effect in the treated (ATT) weighting balanced (SMD ████) all 4 covariates between cohorts. In the full model, however, all 8 of the included factors had absolute SMDs ████, indicating lack of balance, with moderate differences observed for 6 factors (ECOG at index date, history of liver metastases, age, history of smoking, sex, history of other metastases), and substantial differences observed for the remaining 2 factors (history of brain metastases, Asian race). The resulting effective sample size (ESS) in the EGFR TKI cohort was ████ ███ ████ for the base case and full model, respectively, compared to the original ██ observations.

For the PAPILLON trial versus platinum plus IO comparison, before weighting, substantial (absolute SMD ████) differences were observed for ECOG at index date, history of liver metastases, history of brain metastases, age, Asian race, and history of other metastases. In the primary analysis, ATT weighting reduced the proportion of categories with absolute SMDs ████ from ████ ██████ ██ ███ ██████, with only moderate differences observed for the 2 remaining factors (ECOG at index date, age). The balance between populations in the full model, on the other hand, remained the same (6 of 8 factors balanced) following weighting, although moderate differences were observed for 3 factors (ECOG at index date, history of brain metastases, history of smoking) and substantial differences were observed for the remaining 3 factors (history of liver metastases, Asian race, history of other metastases). The resulting ESS in the platinum plus IO cohort was █████ ███ ████ for the base case and full model, respectively, compared to the original ██ observations.

Efficacy Results

The unadjusted comparison of OS for the PAPILLON trial versus EGFR TKI produced an HR of ████ ████ ███ █████ █████ ███████) in favour of amivantamab plus carboplatin-pemetrexed. The primary analysis, using base-case ATT weighting, produced results in favour of amivantamab plus carboplatin-pemetrexed, with an HR of ████ ████ ███ █████ █████ ███████). The unadjusted comparison of OS for the PAPILLON trial versus platinum plus IO, produced an HR of ████ ████ ███ █████ █████ ███████) in favour of amivantamab plus carboplatin-pemetrexed. The primary analysis, using base-case ATT weighting, produced a point estimate favouring amivantamab plus carboplatin-pemetrexed, but the 95% CI crossed the null, with an HR of ████ ████ ███ █████ █████ ███████).

The unadjusted comparison of PFS for the PAPILLON trial versus EGFR TKI produced an HR of ████ ████ ███ █████ █████ ███████); the point estimate favoured amivantamab plus carboplatin-pemetrexed, but the 95% CI crossed the null. Similarly, the primary analysis, using base-case ATT weighting, produced an HR of ████ ████ ███ █████ █████ ███████). The unadjusted comparison of PFS for the PAPILLON trial versus platinum plus IO produced an HR of ████ ████ ███ █████ █████ ███████), favouring amivantamab plus carboplatin-pemetrexed. The primary analysis, using base-case ATT weighting, produced results in favour of amivantamab plus carboplatin-pemetrexed, with an HR of ████ ████ ███ █████ █████ ███████).

The unadjusted comparison of real-world PFS (RW-PFS) for the PAPILLON trial versus EGFR TKI produced an HR of ████ ████ ███ █████ █████ ████████, favouring amivantamab plus carboplatin-pemetrexed. The primary analysis, using base-case ATT weighting, produced results in favour of amivantamab plus carboplatin-pemetrexed, with an HR of ████ ████ ███ █████ █████ ███████). The unadjusted comparison of RW-PFS for the PAPILLON trial versus platinum plus IO produced an HR of ████ ████ ███ █████ █████ ███████), favouring amivantamab plus carboplatin-pemetrexed. The primary analysis, using base-case ATT weighting, produced results in favour of amivantamab plus carboplatin-pemetrexed, with an HR of ████ ████ ███ █████ █████ ███████).

The unadjusted comparison of TTNT for the PAPILLON trial versus EGFR TKI produced an HR of ████ ████ ███ █████ █████ ███████) favouring amivantamab plus carboplatin-pemetrexed. The primary analysis, using base-case ATT weighting, produced results in favour of amivantamab plus carboplatin-pemetrexed, with an HR of ████ ████ ███ █████ █████ ███████). The unadjusted comparison of TTNT for the PAPILLON trial versus platinum plus IO produced an HR of ████ ████ ███ █████ █████ ███████), favouring amivantamab plus carboplatin-pemetrexed. The primary analysis, using base-case ATT weighting, produced results in favour of amivantamab plus carboplatin-pemetrexed, with an HR of ████ ████ ███ █████ █████ ███████).

Results of the other sensitivity analyses (not shown) were mostly in the direction of favouring amivantamab plus carboplatin-pemetrexed, but in some cases the results were imprecise, with CIs that crossed the null.

Harms Results

Harms were not assessed in the nonrandomized studies.

Critical Appraisal

There was no predefined protocol available, and the search and selection criteria, data extraction, and methods to appraise the risk of bias were not described. There was minimal information related to the real-world data source regarding data quality and completeness, suitability, and validity of any algorithms used to identify patients and classify outcomes. Additionally, due to the lack of protocol, there is an increased risk of bias in the selection of the reported results.

Propensity score–based methods aim to reduce the risk of bias due to confounding. However, it is important to note that no comparisons using data from randomized cohorts were used in the sponsor’s analysis. Lack of randomization within the datasets introduces the possibility of imbalance of patient characteristics, which could lead to comparing groups of patients who do not possess similar prognostic risk. While the methods used by the sponsor serve to reduce confounding, the results indicate that a high risk of residual confounding is present in the analyses, even after the adjustments that were made on prognostic and effect-modifying covariates. While a sensitivity analysis of the full model including all 8 factors was conducted, the primary analysis of the base case only adjusted for 4 factors; it is unlikely that this represents all relevant prognostic and effect-modifying variables. In many cases, the full model was associated with high SMDs, indicating evidence of differences in baseline characteristics between the groups being compared. The base-case model, which used 4 factors, also had notable imbalances.

The use of real-world data has several limitations. Participants in the PAPILLON trial were monitored more strictly than were the patients included in the ITC from the real-world databases. Monitoring of patients in the real-world databases was likely to be less rigorous. OS measurements may include errors or missing deaths, or censoring may differ between the clinical and real-word data sources. The sponsor suggested that missing deaths in real-world data may result in an overestimation of OS. The handling of missing data in the databases and in the analyses was not clearly described in the sponsor’s report.

The effective sample sizes were very small in the base case and in several of the full model analyses. For example, the ESS was ██ in the IPTW-ATT full model for the EGFR TKI group. In the full model, the majority of the 8 included factors remained unbalanced (SMDs ████) following ATT weighting for both the the PAPILLON trail versus EGFR TKI (all 8 factors imbalanced) and platinum plus IO comparisons (6 of 8 factors were imbalanced). Therefore, the reliability of the results from the full model is expected to be low due to risk of bias and the small sample size. The results of the base-case model are based on populations that have greater similarity to one another (in comparison to the full model). However, imbalances remained, and the base-case model did not control for all the important baseline prognostic covariates.

The selection of comparators in the analyses lack clinical relevance in the context of Canada. The relevance is also limited by the lack of analyses, including patient-reported outcomes such as HRQoL.

Conclusions

There is evidence from 1 pivotal, phase III, open-label, randomized controlled trial (RCT) (PAPILLON, n = 308) that amivantamab plus carboplatin-pemetrexed results in a clinically important improvement in PFS compared to carboplatin-pemetrexed alone in patients with locally advanced NSCLC with an exon 20ins mutation. Evidence from the trial suggested that amivantamab plus carboplatin-pemetrexed may improve OS over carboplatin-pemetrexed alone. However, conclusions regarding OS are limited because the OS data are not yet mature, a large number of crossovers occurred in this trial, and the estimates were imprecise (i.e., 95% CIs included effects that may not be clinically important and crossed the null). The combination of amivantamab plus carboplatin-pemetrexed likely improves ORR compared to carboplatin-pemetrexed alone, although the clinical significance of this improvement is less clear, given that ORR is a surrogate outcome without strong evidence of a potential to predict the treatment effect on OS. Amivantamab plus carboplatin-pemetrexed may result in little to no difference in HRQoL; the results of this assessment were likely biased by the open-label design of the trial and patient attrition. Rash and infusion-related reactions are more likely to occur when amivantamab is added to carboplatin-pemetrexed. The most common notable harm associated with amivantamab was rash. However, this only rarely resulted in serious harm or prompted the patient to stop therapy.

A nonrandomized study using individual patient data from the PAPILLON trail and real-world databases, as well as IPTW methods to reduce the risk of bias due to confounding, aimed at determining the comparative efficacy of amivantamab plus carboplatin-pemetrexed versus EGFR TKI and chemotherapy plus IO for several efficacy outcomes (OS, PFS, RW-PFS, TTNT). Despite the analyses suggesting superiority of amivantamab plus carboplatin-pemetrexed for most outcomes, the evidence was insufficient to draw definitive conclusions on the relative efficacy of amivantamab plus carboplatin-pemetrexed or the precise magnitude of any potential difference. The clinical experts consulted by the review team suggested that superiority of amivantamab plus chemotherapy over EGFR TKI monotherapy or chemotherapy plus IO is plausible, given that the 2 comparators are not known to be efficacious in patients with NSCLC with EGFR exon 20ins mutations. The nonrandomized study provided no information on HRQoL, an outcome of importance to patients, or on harms.

Introduction

The objective of this report is to review and critically appraise the evidence submitted by the sponsor on the beneficial and harmful effects of amivantamab 1,400 mg IV infusion for the first-line treatment of adult patients with locally advanced (not amenable to curative therapy) or metastatic NSCLC with activating EGFR exon 20ins mutations.

Disease Background

Contents within this section have been informed by materials submitted by the sponsor and clinical expert input. The following have been summarized and validated by the CDA-AMC review team.

Lung cancer is the most commonly diagnosed cancer type in Canada, as well as the leading cause of cancer-related death in the country.¹ Canadian Cancer Statistics (2023) estimate that 1 in 14 Canadians will be diagnosed with lung cancer in their lifetime and that 1 in 4 cancer-related deaths in Canada will be attributed to the disease.¹ It is estimated that in 2024 there were 32,100 Canadians diagnosed with lung and bronchus cancer.³² The overwhelming majority of newly diagnosed lung cancer cases in Canada are attributed to NSCLC (88%), which is a serious terminal illness that is associated with a poor prognosis, as the 5-year net survival rate for patients in Canada with advanced (stage IV) NSCLC is only 3%.¹ Goals of treatment for advanced NSCLC include delaying progression, prolonging survival, palliation of symptoms, and improving quality of life.²

Scientific advances over the past decade have allowed for a better understanding of advanced NSCLC and its mechanisms of action.^19,33 One key mechanism are “driver” mutations, which activate progrowth signalling pathways. The most prevalent driver mutation in NSCLC results in the activation of EGFR, which has been identified in 15% of Western populations³⁴ and 40% to 50% of Asian populations.³⁵ EGFR mutations can be classified into 2 distinct groups: common EGFR mutations, which comprise 85% to 90% of all EGFR mutations and include exon 19 deletions and exon 21 L858R point mutations, and uncommon EGFR mutations, which constitute the remaining 10% to 15% and include exon 20ins mutations.^{8,19, 36-38} EGFR exon 20ins is a rare mutation that is associated with aggressive, highly symptomatic disease and significant clinical burden. EGFR exon 20ins mutations are heterogenous at the molecular level, with more than 70 types of mutations described to date.^3,19,36,39 Multiple studies have found that patients with EGFR exon 20ins-positive NSCLC are typically female, nonsmokers, and diagnosed with metastatic disease at approximately 60 years of age.^3-6 In Canada, it has been estimated that EGFR exon 20ins account for approximately 5% of EGFR mutations and between 0.4% to 1.2% of all NSCLC cases, with provincial variation likely being driven by differences in population demographics (exon 20ins mutations are more prevalent in those with East Asian ethnicity).^7-10,12-20 Patients with EGFR exon 20ins experience a poor HRQoL due to frequent disease-related symptoms, such as fatigue, pain, shortness of breath and cough, and negative impacts on daily activities, including self-care, social activities, and family life.^40-43

Diagnosis of NSCLC often occurs at late-stage disease and requires a biopsy in which sufficient tissue is obtained for histologic confirmation.¹ Following the biopsy, a diagnosis of 1 of the 3 distinct histological types of NSCLC (i.e., adenocarcinoma, squamous cell carcinoma, and large cell carcinoma) is made, and the tumour is staged.⁴⁴ Assessment of EGFR mutation status is recommended as part of a sequential or parallel approach to facilitate treatment decision-making.⁴⁵

Standards of Therapy

Contents within this section have been informed by materials submitted by the sponsor and clinical expert input. The following have been summarized and validated by the CDA-AMC review team.

Treatment of EGFR-mutated NSCLC is not a “one size fits all” approach, as key differences exist in the mechanisms of action between common and uncommon mutations.¹⁹ While common EGFR mutations respond well to EGFR TKIs, they are largely ineffective against EGFR exon 20ins mutations.^46,47 Given the paucity of effective treatment options and lack of approved first-line targeted therapies, survival outcomes are worse in patients diagnosed with EGFR exon 20ins mutations versus common EGFR mutations.⁸ A population-based study of patients with EGFR-positive NSCLC in Alberta, Canada, found real-world median OS in patients with EGFR exon 20ins was significantly worse than in patients with EGFR exon 19 deletions or exon 21 L858R mutations.⁷

As per the clinical experts consulted by the review team, current treatment goals for patients with advanced NSCLC include prolonging PFS and OS and maintaining or improving HRQoL. In 2023, the American Society of Clinical Oncology and the European Society for Medical Oncology published treatment guidelines recommending chemotherapy for patients with EGFR exon 20ins-positive NSCLC in the first-line setting.^48,49 According to the clinical experts consulted by the review team, platinum-doublet chemotherapy in first-line treatment, and then other cytotoxic chemotherapies (e.g., docetaxel), would be offered to the present patient target population. Patients with EGFR exon 20ins mutations do not respond to standard EGFR targeting TKIs, such as osimertinib, that have led to much better outcomes in patients with sensitizing EGFR mutations, i.e., exon 19 deletion and exon 21 L858R substitution, compared to platinum-doublet chemotherapy. Further, the patient target population does not have meaningful responses to IO checkpoint inhibitors. Current therapy with cytotoxic chemotherapy provides a modest survival benefit over best supportive care. However, benefits are typically short-lived, and chemotherapies can have significant toxicities, according to the clinical experts consulted by the review team. Treatment with chemotherapy in patients with EGFR exon 20ins has been associated with poor survival outcomes (median PFS ranging from 4.2 to 6.9 months and median OS from 16.1 to 22.4 months).^8,23-28

Currently, there are no approved targeted therapies for patients with exon 20ins in the first-line setting. The current recommended standard of care remains chemotherapy (cisplatin or carboplatin, generally in combination with pemetrexed, followed by pemetrexed maintenance).^21,22

Drug Under Review

Key characteristics of amivantamab are summarized in Table 3 with other treatments available for locally advanced (not amenable to curative therapy) or metastatic NSCLC with activating EGFR exon 20ins mutations.

The recommended dosage for amivantamab is a once weekly IV infusion, at a dose of 1,400 mg (1,750 mg if body weight is 80 kg or greater) for 4 weeks (first dose split on days 1 and 2), then 1,750 mg (2,100 mg if body weight is ≥ 80 kg) on day 1 of each 21-day cycle, starting with cycle 3. Amivantamab has received Health Canada approval for the first-line treatment of adult patients with locally advanced (not amenable to curative therapy) or metastatic NSCLC with activating EGFR exon 20ins mutations. The sponsor’s reimbursement request is consistent with the indication. Amivantamab is a bispecific antibody that binds to both the EGFR and mesenchymal-epithelial transition (MET) receptor. It is thought that the EGFR and MET pathways compensate for each other in situations where 1 pathway is inhibited, leading to the so-called “kinase switch” drug resistance (seen with EGFR TKIs).⁵⁰ Thus, simultaneous inhibition of both EGFR and MET may improve overall treatment efficacy by limiting the compensatory pathway activation.⁵⁰

Table 3: Key Characteristics of Amivantamab Plus Carboplatin-Pemetrexed, Platinum-Based Chemotherapy, EGFR TKI, and IO

AUC = area under the curve; CP = carboplatin-pemetrexed; exon 20ins = exon 20 insertion; IO = immunotherapy; NSCLC = non–small cell lung cancer; PD-1 = programmed cell death protein 1; TKI = tyrosine kinase inhibitor.

^aAlthough IO plus chemotherapy is not currently indicated for the first-line treatment of patients with EGFR exon 20ins NSCLC, it is commonly prescribed to patients while waiting to receive next-generation sequencing (NGS) testing results for EGFR tumour aberrations. Refer to page 17 for more information on this treatment regimen.

Source: Product monographs for amivantamab, carboplatin, pemetrexed, afatinib, gefitinib, erlotinib, and pembrolizumab.^53-59

Testing Procedure Considerations

The clinical experts consulted for this review confirmed that testing for EGFR exon 20ins mutations is currently done as routine care in patients with advanced nonsquamous NSCLC. Although the process varies across different provinces and territories, reflex testing for EGFR mutations, including exon 20ins, is the recommended standard of care in Canada.^60,61 According to the clinical experts, depending on the jurisdiction, EGFR testing can be ordered directly by the pathologist (reflex testing), referred directly by the lab or pathologist to an external testing site (reflex referral), or ordered by a health care provider such as oncologist or surgeon.

Testing for EGFR exon 20ins mutations is conducted on tissue samples collected through biopsy as part of the routine diagnostic work-up. EGFR exon 20ins mutations can be identified by either polymerase chain reaction or next-generation sequencing (NGS) testing methods.⁶⁰ NGS is the preferred method, since targeted polymerase chain reaction (PCR) may under-detect EGFR exon 20ins mutations.^60,61 In instances where there is inadequate biopsy tissue available to test, NGS testing of cell-free circulating tumour DNA (ctDNA) drawn using a noninvasive blood sample could be a comparable alternative.^61,62 However, the clinical experts consulted for this review pointed out that patients may have to pay out of pocket for ctDNA testing. According to the clinical experts, about 10% of nonsquamous NSCLC patients in Canada are tested for EGFR exon 20ins mutations by ctDNA analysis.

We considered the potential impacts of EGFR exon 20ins mutation testing to ascertain eligibility for the first-line treatment with amivantamab plus PC for locally advanced or metastatic NSCLC, including impacts to health systems, patients (including families and caregivers), and costs; these impacts are not anticipated to be substantial. Key considerations and relevant information available from materials submitted by the sponsor, input from the clinical experts consulted by the review team, and sources from the literature were validated by the review team and are summarized in Table 4.

Table 4: Considerations for EGFR Exon 20ins Mutation Testing for Establishing Treatment Eligibility for Amivantamab in Locally Advanced or Metastatic NSCLC

exon 20ins = exon 20 insertion; NGS = next-generation sequencing; NSCLC = non–small cell lung cancer.

^aCDA-AMC have not evaluated or critically appraised this evidence to determine its validity or reliability.

Source: Sponsor’s Summary of Clinical Evidence.³¹

Perspectives of Patients, Clinicians, and Drug Programs

Patient Group Input

This section was prepared by the CDA-AMC review team based on the input provided by patient groups. The full patient and clinician group submissions received by CDA-AMC are available in the consolidated patient and clinician group input document for this review on the project website.

Patient Input

The patient input submission was jointly submitted by LCC, the CCSN, and the LHF. LCC serves as a leading resource for lung cancer education, patient support, research, and advocacy in Canada. The CCSN is a national network of patients, families, survivors, friends, community partners, funders, and sponsors who have come together to take action to promote the very best standard of care. The LHF (previously named the Ontario Lung Association) assists and empowers people living with or caring for others with lung disease. Nine patients with NSCLC provided inputs through virtual interviews by LCC and the CCSN. Among them, 7 patients were from Canada, 1 from the US, and 1 from the UK. All participants (n = 9) had experience with amivantamab. All data were collected between May and July in 2024.

Most (6 of 9) patients had stage 4 NSCLC, with metastases to bone, liver, lymph nodes, and/or brain when first diagnosed with NSCLC. Patient groups highlighted that the most common subsets of EGFR mutations include exon 19 deletion and exon 21 L858R mutations, while EGFR exon 20 mutations are much less common, with a frequency of roughly 0.1% to 4% in NSCLC. Patients with exon 20 mutations face a unique challenge as the mutation is insensitive to conventional TKIs and, thus, face a poorer prognosis, necessitating different treatment options. Platinum-doublet chemotherapy followed by single-drug docetaxel is the current standard of care for patients with EGFR-positive exon 20ins NSCLC. However, chemotherapy is limited as a viable long-term treatment option due to its nature as a systemic treatment with harsh and toxic side effects, which often creates additional burden on patients, leading to decreased function, poorer quality of life, and increased dependence on caregivers for daily activities. Immunotherapies, such as EGFR-specific TKIs, are available treatment options for patients with EGFR-positive exon 20ins NSCLC but have limited benefit, as patients may experience insensitivity. Additionally, radiotherapy, such as whole-brain radiation therapy, is effective in stabilizing brain metastases but may cause memory loss. Therefore, patient groups indicated that there is an urgent unmet need for novel treatment options for patients with EGFR-positive exon 20ins NSCLC. According to the patient group input, improved management of disease symptoms, quality of life, and survival, as well as delayed disease progression and manageable side effects, are considered important outcomes by patients with NSCLC. Generally, patients indicated that amivantamab was effective in stabilizing disease and maintaining quality of life with manageable side effects. The most common side effects reported were facial and scalp rashes, cuts on fingers and toes, paronychia, eye dryness, sensitivity to the sun, fatigue, skin sensitivity, and nausea. Other less common side effects noted by patients included low appetite, mild peripheral neuropathy, muscle aches, tinnitus, and gastrointestinal issues (constipation or diarrhea).

Clinician Input

Input From Clinical Experts Consulted by CDA-AMC

All CDA-AMC review teams include at least 1 clinical specialist with expertise regarding 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 by CDA-AMC on this review noted that chemotherapy is not well-tolerated and the benefit is modest and short-lived; not all patients respond to chemotherapy and most who do respond quickly become refractory. Targeted therapies with better efficacy are needed, as they directly target the underlying mechanisms of malignant proliferation. According to the clinical experts consulted for this review, the goals of therapy would be to prolong survival, delay time to progression of disease, and maintain maximum quality of life.

Place in Therapy

According to the clinical experts, amivantamab would be added to current standard doublet chemotherapy in the first-line setting, per the evidence from the PAPILLON trail.⁶⁵ If a patient did not tolerate the cytotoxic chemotherapy component, that could be dropped, and the patient could continue on amivantamab monotherapy, as the mechanism is independent from that of chemotherapy. According to the clinical experts, this would preclude patients who do not tolerate chemotherapy from having to move on to a distinct second line of chemotherapy due to intolerance to first-line chemotherapy.

Patient Population

The clinical experts believed that the patients most suited to receive amivantamab would be those with an identified exon 20ins mutation, noting that testing for this mutation is widely available in Canada, although the wait time for EGFR exon 20ins mutation testing can be variable, depending on the region of Canada. One clinical expert noted that some jurisdictions are still struggling with timely testing results, meaning that mutation testing results are not available when a patient needs to start therapy. This clinical expert also noted that, if an exon 20ins mutation was identified after a patient had already started platinum-doublet chemotherapy, the patient should be allowed to add amivantamab at that point in time. The clinical experts noted that underdiagnosis may occur in settings with inadequate tissue for NGS testing; ctDNA testing may play a role to fill that gap.

Assessing the Response to Treatment

The clinical experts noted that tumour response is determined in clinical practice with serial physical and symptom assessments and serial radiography, typically with serial CT scans every 8 to 12 weeks during therapy or in the setting of new symptoms concerning disease progression during therapy. The clinical experts added that a meaningful response to treatment would involve a patient whose cancer is shrinking or has stopped growing, or even a rate of growth that has slowed significantly in the setting of a clinical benefit.

Discontinuing Treatment

The clinical experts believed that treatment should be discontinued when patients experience unacceptable toxicity despite appropriate supportive care and dose reductions, or significant progression of disease, or when patients choose to stop treatment.

Prescribing Considerations

With respect to setting, the clinical experts believed that amivantamab should be administered in systemic therapy units with experience in managing acute infusion reactions, under the supervision of a medical oncologist.

Clinician Group Input

This section was prepared by the CDA-AMC review team based on the input provided by clinician groups. The full patient and clinician group submissions received by CDA-AMC are available in the consolidated patient and clinician group input document for this review on the project website.

Two clinician groups submitted inputs, the LCC MAC and the OH-CCO Lung Cancer Drug Advisory Committee. In total, 32 clinicians from the LCC MAC and 7 clinicians from OH-CCO Lung Cancer Drug Advisory Committee provided input to the submissions.

The clinician groups agreed with the clinical experts consulted by CDA-AMC that, because of poor outcomes with current available treatments, there is a significant unmet need for novel targeted therapies with better efficacy. The clinician groups agreed with the clinical experts consulted by CDA-AMC that amivantamab plus carboplatin-pemetrexed should be used in the first-line setting, and clinicians from the OH-CCO Lung Cancer Drug Advisory Committee indicated that amivantamab plus carboplatin-pemetrexed can replace pembrolizumab or ipilimumab and/or nivolumab. All clinicians agreed that patients with EGFR exon 20ins mutations are best suited for treatment with amivantamab plus carboplatin-pemetrexed. Clinicians from the LCC MAC suggested radiological response assessments every 6 to 9 weeks, and clinicians from the OH-CCO Lung Cancer Drug Advisory Committee suggested response assessments every 9 to 12 weeks. All clinicians agreed that disease progression and unacceptable toxicities should be considered when deciding to discontinue treatment. All clinicians agreed that specialists with experience in using systemic therapy in cancer care are required for the treatment with amivantamab; outpatient cancer centres, satellite facilities, or hospitals would be appropriate settings.

Drug Program Input

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

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

CNS = central nervous system; ECOG PS = Eastern Cooperative Oncology Group Performance Status; NGS = next-generation sequencing; NSCLC = non–small cell lung cancer; pERC = pan-Canadian Oncology Review Expert Review Committee; TKI = tyrosine kinase inhibitor.

Clinical Evidence

The objective of the CDA-AMC Clinical Review report is to review and critically appraise the clinical evidence submitted by the sponsor on the beneficial and harmful effects of amivantamab in the first-line treatment of adult patients with locally advanced (not amenable to curative therapy) or metastatic NSCLC in patients with activating EGFR exon 20ins mutation. The focus will be placed on assessment of comparative effectiveness and safety comparing amivantamab to relevant comparators, including identifying gaps in the current evidence.

A summary of the clinical evidence included by the sponsor in the review of amivantamab is presented in 4 sections, with the CDA-AMC critical appraisal of the evidence included at the end of each section. The first section, the systematic review, includes pivotal studies and RCTs that were selected according to the sponsor’s systematic review protocol. The CDA-AMC assessment of the certainty of the evidence in this first section using the GRADE approach follows the critical appraisal of the evidence. The second section included indirect evidence from the sponsor; the sponsor submitted an ITC report that used individual patient data for both arms. This was considered as a nonrandomized study by the review team and presented in the Studies Addressing Gaps section. No long-term extension studies were submitted.

Included Studies

Clinical evidence from the following are included in the CDA-AMC review and appraised in this document:

• 1 pivotal RCT identified in systematic review

• 1 study addressing gaps in the systematic review evidence (submitted by the sponsor as an ITC report).

Systematic Review

Contents in this section have been informed by materials submitted by the sponsor. The following have been summarized and validated by the CDA-AMC review team.

Description of Studies

Characteristics of the included studies are summarized in Table 6.

The PAPILLON trial is an ongoing, phase III, multicentre, open-label trial conducted in 131 centres across 25 countries, including 3 sites in Canada. Patients were 18 years of age or older, with treatment-naive EGFR exon 20ins mutated locally advanced or metastatic NSCLC. The primary objective of the study was to assess the efficacy of amivantamab plus carboplatin-pemetrexed compared with carboplatin-pemetrexed in the first-line treatment of patients with EGFR exon 20ins mutated NSCLC. A total of 308 patients were randomized 1:1 to either the amivantamab plus carboplatin-pemetrexed or carboplatin-pemetrexed arm using an interactive web response system, from December 2020 to November 2022. The clinical cut-off date for data inclusion was May 2023. A screening period 28 days before randomization was used to ensure study eligibility, which was assessed through a baseline radiographic assessment of disease sites and a review of test results and/or tumour biopsy to confirm EGFR exon 20ins mutation status. The treatment phase for each patient started at cycle 1 day 1 and continued in 21-day cycles until the end of treatment visit (approximately 30 days after discontinuation of study treatment).

Eligible patients in the carboplatin-pemetrexed arm who experienced disease progression were given the option to enter the crossover phase and receive amivantamab monotherapy in 21-day cycles (described in additional detail in Appendix 3). Patients discontinued study treatment for documented radiographic disease progression (confirmed by BICR, according to Response Evaluation Criteria in Solid Tumours [RECIST] v1.1) or if they met another criterion for discontinuation of study treatment. Patients who discontinued assigned study treatment were followed for subsequent therapy, disease status, survival, and symptomatic progression in the follow-up phase. This phase started after the end of treatment visit and continued until the end of the study, death, loss to follow-up, or withdrawal of consent from participation in the study, whichever came first (refer to Figure 1 for a schematic overview of the study design).

Figure 1: Schematic Overview of the PAPILLON Study Design

AUC = area under the curve; C2D1 = cycle 2 day 1; C3D1 = cycle 3 day 1; C4D1 = cycle 4 day 1; QW = once weekly.

* Stratification factors: Brain metastases (yes or no); ECOG PS (0 or 1); prior EGFR TKI use (yes or no).

† Doses shown by body weight (< 80 kg/≥ 80 kg).

‡ Cycle 1: first dose split between days 1 and 2, then doses on days 8 and 15; cycle 2: day 1.

Table 6: Details of Studies Included in the Systematic Review

ACP = amivantamab + carboplatin-pemetrexed; AUC = area under the curve; BICR = blinded independent central review; CCO = clinical cut-off date; CP = carboplatin-pemetrexed; DOR = duration of response; ECOG = Eastern Cooperative Oncology Group; EORTC QLQ-C30 = European Organization for the Research and Treatment of Cancer Quality of Life Questionnaire Core 30; exon 20ins = exon 20 insertion; ILD = interstitial lung disease; NSCLC = non–small cell lung cancer; ORR = objective response rate; OS = overall survival; PFS = progression-free survival; PFS2 = progression-free survival after first subsequent therapy; PRO = patient-reported outcomes; PROMIS PF = Patient-Reported Outcomes Measurement Information System — Physical Function; RECIST = Response Evaluation Criteria in Solid Tumours; RCT = randomized controlled trial; TTD = time to discontinuation; TTSP = time to symptomatic progression; TTST = time to subsequent therapy.

Note: 2 additional reports were included (CSR for PAPILLON, sponsor submission to CDA-AMC).

Source: Details included in the table are from the sponsor’s Summary of Clinical Evidence.³¹

Populations

Inclusion and Exclusion Criteria

Key inclusion criteria include patients 18 years of age or older, with histologically or cytologically confirmed, locally advanced or metastatic, nonsquamous NSCLC with documented primary EGFR exon 20ins mutations, measurable disease according to RECIST v1.1, ECOG PS 0 or 1, and adequate organ and bone marrow function. Key exclusion criteria include having received any prior systemic treatment for locally advanced or metastatic disease, with evidence of synchronous NSCLC disease, or with untreated brain metastases.

Interventions

Patients in the amivantamab plus carboplatin-pemetrexed arm received the following study treatment:

• amivantamab 1,400 mg (1,750 mg if body weight is ≥ 80 kg) by IV infusion once weekly through cycle 1 (day 1 on weeks 1, 2, and 3) and on cycle 2 day 1 (day 1 week 4); then 1,750 mg (2,100 mg if body weight is ≥ 80 kg) on day 1 of each 21-day cycle, starting at week 7 (cycle 3 day 1)

• pemetrexed 500 mg/m² (with vitamin supplementation) on day 1 of each 21-day cycle, in combination with carboplatin for up to 4 cycles, and then as maintenance monotherapy until disease progression

• carboplatin area under the curve (AUC) 5 on day 1 of each 21-day cycle, for up to 4 cycles.

The entire regimen was infused in an outpatient chemotherapy suite and took between 2 and 6 hours depending on cycle, day, and dose. Patients in the carboplatin-pemetrexed arm received the following study treatment:

• pemetrexed 500 mg/m² (with vitamin supplementation) on day 1 of each 21-day cycle, in combination with carboplatin for up to 4 cycles, and then as maintenance monotherapy until disease progression

• carboplatin AUC 5 of day 1 of each 21-day cycle, for up for 4 cycles.

Assigned study treatment was continued until documented radiographic (RECIST v1.1) disease progression, confirmed by BICR or another criterion for discontinuation, was met. Continuation of study treatment after confirmed disease progression was allowed after approval from the medical monitor, if the investigator believed the participant was deriving clinical benefit.

Patients in the carboplatin-pemetrexed group were allowed to switch to amivantamab monotherapy once they had progressed, referred to as the crossover phase. All participants in the crossover phase received amivantamab in 21-day cycles at a dose of 1,400 mg (1,750 mg if body weight ≥ 80 kg) by IV infusion once weekly up to cycle 2 day 1, then 1,750 mg (2,100 mg if body weight ≥ 80 kg) on day 1 of each 21-day cycle, starting with cycle 3.

Participants who discontinued assigned study treatment for any reason were followed for subsequent therapy, disease status, survival, and symptomatic progression in the follow-up phase. For the secondary end points related to subsequent therapy, a crossover from carboplatin-pemetrexed to amivantamab monotherapy was also considered as initiation of a subsequent systemic therapy.

Throughout the study, investigators were allowed to prescribe any concomitant medications or treatments deemed necessary to provide adequate supportive care, except for those listed as prohibited therapies. Prohibited therapies included any chemotherapy, anticancer therapy, experimental therapies, radiotherapy to tumour lesions being assessed for response, phenytoin or phosphenytoin (with carboplatin), and live or attenuated vaccines. Medications that were permitted (with caution) included concurrent use of a nonsteroidal anti-inflammatory with pemetrexed in patients with creatinine clearance less than 80 mL/min, ibuprofen specifically, within 2 days (before or after) of administration of pemetrexed, and any drug that may decrease serum magnesium, due to the hypomagnesemia that can occur with EGFR inhibitors.

There were both required and optional concomitant medications given specifically with amivantamab to prevent infusion-related reactions. The required medications were IV glucocorticoids (cycle 1 day 2), IV or oral antihistamines (before amivantamab infusion), and IV or oral antipyretics (before amivantamab infusion). Optional medications included IV or oral glucocorticoids (cycle 1 day 8 and onwards), IV histamine H₂-receptor antagonists (any cycle), and antiemetics (ondansetron oral or IV, any cycle). The pre- and postinfusion medications given for chemotherapy included corticosteroids, folic acid, and vitamin B12, given concomitantly with pemetrexed.

Outcomes

A list of efficacy end points assessed in this Clinical Review Report is provided in Table 7, followed by descriptions of the outcome measures. Summarized end points are based on outcomes included in the sponsor’s Summary of Clinical Evidence as well as any outcomes identified as important to this review, according to the clinical expert(s) consulted by CDA-AMC and patient, clinician, and drug program input. Using the same considerations, the CDA-AMC review team selected end points that were considered to be most relevant to inform the CDA-AMC expert committee deliberations and finalized this list of end points in consultation with members of the expert committee. All summarized efficacy end points were assessed using GRADE. Select notable harms outcomes considered important for informing the CDA-AMC expert committee deliberations were also assessed using GRADE. Based on the input received, PFS, OS, ORR, and HRQoL were identified as important outcomes. The clinical experts consulted for this review selected these end points based on their relevance to patients and their use in making treatment decisions. For HRQoL, the EORTC QLQ-C30 GHS was chosen as the outcome to assess with GRADE as the clinical experts believed it to be a key HRQoL measure with relevance to patients. Patient-Reported Outcomes Measurement Information System — Physical Function (PROMIS PF) total score was included in the Appendix 1 of this Clinical Review Report in support of patients’ HRQoL. The time points chosen to perform GRADE analyses were chosen to balance the need to obtain as late a follow-up as possible (particularly for outcomes like OS and PFS) with the reality that, at these later time points, very few patients remain at risk, therefore limiting any conclusions that can be drawn from these data. The clinical experts believed that rash and infusion-related reactions were key notable harms.

Progression-Free Survival

The primary end point was PFS, defined as the time from randomization until the date of objective disease progression (based on BICR using RECIST v1.1) or death (by any cause), whichever comes first. Participants who had not progressed or died at the time of analysis were censored at the time of the latest date of their last evaluable RECIST v1.1 assessment. If the participant progressed or died after 2 or more consecutive missed disease assessments, the participant was censored at the time of the last evaluable RECIST v1.1 assessment. If the participant had no evaluable visits or did not have any baseline data, they were censored at day 1, unless they died within 2 visits of baseline. Radiographic disease progression was assessed during imaging at regular 6-week intervals. The sponsor is not aware of a reported MID specific to the EGFR exon 20ins population.

PFS After First Subsequent Therapy 2

PFS2 by investigator was defined as the time from randomization until the date of second objective disease progression, after initiation of subsequent anticancer therapy, based on investigator assessment (after that used for PFS) or death, whichever came first. The sponsor is not aware of a reported MID specific to the EGFR exon 20ins population.

Objective Response Rate

ORR was defined as the proportion of participants who achieved either a complete response or partial response, as defined by BICR using RECIST v1.1. Data obtained up until progression or last evaluable disease assessment in the absence of progression was included in the assessment of ORR. However, any complete response or partial response that occurred after a further anticancer therapy was received was not included in the numerator for the ORR calculation. Participants who did not have a tumour response assessment for any reason were considered nonresponders and were included in the denominator when calculating the response rate. The sponsor is not aware of a reported MID specific to the EGFR exon 20ins population.

Overall Survival

OS was defined as the time from the date of randomization until the date of death due to any cause. Any participant not known to have died at the time of analysis was censored based on the last recorded date on which the participant was known to be alive. The sponsor is not aware of a reported MID specific to the EGFR exon 20ins population. The clinical experts consulted by the review team considered a 5% between-group difference as the smallest clinically important difference.

European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30

EORTC QLQ-C30 is used to measure self-reported functioning for all cancer types in patients with cancer. It includes 30 items resulting in 5 functional scales (physical functioning, role functioning, emotional functioning, cognitive functioning, and social functioning), 1 GHS and quality of life scale, 3 symptom scales (fatigue, nausea and vomiting, and pain), and 6 single items (dyspnea, insomnia, appetite loss, constipation, diarrhea, and financial difficulties). The instrument contains 28 items using a verbal rating scale with 4 response options: “Not at All,” “A Little,” “Quite a Bit,” and “Very Much” (scored 1 to 4). Two additional items use response options (1 to 7): 1 = Very Poor to 7 = Excellent. All scale and item scores were linearly transformed to be in the range from 0 to 100 according to the algorithm in EORTC QLQ-C30 scoring manual, version 3.0.⁶⁷ A higher score represents a higher (“better”) level of functioning, or a higher (“worse”) level of symptoms. This was measured as change from baseline and time to deterioration for symptom, GHS, and physical functioning scores. The meaningful worsening threshold for time to deterioration in the PAPILLON trail was specified by the sponsor as an increase of 10 points or more for symptom scales and a decrease of 10 points or more for health status and functioning scales.

PROMIS PF Short Form V2.0

Another measure of patient-reported outcomes supportive of HRQoL outcomes in the target patient population was the PROMIS PF short form. In the presence of survival and EORTC QLQ-C30 data, this instrument was not considered to be most critical to guide treatment selection in clinical practice. Results are included in Appendix 1 and not assessed with GRADE. For a description of the measure and its meaningful worsening threshold identified in the literature, please refer to Table 8.

Safety Evaluation Plan

The verbatim terms used in the case report form by investigators to identify AEs were coded using the Medical Dictionary for Regulatory Activities (MedDRA) Version 25.0. Any AE occurring at or after the initial administration of study treatment through the day of last dose plus 30 days or until the start of subsequent anticancer therapy (if earlier) was considered to be treatment-emergent. Treatment-related AEs were those judged by the investigator to be at least possibly related to the study medication. Patients with multiple occurrences of events were only counted once at the maximum severity to study medication for each preferred team, system organ class, and overall. Deaths that occurred within 30 days after the last dose of study medication were defined as on-study deaths.

Data for PFS2 were also summarized in the main section of the report because the clinical experts believed this outcome to be of some relevance in the population. However, a GRADE assessment was not performed because this was not considered an outcome of critical importance for decision-making.

Table 7: Outcomes Summarized From the Studies Included in the Systematic Review

EORTC QLQ-C30 = European Organization for the Research and Treatment of Cancer Quality of Life Questionnaire Core 30; HRQoL = health-related quality of life; ORR = objective response rate; OS = overall survival; PFS = progression-free survival; PFS2 = progression-free survival after first subsequent therapy.

^aStatistical testing for these end points was adjusted for multiple comparisons (e.g., hierarchal testing).

Source: Details included in the table are from the sponsor’s Summary of Clinical Evidence.³¹

Table 8: Summary of Outcome Measures and Their Measurement Properties

EORTC QLQ-C30 = European Organization for the Research and Treatment of Cancer Quality of Life Questionnaire Core 30; HRQoL = health-related quality of life; MID = minimal important difference; PF = physical function; PRO = patient-reported outcomes.

Statistical Analysis

Sample Size and Power Calculation

The total sample size needed for the PAPILLON trail was approximately 300 participants (150 per group). This was calculated given that a total of 200 PFS events would provide approximately 90% power to detect a hazard ratio (HR) of 0.625 at a 2-sided alpha level of 0.05. This estimation corresponded to an extension of at least 3 months in the median PFS, which was estimated at 8 months in the amivantamab plus carboplatin-pemetrexed group and 5 months in the carboplatin-pemetrexed group. The basis for these estimates was not reported. The sample size calculation took into consideration an annual dropout rate of 5%.

Statistical Test or Model

In the PAPILLON trail, all primary and secondary efficacy end points were analyzed using the full analysis set.

The primary end point (PFS by BICR) was analyzed using a log-rank test at the overall 2-tailed significance level of 0.05. The log-rank test was stratified by ECOG PS (0 or 1) and history of brain metastases (yes or no) using the Breslow approach for handling ties. The Kaplan-Meier product limit was used to estimate the median PFS, and a stratified Cox model using the same stratification factors as the log-rank test was used to estimate the HR, along with the associated 95% CIs. The proportional hazards assumption was examined by plotting log(-log[estimated survival distribution function]) against log(survival time). In addition, a treatment by logarithm-transformed time interaction term was added into the primary Cox model and tested. A P value greater than 0.05 for the interaction term was interpreted as no statistical evidence against the proportional hazards assumption.

Multiple Testing Procedure

A hierarchical fixed-sequence testing strategy was used to control the overall type I error rate for hypotheses testing of primary and key secondary end points ORR and OS at 5%. If the testing for the primary end point of PFS was statistically significant, key secondary end points, including ORR and OS, were sequentially tested, with an overall 2-sided alpha of 0.05. The test for ORR was to be conducted before the test for OS.

Based on the O’Brien Fleming alpha spending approach, a 2-sided alpha of 0.0008 was allocated to the interim analysis, approximately 48 months after the first participant was randomized, when approximately 210 deaths overall are anticipated. The final analysis will be conducted at a 2-sided alpha of 0.0498.

Data Imputation Methods

No imputation method was used for handling missing or incomplete data.

Subgroup Analyses

Prespecified subgroup analyses were performed for the efficacy and/or safety end points. Only the subgroup data reported for the primary outcome was of interest for this report. The following subgroups were used:

• age group (< 65 years, ≥ 65 years; < 75 years, ≥ 75 years)

• sex (male, female)

• race (Asian, non-Asian)

• weight (< 80 kg, ≥ 80 kg)

• history of brain metastasis (yes, no)

• ECOG PS (0, 1)

• history of smoking (yes, no).

HRs with 95% CI were reported for each of the individual subgroups. However, no P values were reported and no tests for interactions appeared to have been performed.

Sensitivity Analyses

The sensitivity analyses performed for the primary outcome included an unstratified log-rank test and an assessment of the proportional hazards assumption by plotting log (-log[estimated survival distribution function]) against log (survival time). Additionally, a treatment by logarithm-transformed time interaction term was added into the primary Cox model and tested. A P value greater than 0.05 for the interaction term was interpreted as no statistical evidence against the proportional hazards assumption. An additional sensitivity analysis, which used investigator assessments for radiographic progression, does not appear to have been prespecified.

Secondary Outcomes

The secondary end point of ORR was analyzed using a logistic regression model at the overall 2-tailed significance level of 0.05 and stratified by ECOG PS (0 or 1) and history of brain metastases (yes or no). Results were presented as an OR with its associated 95% CIs.

The secondary end point of OS was tested sequentially following significance of PFS and ORR end points. OS was tested with a nonstratified log-rank test as the primary analysis and a sensitivity analysis using a stratified log-rank test (using the same methodology and model as for the analysis of PFS). The analysis of OS was a priori planned to be conducted at 2 time points: (1) at the time of the primary analysis of PFS, when approximately 85 deaths overall were anticipated, and (2) approximately 48 months after the first patient was randomized, when about 210 deaths overall are anticipated. To try to account for the large number of crossovers, the sponsor performed 3 different sensitivity analyses specifically for OS: IPCW, TSE methods, and RPSFT models. IPCW artificially censors patients whose NSCLC had progressed when they switched from carboplatin-pemetrexed to amivantamab, and the remaining observations from patients whose NSCLC had not yet progressed when they switched to amivantamab are upweighted to represent the censored patients. This weighting process aims to account for the similarities in baseline and time-varying factors between the patients whose NSCLC had progressed who were censored and those who were not, thereby mitigating the selection bias caused by censoring. In the TSE method, the impact of crossover is estimated by comparing those patients who crossed over to those who did not after a disease-related secondary baseline. The TSE method uses a parametric survival regression model to estimate the effect of crossover by accounting for confounders (baseline and at secondary baseline) as covariates and is therefore reliant upon the “no unmeasured confounding” assumption. RPSFT uses a counterfactual framework to estimate the causal effect of the intervention. It uses the g-estimation method and assumes that the treatment effect in those who crossed over is the same as those who were originally assigned to the intervention. The estimated treatment effect (in terms of time shrinkage) is later applied to adjust for the benefit that those who crossed over obtained from receiving the active treatment, to remove the impact of crossover on the observed results.

The remaining secondary end points, including PFS2, were analyzed using the same methodology and model as for the analysis of PFS. For end points related to subsequent therapy, a crossover from the carboplatin-pemetrexed arm to amivantamab monotherapy was also considered as an initiation of a subsequent systemic therapy.

For EORTC QLQ-C30 domain scores and the PROMIS Physical Function Tscore, descriptive statistics were reported at baseline and at each visit for absolute value and for change from baseline. Statistical tests were not adjusted for multiplicity. All P values were nominal. Instrument level compliance for patient-reported outcomes was summarized for patients expected to complete patient-reported outcome assessments at each time point. The change of scores from baseline over time was assessed using a mixed-effects model for repeated measures analysis based on restricted maximum likelihood. The model included participants as a random effect and baseline value, treatment group, time in week, treatment-by-time interaction, and stratification factors as fixed effects. The treatment comparison was based on the least squares means, and the 2-sided 95% CI was estimated. Randomized participants who received at least 1 dose of study treatment and had at least 1 evaluable postbaseline measurement were included in the analysis.

With respect to harms, all reported TEAEs were included in the analysis. For each AE, the number and percentage of patients who experienced at least 1 occurrence of the given event were summarized by treatment arm. For each treatment group, AE incidence rates were summarized with frequency and percentage by system organ class and preferred term, with all patients treated in that treatment group as the denominator, unless otherwise specified. In addition, AE incidence rates were summarized by severity and relationship to study medication.

Analysis Populations

Refer to Table 9 for analysis populations in the PAPILLON trail.

Table 9: Analysis Populations of PAPILLON

ACP = amivantamab + carboplatin-pemetrexed; CP = carboplatin-pemetrexed.

Source: PAPILLON Clinical Study Report.³³

Results

Patient Disposition

There were more patients in the carboplatin-pemetrexed group who discontinued treatment compared to amivantamab plus carboplatin-pemetrexed (84.5% versus 52.9%), with the most common reason being disease progression (69.0% versus 33.1%) (Table 10). A total of █████ of patients in the carboplatin-pemetrexed group crossed over to amivantamab monotherapy, as was allowed in the protocol. An additional 6 patients in the carboplatin-pemetrexed group crossed over to amivantamab monotherapy off protocol. All patients who crossed over were treated in the crossover phase.

Baseline Characteristics

The baseline characteristics outlined in Table 11 are limited to those that are most relevant to this review or were felt to affect the outcomes or interpretation of the study results.

Patients in the PAPILLON trail were 60 years of age, on average, and the majority were female (57.8%) and had no smoking history (58%). There were 65% of patients with an ECOG PS of 1 (the remainder had an ECOG PS of 0), and the majority of patients (77%) had no history of brain metastases. Almost all patients had adenocarcinoma (98.7%) and had not used prior EGFR inhibitors (99% each). Almost all patients were either stage IVA (███) or IVB (███) at screening. ███ patients had prior surgery for lung cancer and ███ had prior radiotherapy. Although there were some differences in specific baseline characteristics between amivantamab plus carboplatin-pemetrexed and carboplatin-pemetrexed groups (location of metastasis at screening, time since diagnosis, histology, and prior radiotherapy), the clinical experts consulted by CDA-AMC on this review did not believe them to be clinically relevant.

Table 10: Summary of Patient Disposition From PAPILLON

ACP = amivantamab + carboplatin-pemetrexed; CP = carboplatin-pemetrexed; FAS = full analysis set; NA = not applicable; PP = per protocol.

^aThe reason for discontinuation for all study treatment is the reason for discontinuation for the last study treatment received.

^bNumbers and percentages represent analysis conducted on safety analysis set.

Source: Details included in the table are from the sponsor’s Summary of Clinical Evidence.³¹

Table 11: Summary of Baseline Characteristics From PAPILLON, FAS

ACP = amivantamab + carboplatin-pemetrexed; CP = carboplatin-pemetrexed; ECOG = Eastern Cooperative Oncology Group; NSCLC = non–small cell lung cancer; SD = standard deviation; TKI = tyrosine kinase inhibitor.

Note: Percentages calculated with the number of patients in each treatment group as denominator.

^aIn some regions, reporting of race is not required.

^b“Multiple” includes 1 patient who selected Black or African American and white.

^cProtocol allowed for prior TKI therapy if certain criteria were met per exclusion criterion 1.

^d“Other” includes unknown histology grade.

^ePatients can be counted in more than 1 category.

^fAnalyses completed on safety set instead of full analysis set.

^gPrior cancer-related surgery includes tumour biopsies performed for diagnosis.

^hPalliative was captured on the electronic case report form as Curative/Palliative/Any other intent.

ⁱOne patient received both adjuvant and palliative therapy.

Source: Details included in the table are from the sponsor’s Summary of Clinical Evidence.³¹

Exposure to Study Treatments

Overall, there was a difference in exposure to regimen between amivantamab plus carboplatin-pemetrexed and carboplatin-pemetrexed, █████ patient-years versus ████ patient-years. While exposure to carboplatin was similar between groups, exposure to pemetrexed was █████ patient-years in amivantamab plus carboplatin-pemetrexed and ████ patient-years in carboplatin-pemetrexed.

With respect to concomitant medications, there was a relatively large numerical difference in frequency of use between amivantamab plus carboplatin-pemetrexed and carboplatin-pemetrexed groups for antibacterials for systemic use (88.1% and 35.5%, respectively), corticosteroids for systemic use (74.2% and 27.1%, respectively), corticosteroids and dermatological preparations (59.6% and 11.6%, respectively), ophthalmologicals (46.4% and 18.7%, respectively), and emollients and protectives (33.8% and 5.2%, respectively).

Table 12: Summary of Patient Exposure From PAPILLON

ACP = amivantamab + carboplatin-pemetrexed; CP = carboplatin-pemetrexed; SD = standard deviation.

Notes: Analyses conducted on the safety analysis set. Percentages are calculated with the number of patients in each treatment group as the denominator. Amivantamab is administered once weekly (with the first dose split over days 1 and 2) up to cycle 2 day 1, and then administered once every 3 weeks.

^aTotal exposure is the sum of treatment duration for each patient.

^bRelative dose intensity (%) is calculated as the ratio of total received dose versus total prepared dose.

Source: Details included in the table are from the sponsor’s Summary of Clinical Evidence.³¹

Table 13: Patient Exposure to Concomitant Medications in PAPILLON (Safety Analysis Set)

ACP = amivantamab + carboplatin-pemetrexed; CP = carboplatin-pemetrexed.

Source: PAPILLON Clinical Study Report.⁷⁷

Subsequent therapies received by patients in the PAPILLON trail are summarized in Table 14. There were fewer patients in the amivantamab plus carboplatin-pemetrexed than in the carboplatin-pemetrexed group that had received subsequent therapy, 28% versus 61%, respectively, and the most common subsequent therapy was a drug that targeted the EGFR or a TKI-based regimen (8% versus 53%, respectively).

Table 14: Summary of Subsequent Treatment From PAPILLON

ACP = amivantamab + carboplatin-pemetrexed; CP = carboplatin-pemetrexed; IO = immunotherapy; TKI = tyrosine kinase inhibitor; VEGFi = vascular endothelial growth factor inhibitor.

^aTKI in combination with chemotherapy, IO, or vascular endothelial growth factor inhibitor therapies are only counted in the TKI-based regimen category.

Source: Details included in the table are from the sponsor’s Summary of Clinical Evidence.³¹

Efficacy

As of the data cut-off in May 2023, there was a median follow-up of 14.92 (range, 0.3 to 27.0) months.

Progression-Free Survival

Analyses were performed after 216 events had been observed (amivantamab plus carboplatin-pemetrexed arm: 84, carboplatin-pemetrexed arm: 132). The HR for amivantamab plus carboplatin-pemetrexed versus carboplatin-pemetrexed was 0.395 (95% CI, 0.296 to 0.528; P < 0.0001), favouring amivantamab plus carboplatin-pemetrexed (Table 15). The median PFS by BICR in the amivantamab plus carboplatin-pemetrexed arm was 11.37 months (95% CI, 9.79 to 13.70) compared with 6.70 months (95% CI, 5.59 to 7.33) in the carboplatin-pemetrexed arm. The Kaplan-Meier plot of PFS by BICR is in Figure 2.

At the 6-month time point, the absolute difference in PFS rate between amivantamab plus carboplatin-pemetrexed and carboplatin-pemetrexed groups was █████ ████ ███ ████ ██ █████), at 12 months it was █████ ████ ███ ████ ██ █████), and at 18 months the PFS rate was █████ ████ ███ ████ ██ ███████ Note that by 18 months there were only 17 patients (15 in the amivantamab plus carboplatin-pemetrexed group and 2 in the carboplatin-pemetrexed group) remaining at risk. A preplanned sensitivity analysis evaluating PFS, as assessed by the treating investigator, was consistent with the primary analysis. The test for proportional hazards and the residual plots showed that the assumption was satisfied.

Figure 2: Primary End Point: PFS by BICR for Amivantamab Plus Carboplatin and Pemetrexed Versus Carboplatin and Pemetrexed

BICR = blinded independent central review; CI = confidence interval; mo = months; PFS = progression-free survival.

Note: Censoring of data is indicated by tick marks, and median PFS is indicated by dashed lines.

Source: PAPILLON Clinical Summary.⁶⁵

Subgroup analyses, based on the primary outcome of PFS by BICR, are reported in Table 33, Appendix 1. The reported HR with CIs suggest that the treatment effect is consistent across various subgroups based on sex, race (Asian versus non-Asian), weight (< 80 kg versus ≥ 80 kg), ECOG PS (0 or 1), and history of smoking. However, there was some indication that the magnitude of effects of amivantamab plus carboplatin-pemetrexed versus carboplatin-pemetrexed may differ depending on history of brain metastasis (HR of 0.63; 95% CI, 0.38 to 1.06) or no history of brain metastasis (HR of 0.33; 95% CI, 0.23 to 0.46) and possibly the age of patients, i.e., 75 years of age or older (HR of ████ ████ ███ █████ ████), and age younger than 75 (HR of ████ ████ ███ █████ ███████, although it should be noted that the 75 or older age group had a relatively small sample of only ██ patients across groups, compared to ███ patients in those aged younger than 75. The sample for the subgroup of patients with brain metastases was also relatively small (N = ██) across both groups. The purpose of the subgroup analyses was not to test for subgroup differences.

Key Secondary End Points

Objective Response Rates

Based on BICR assessment in patients with measurable disease at baseline (n = 304), there was a higher percentage of responders in the amivantamab plus carboplatin-pemetrexed arm (111 patients [72.5%]), compared with the carboplatin-pemetrexed arm (72 patients [46.4%]) at the data cut-off, for an OR of █████ ████ ███ ██████ ██████ █████████ favouring amivantamab plus carboplatin-pemetrexed (Table 15). The percentage of patients with progressive disease as the best response in the amivantamab plus carboplatin-pemetrexed arm was ████, compared with █████ in the carboplatin-pemetrexed arm. Results of the sensitivity analysis evaluating ORR, as assessed by the treating investigator, were consistent with the assessment by BICR.

Overall Survival

As of the data cut-off in May 2023, after a median follow-up time of 14.92 months, a total of 70 deaths were reported across both groups (amivantamab plus carboplatin-pemetrexed arm: 28 deaths, carboplatin-pemetrexed arm: 42 deaths). At this point, 65 patients from the carboplatin-pemetrexed group had crossed over to amivantamab monotherapy and the HR was 0.675 (95% CI, 0.418 to 1.090; P = 0.106) (Table 15). Median OS in the amivantamab plus carboplatin-pemetrexed arm was NE, and the median OS in the carboplatin-pemetrexed arm was 24.38 (95% CI, 22.08 to NE) months. At the 12-month time point, the absolute difference in OS rate between amivantamab plus carboplatin-pemetrexed and carboplatin-pemetrexed groups was ████ ████ ███ ████ ██ ███████; at 18 months, it was ████ ████ ███ ████ ██ ██████; and at 24 months, it was █████ ████ ███ ████ ██ ███████ Note that by 24 months there were only 11 patients (5 in the amivantamab plus carboplatin-pemetrexed group and 6 in the carboplatin-pemetrexed group) who contributed to the analysis.

The Kaplan-Meier plot of OS is depicted in Figure 4, Appendix 1 of this clinical report. After a slight imbalance in deaths during the first 3 months of the study (██ in the amivantamab plus carboplatin-pemetrexed group and ██ in the carboplatin-pemetrexed group; refer to Harms — Mortality section in this report), the Kaplan-Meier curves described a favourable trend in OS, starting from approximately 9 months, in patients treated with amivantamab plus carboplatin-pemetrexed. The curves are seen to come together approximately between 10 and 11 months before showing evidence of separating again.

The final planned OS analysis will be conducted on more mature OS data, approximately 48 months after the first participant was randomized, when approximately 210 deaths overall are anticipated. Results of the stratified sensitivity analysis were consistent with unstratified analysis.

The sponsor also conducted sensitivity analyses to adjust for patients who crossed over to amivantamab monotherapy from carboplatin-pemetrexed. The OS results were similar in the IPCW (adjusted HR of ████ ████ ███ █████ ██████ and TSE (adjusted HR of ████ ████ ███ █████ ████) analyses and differed slightly from those in the RPSFT (adjusted HR of ███ ████ ███ █████ ████).

PFS After First Subsequent Therapy

At the data cut-off, there were ██ of 153 patients (██████ in the amivantamab plus carboplatin-pemetrexed group and ██ of 155 patients in the carboplatin-pemetrexed group (██████ who had a PFS2 event, for an HR of 0.493 (95% CI, 0.32 to 0.759; P = 0.001) (Table 15). This end point was not controlled for multiplicity, and there is an increased risk of type I error (i.e., erroneously rejecting the null hypothesis). The median PFS2 was NE (95% CI, 22.77 to NE) in the amivantamab plus carboplatin-pemetrexed arm and was 17.25 months (95% CI, 13.96 to 21.52) in the carboplatin-pemetrexed arm.

Health-Related Quality of Life

A summary of the EORTC QLQ-C30 GHS score is presented in Table 15. Baseline compliance was ████ in both the amivantamab plus carboplatin-pemetrexed and carboplatin-pemetrexed treatment groups; compliance exceeded ███ through cycle 31 in both study arms. The number of patients available to complete the measure ██████████ with each cycle. After cycle 9, the number of patients available to complete assessments had ███████ ██ ████ ████ ███ after cycle 9 and after cycle 13 in the carboplatin-pemetrexed and the amivantamab plus carboplatin-pemetrexed groups, respectively. At the 12-month follow-up (cycle 17 day 1), the mean change from baseline on the EORTC QLQ-C30 was ███ ████ ███ █████ ████ in the amivantamab plus carboplatin-pemetrexed group ███ ████ ████ ███ █████ ████ in the carboplatin-pemetrexed group, with a mean difference of change between amivantamab plus carboplatin-pemetrexed and carboplatin-pemetrexed groups of ███ ████ ███ █████ ███) points. The point estimates for the between-group difference in this score did not meet the MID identified from the literature (a ≥ 10-point decrease for health status).

A summary of the PROMIS PF total score is presented in Table 33, Appendix 1. Baseline compliance was ████ in both the amivantamab plus carboplatin-pemetrexed and carboplatin-pemetrexed treatment groups; compliance exceeded ███ through cycle 31 in both study arms. The number of patients available to complete the measure ██████████ with each cycle. The number of patients available to complete assessments had ███████ ██ ████ ████ ███ after cycle 9 and after cycle 13 in the carboplatin-pemetrexed and the amivantamab plus carboplatin-pemetrexed groups, respectively. At the 12-month follow-up (cycle 17 day 1), the mean change from baseline on the PROMISE-PF total score was ████ ████ ███ █████ █████ in the amivantamab plus carboplatin-pemetrexed group and ████ ████ ███ ██████ ████ in the carboplatin-pemetrexed group, with a mean difference in change between amivantamab plus carboplatin-pemetrexed and carboplatin-pemetrexed groups of ████ ██████ ████ points.

Table 15: Summary of Key Efficacy Results From PAPILLON

ACP = amivantamab + carboplatin-pemetrexed; BICR = blinded independent central review; CI = confidence interval; CP = carboplatin-pemetrexed; EORTC QLQ-C30 = European Organization for the Research and Treatment of Cancer Quality of Life Questionnaire Core 30; HR = hazard ratio; HRQoL = health-related quality of life; NE = not estimable; OR = odds ratio; ORR = objective response rate; OS = overall survival; PFS = progression-free survival; PFS2 = progression-free survival after first subsequent therapy; SD = standard deviation.

Note: + = censored observation.

^aHR is from stratified proportional hazards model. HR less than 1 favours ACP.

^bP value is from a log-rank test stratified by ECOG PS (0 or 1) and history of brain metastases (yes or no).

^cHR is from proportional hazards mode. HR less than 1 favours ACP.

^dP value is from a log-rank test.

^eP value is from a logistic regression model stratified by ECOG PS (0 or 1) and history of brain metastases (yes or no).

^fMeans and 95% CIs are derived based on the mixed-effects model with repeated measures, in which the dependent variable is change from baseline in score, and independent variables are baseline, visit, treatment, and visit by treatment interaction as fixed effects, and individual patient as random effect.

^gNot adjusted for multiple comparisons. There is an increased risk of type I error (i.e., erroneously rejecting the null hypothesis).

Source: Details included in the table are from the sponsor’s Summary of Clinical Evidence.³¹

Harms

Refer to Table 16 for harms data.

Adverse Events

Overall, all 151 patients (100%) in the amivantamab plus carboplatin-pemetrexed group and 152 patients (98.1%) in the carboplatin-pemetrexed group experienced at least 1 TEAE. The most common TEAE was rash, which occurred in 81 patients (53.6%) in the amivantamab plus carboplatin-pemetrexed group and 12 patients (7.7%) in the carboplatin-pemetrexed group. Other common TEAEs with large differences between groups included dermatitis acneiform, in 47 patients (31.1%) in the amivantamab plus carboplatin-pemetrexed group and 5 (3.2%) in the carboplatin-pemetrexed group; hypoalbuminemia, in 62 (41.1%) patients in amivantamab plus carboplatin-pemetrexed and 15 (9.7%) patients in carboplatin-pemetrexed; peripheral edema, in 45 (29.8%) patients in amivantamab plus carboplatin-pemetrexed and 16 (10.3%) patients in carboplatin-pemetrexed; and infusion-related reaction, in 63 (41.7%) patients in amivantamab plus carboplatin-pemetrexed and 2 (1.3%) patients in carboplatin-pemetrexed.

Grade 3 and Higher

Overall, there were 114 patients (75.5%) in the amivantamab plus carboplatin-pemetrexed group and 83 patients (53.5%) in the carboplatin-pemetrexed group who reported at least 1 TEAE that was grade 3 or higher. The most common grade 3 or higher TEAE was neutropenia, occurring in 50 patients (33.1%) in the amivantamab plus carboplatin-pemetrexed group and 35 patients (22.6%) of the carboplatin-pemetrexed group. Other common grade 3 or higher TEAEs where there was a notable (> 5%) difference in incidence between groups were paronychia, occurring in 10 patients (6.6%) in the amivantamab plus carboplatin-pemetrexed group and no patients in the carboplatin-pemetrexed group; rash, which occurred in 17 patients (11.3%) in the amivantamab plus carboplatin-pemetrexed group and no patients in the carboplatin-pemetrexed group; and hypokalemia, occurring in 13 patients (8.6%) in the amivantamab plus carboplatin-pemetrexed group and 2 patients (1.3%) in the carboplatin-pemetrexed group.

Serious Adverse Events

Treatment-emergent SAEs were reported in 56 patients (37.1%) in the amivantamab plus carboplatin-pemetrexed group and 48 patients (31.0%) in the carboplatin-pemetrexed group. The most common treatment-emergent SAEs in the amivantamab plus carboplatin-pemetrexed group were pneumonia (6 patients [4.0%] versus 4 patients [2.6%] in the carboplatin-pemetrexed group), pneumonitis (4 patients [2.6%] and no patients in the carboplatin-pemetrexed group), and pulmonary embolism (4 patients [2.6%] in each group). All other SAEs in the amivantamab plus carboplatin-pemetrexed group had an incidence of less than 2%. The most common SAE in the carboplatin-pemetrexed arm was anemia (6 patients [3.9%] and 1 patient [0.7%] in amivantamab plus carboplatin-pemetrexed group).

Withdrawals Due to Adverse Events

Overall, from study initiation through the data cut-off, 36 patients (23.8%) in the amivantamab plus carboplatin-pemetrexed arm and 16 patients (10.3%) in the carboplatin-pemetrexed arm had TEAEs leading to discontinuation of at least 1 study treatment. Of the 36 patients in the amivantamab plus carboplatin-pemetrexed arm who discontinued any study treatment, 17 discontinued amivantamab (discontinuations were considered to be drug-related in 10 patients). Pneumonitis was the most common cause of discontinuation of amivantamab (4 patients [2.6%]), followed by dermatitis acneiform (2 patients [1.3%]). Thrombocytopenia (3 patients [1.9%]) and neutropenia (████████ ███████ were the most common reasons for treatment discontinuation of either carboplatin or pemetrexed in the carboplatin-pemetrexed arm. Most of the TEAEs leading to study treatment discontinuation occurred at a frequency of less than 2% in both treatment arms.

Mortality

There were 4 patients (2.6%) in the amivantamab plus carboplatin-pemetrexed group and 9 patients (5.8%) in the carboplatin-pemetrexed group who died due to an AE during the study, and 3 patients (2.0%) in the amivantamab plus carboplatin-pemetrexed group and 4 patients (2.6%) in the carboplatin-pemetrexed group who died due to an AE within 30 days of the last dose. Excluding the crossover phase, there were 7 patients (4.6%) and 4 (2.6%) patients in the amivantamab plus carboplatin-pemetrexed and carboplatin-pemetrexed groups, respectively, who died due to an AE within 30 days of the last dose. Six out of the 7 AEs in the amivantamab plus carboplatin-pemetrexed group leading to death occurred within the first 90 days of cycle 1 day 1, compared with 1 death in the carboplatin-pemetrexed group. In the amivantamab plus carboplatin-pemetrexed group, these 6 deaths included 1 event each of cardiorespiratory arrest, sepsis, stroke, death, COVID-19, and COVID-19 pneumonia. Each of these deaths was reported to be a unique event in patients with comorbidities with no clear pattern of etiology identified.

Notable Harms

The notable harms identified for this review were rash and infusion reactions. As noted previously, rash was the most common AE, occurring in 81 (53.6%) patients in the amivantamab plus carboplatin-pemetrexed group and 12 (7.7%) patients in the carboplatin-pemetrexed group. Under the grouping skin and subcutaneous disorders, there were ███ ████████ ███████ in the amivantamab plus carboplatin-pemetrexed group and ██ ███████ patients in the carboplatin-pemetrexed group with such an event. Infusion-related reactions were reported in 63 (41.7%) patients in the amivantamab plus carboplatin-pemetrexed group and 2 (1.3%) patients in the carboplatin-pemetrexed group.

Table 16: Summary of Harms Results From PAPILLON

AE = adverse event; CP = carboplatin-pemetrexed; ILD = interstitial lung disease; SAE = serious adverse event.

Note: Subjects are counted only once for any given event, regardless of the number of times they actually experienced the event. Adverse events are coded using MedDRA Version 25.0. Deaths due to COVID-19 are not included in the other causes of death (i.e., subjects are only counted in one row).

^aSubjects who died during the Crossover phase are included.

^bSubjects who died within 30 days will be counted in deaths during study as well.

^cSubjects who died during the Crossover phase are excluded.

Source: Details included in the table are from the sponsor’s Summary of Clinical Evidence.³¹

Table 17: Summary of Discontinuations Due to AEs

AE = adverse event; CP = carboplatin-pemetrexed; ILD = interstitial lung disease; SAE = serious adverse event.

Source: Details included in the table are from the sponsor’s Summary of Clinical Evidence.³¹

Critical Appraisal

Internal Validity

The methods used in the PAPILLON trail to generate the randomization sequence and conceal the allocation until treatment assignment were adequate. Baseline characteristics were relatively well balanced between groups, which suggests that the randomization was successful. The few baseline imbalances that were identified (e.g., location of metastasis at screening, time since diagnosis, histology, and prior radiotherapy) were not believed to have an important impact on the efficacy and harm evaluation. There were some imbalances observed in the use of some concomitant medications. However, the clinical experts consulted on this review did not believe these differences would result in a differential impact on treatment effect.

Following randomization, the study was not blinded. Patients’ knowledge of their assigned treatment may bias the measurement of subjective outcomes, including patient-reported outcomes such as HRQoL. Assessments of disease progression and tumour response are based on subjective interpretation of radiographic images and clinical evaluation. However, both PFS and ORR were ascertained by BICR, which would mitigate the impact of bias in measurement of those outcomes. Patients’ knowledge of treatment assignment also has potential to bias assessment of harms; for example, diarrhea is a relatively common adverse effect of EGFR inhibitors and perhaps more subjective in its reporting.

A power calculation was performed; however, no rationale or source was provided for the underlaying assumptions of expected effect size. A hierarchical fixed-sequence testing strategy was used to control the overall type I error rate for hypotheses testing of primary outcome, PFS (by BICR), and key secondary end points, ORR and OS. Assessment of other important outcomes, such as PFS2, were not adjusted for multiple comparisons and therefore any analyses reported, and any P values, should be considered supportive in nature.

There was a considerable amount of missing data for the patient-reported outcomes, such as the EORTC QLQ-C30 and PROMIS PF. The number of patients available to complete the measure diminished with each cycle. After cycle 9, the number of patients available to complete assessments had ███████ ██ ████ ████ ███ after cycle 9 and after cycle 13 in the carboplatin-pemetrexed and the amivantamab plus carboplatin-pemetrexed groups, respectively. Only █████████ of the original randomized population (██ ██ ███ █████████ was available to provide assessments at the 12-month time point (cycle 17 day 1) in the amivantamab plus carboplatin-pemetrexed group and only slightly more than ███ ███ ██ ███ ████████) in the carboplatin-pemetrexed group. The sponsor employed mixed-effects model for repeated measures to estimate the change from baseline, an approach that assumes data to be missing at random. Although most of these missing data could be accounted for by patients leaving the study or crossing over to amivantamab monotherapy, it means that the remaining patients available to provide assessments were all responders, which may bias results in both groups, and does not suggest data were missing at random.

OS results are a reflection of both the study drug and subsequent treatments. Switching patients in the control group whose disease is not responding to their carboplatin-pemetrexed regimen to amivantamab monotherapy may have improved outcomes in these patients. Thus, there is a risk of bias toward the null. Given the large number of patients who switched from carboplatin-pemetrexed to amivantamab, the potential for underestimation of the treatment effect may be significant. For example, of the 155 patients in the carboplatin-pemetrexed group, ███ had a PFS event (death or progression), ███ of them had progressive disease, and 71 of 107 (66.4%) patients with disease progression switched to amivantamab monotherapy. On average, patients who switched were exposed to amivantamab for nearly as much time as they were exposed to carboplatin-pemetrexed, suggesting that switching may have had a significant impact on OS results, as well as influencing subsequent assessment of PFS2. In an effort to account for this potential bias, the sponsor performed several post hoc sensitivity analyses of OS results, including IPCW, TSE, and RPSFT. The OS results were similar in the IPCW (adjusted ██ ██ ████ ████ ███ █████ ████) and TSE (adjusted HR of ████ ████ ███ █████ ████) analyses and differed slightly from those in the RPSFT (adjusted HR ██ ███ ████ ███ █████ ████). CIs were wide in all cases (i.e., including small effects) and crossed the null for the RPSFT model. A key assumption of the IPCW and TSE approaches is that there is no unmeasured confounding, while a key assumption of the RPSFT model is that there is a common treatment effect. These assumptions cannot be objectively verified, and, if deemed implausible, would undermine the validity of the estimated effects. Several potential confounding variables were considered in the IPCW and TSE analyses. However, numerous subjective factors may influence which patients cross over (versus those who do not), which makes it difficult to fully justify the assumption that all potential confounding has been accounted for.⁷⁸ Similarly, in relation to the RPSFT model, it might not be clinically reasonable to assume the treatment effect to be equivalent regardless of whether it is received at randomization or postprogression. Ultimately, all of these sensitivity analyses are considered to provide supportive information about the true treatment effect of amivantamab plus carboplatin-pemetrexed compared to carboplatin-pemetrexed but are limited due to their dependence on unverifiable assumptions.

External Validity

The clinical experts believed that the patients enrolled in the PAPILLON trail were generally representative of the patients who would receive this drug in Canada. The clinical experts consulted for this review noted that, although inclusion in the PAPILLON trail was restricted to patients with an ECOG performance status of 0 or 1, they would likely also offer amivantamab plus carboplatin-pemetrexed to patients with an ECOG of 2.

The clinical experts believed OS to be an important outcome in assessing response in these patients; however, these data are still immature. Specifically, there were a total of 70 (about 23%) patients who had died at the time of the data cut-off (median follow-up of 14.92 months), while the final analysis of OS is planned to be conducted 48 months after the first patient was randomized, with an anticipated 210 deaths. The clinical experts believed that PFS is an appropriate primary outcome, particularly since OS is difficult to assess in clinical trials in this setting, given the tendency to allow patients whose disease progressed on their control treatment, to cross over to the treatment group.

The background therapy, carboplatin-pemetrexed, is appropriate for these patients, according to the clinical experts consulted for this review. The clinical experts suggested that some clinicians might prefer to use cisplatin over carboplatin, which was felt to be reasonable, given cisplatin is anticipated to have a similar side effect profile and efficacy as carboplatin.

GRADE Summary of Findings and Certainty of the Evidence

Methods for Assessing the Certainty of the Evidence

For pivotal studies and RCTs identified in the sponsor’s systematic review, GRADE was used to assess the certainty of the evidence for outcomes considered most relevant to inform the CDA-AMC expert committee deliberations, and a final certainty rating was determined as outlined by the GRADE Working Group.^79,80

• High certainty: We are very confident that the true effect lies close to that of the estimate of the effect.

• Moderate certainty: We are moderately confident in the effect estimate. The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. We use the word “likely” for evidence of moderate certainty (e.g., “X intervention likely results in Y outcome”).

• Low certainty: Our confidence in the effect estimate is limited. The true effect may be substantially different from the estimate of the effect. We use the word “may” for evidence of low certainty (e.g., “X intervention may result in Y outcome”).

• Very low certainty: We have very little confidence in the effect estimate. The true effect is likely to be substantially different from the estimate of effect. We describe evidence of very low certainty as “very uncertain.”

Following the GRADE approach, evidence from RCTs started as high-certainty evidence and could be rated down for concerns related to study limitations (which refers to internal validity or risk of bias), inconsistency across studies, indirectness, imprecision of effects, and publication bias.

When possible, certainty was rated in the context of the presence of an important (nontrivial) treatment effect; if this was not possible, certainty was rated in the context of the presence of any treatment effect (i.e., the clinical importance is unclear). In all cases, the target of the certainty-of-evidence assessment was based on the point estimate and where it was located relative to the threshold for a clinically important effect (when a threshold was available) or to the null. The target of the certainty-of-evidence assessment was the presence or absence of a clinically important effect based on thresholds informed by the clinical experts consulted for this review for PFS and OS. The literature-based MID of 10 points was used for the EORTC QLQ-C30 GHS.^29,30 This MID has been estimated for within-group changes and was applied in the absence of an estimate of the MID for a between-group difference. The target of the certainty of evidence was the presence or absence of any (non-null) effect for the ORR because a threshold for a clinically important between-group difference could not be estimated.

Results of GRADE Assessments

Table 2 presents the GRADE summary of findings for amivantamab plus carboplatin-pemetrexed versus carboplatin-pemetrexed.

Long-Term Extension Studies

Contents in this section have been informed by materials submitted by the sponsor. The following have been summarized and validated by the CDA-AMC review team.

No long-term extension studies were available at the time of filing the application.

Indirect Evidence

The sponsor submitted an ITC report that included analyses that used individual patient data from the PAPILLON trail and real-world databases and IPTW methods to reduce the risk of bias due to confounding. This was considered by the review team as a nonrandomized study and is reported in the Studies Addressing Gaps section.

This section includes the contents of the ITC report submitted by the sponsor, which was considered by the review team to include a nonrandomized study, as it utilized individual patient data for each arm in the comparisons.

Objectives for the Nonrandomized Study

The objective of this section is to summarize and critically appraise the methods and findings of the nonrandomized study submitted by the sponsor. The sponsor submitted a nonrandomized study that assessed the relative effectiveness between amivantamab plus carboplatin-pemetrexed (assessed in the PAPILLON study), and other common treatment options used in real-world clinical practice (EGFR TKI monotherapy and platinum chemotherapy plus IO).

The sponsor stated that comparisons of amivantamab plus carboplatin-pemetrexed to EGFR TKIs, and IO, in head-to-head clinical studies of patients with EGFR exon 20ins mutated advanced NSCLC in the first-line setting, have not been conducted.

Description of the Nonrandomized Study

The sponsor’s analysis included a nonrandomized study using propensity score weighting to reduce the impact of bias due to confounding. The analyses used individual patient data from the PAPILLON trail for amivantamab plus carboplatin-pemetrexed and from real-world databases COTA and ConcertAI for EGFR TKI monotherapy and platinum chemotherapy plus IO.

Nonrandomized Study

Objectives

The sponsor submitted a nonrandomized study that assessed the relative effectiveness between amivantamab plus carboplatin-pemetrexed (assessed in the PAPILLON study) and other common treatment options used in real-world clinical practice (EGFR TKI monotherapy and platinum chemotherapy plus IO). The comparative clinical efficacy of chemotherapy alone versus EGFR TKIs or chemotherapy alone versus IO plus chemotherapy was not the purpose of assessment.

Study Selection Methods

The sponsor did not submit a predetermined protocol or undertake a systematic review to identify studies for inclusion in the analysis. There was no evidence that the sponsor employed a literature search or utilized a priori selection criteria to select the studies for the analysis. There was no description of data extraction methods or appraisal of risk of bias of the contributing data sources.

Data Sources

PAPILLON Study

Individual patient data for amivantamab plus carboplatin-pemetrexed consisted of data from 1 arm of the PAPILLON trial. the PAPILLON trail is an ongoing, phase III, open-label, parallel, multicentre RCT that enrolled 308 patients. Treatment was administered in cycles of 21 days (3 weeks) until patient discontinuation or disease progression. This study is described in detail in other sections of this report.

Real-World Database: COTA

COTA’s Real-World Evidence database comprises longitudinal, Health Insurance Portability and Accountability Act (HIPAA)-compliant data pertaining to the diagnosis, clinical management, and outcomes of patients with cancer in the US. Data are abstracted from the electronic health records of health care provider sites, representing diverse treatment settings, including academic, for-profit, community sites, and hospital systems. COTA leverages both human abstraction and technologic methods to transform structured and unstructured data into a standard data model. Abstractors receive in-depth training on oncology, data abstraction, and electronic medical record navigation and are required to achieve an acceptable accuracy rate on abstraction of test records against a gold standard before abstracting in COTA’s production database. The COTA deidentified retrospective dataset includes data on a cohort of patients with NSCLC and confirmed EGFR exon 20ins mutation from 5 unique sites of care; 79% of patients were treated at academic medical centres and 21% were treated in the community oncology setting. Patients are excluded from the dataset if they:

• are younger than aged 18 years at the time of diagnosis

• do not have the malignancy of interest

• are not evaluated at the accessible provider site for the malignancy

• have concurrent primary malignancies

• have no date of diagnosis in the electronic health record

• have no clinician note available in the electronic health record

• are diagnosed with a malignancy after the diagnosis of the malignancy of interest before evaluation for the malignancy of interest at the accessible provider site

• are initially misdiagnosed and treated for another malignancy but are later confirmed to have the malignancy of interest

• are on therapy, active surveillance, or observation for a malignancy diagnosed before the diagnosis of the malignancy of interest at the time of diagnosis of the malignancy of interest

• have metastatic malignancy, including leukemias and multiple myeloma, diagnosed before the malignancy of interest

• were only seen once at the provider site for the malignancy of interest

• had low-grade (inclusive of grades 1 and 2) neuroendocrine histology.

Real-World Database: ConcertAI

ConcertAI sources clinical data from multiple partnerships spanning a broad range of US locations, regardless of electronic medical record (EMR) provider. This includes data from various organizations aggregating data from oncology practices to create rapid learning systems for providers to improve patient care. ConcertAI’s data source partners receive copies of the practice’s complete EMR, including unstructured notes and scanned documents attached to the EMR. ConcertAI’s data partners are EMR agnostic and have customer-implementation teams assigned to each practice for procurement of structured and unstructured information. ConcertAI processes all available oncology data (patient documents, physician notes, structured EMR data) into an Observational Medical Outcomes Partnership (OMOP)-based Common Data Model designed for easier query and analyses. The split of organization type across the data product varies by indication (with an overall average of approximately 70% community and approximately 30% academic). The geographic split across the data product also varies by indication (with an overall average of approximately 15% northeast, approximately 25% midwest, approximately 40% south, and approximately 20% west). Patients were extracted from both the Patient360 and Genome360 data samples and pooled into a single “ConcertAI” cohort.

Patient360

ConcertAI screens patients for curation based on a structured International Classification of Diseases (ICD)-10 diagnosis for the indication of interest and, at a minimum, the diagnosis year. Once a screening list has been generated with these 2 criteria, ConcertAI randomly pulls in patients by first curating and confirming diagnosis and diagnosis date (based on pathology reports and other unstructured data). All patients must be older than 18 years of age at first diagnosis date. Patients are selected for inclusion in data products through the quality-control processes, identifying whether there are conflicts in the patient record, using structured or both structured and curated data, depending on the data product. Patients were not filtered by ConcertAI based on data completeness to avoid introduction of selection bias.

Genome360

Apart from the criteria used to select patients for Patient360 curation, as detailed previously, patients were further selected for Genome360 only if the patient record contained a somatic NGS report, which is determined through natural language processing algorithms and further confirmed by manual screening.

Table 18: Analysis Populations of the Included Studies in the Nonrandomized Study

CP = carboplatin-pemetrexed; ECOG = Eastern Cooperative Oncology Group; exon 20ins = exon 20 insertion; IO = immunotherapy; NSCLS = non–small cell lung cancer; RECIST = Response Evaluation Criteria in Solid Tumours; TKI = tyrosine kinase inhibitor; ULN = upper limit of normal; VEGFi = vascular endothelial growth factor inhibitor.

Source: Sponsor’s ITC.⁸¹

Feasibility Assessment

The sponsor did not provide a feasibility assessment or a systematic review of the literature.

Analysis Methods

Outcomes Analyzed

Overall Survival

In both data sources, OS was defined as the time between the index date and death from any cause. Patients still alive at the end of the study period were censored at the study end date. For patients who were alive or for whom their vital status was unknown, this interval was censored at the date the patient was last known to be alive.

Progression-Free Survival

PFS was defined as the time from index date to date of disease progression (as recorded in the database) or death (from any cause), whichever occurred first. Disease progression in the PAPILLON trail assessed by BICR was used. A definition for “progression” in the real-world data sources was not provided, but it was noted that the databases did not have a systematic capture of progressive disease or RECIST v.1.1 criteria.

Real-World PFS

As the real-world COTA and ConcertAI databases do not have a systematic capture of progressive disease or RECIST v1.1. criteria, which are not typically captured in the real-world, the start of subsequent treatment was also considered a progression event. Therefore, RW-PFS for COTA and ConcertAI was defined as the time from index date to date of disease progression, start date of subsequent treatment, or death (from any cause), whichever occurred first. Specifics of the estimation of RW-PFS in the PAPILLON trail were not provided.

Time to Next Treatment

In both data sources, TTNT was defined as the time between the index date and the start date of subsequent treatment or death, whichever came first. Patients without a record of subsequent therapy were censored at the date of last contact.

Nonrandomized Study Methods

Baseline Covariates

The identification of potential treatment effect modifiers and prognostic factors was based on an unpublished systematic review with expert input. Thereafter, selected variables for adjustment included prognostic factors that were mutually reported in both the PAPILLON trail as well as the COTA and ConcertAI real-world databases. Prognostic factors were assessed at baseline in the real-world databases, defined as 30 days before to 15 days after the start of therapy, and at baseline in the PAPILLON trial. It was not clear whether some information was missing for the covariate data in the sponsor’s analyses or how that might have been handled. The final list and ranking of covariates are outlined in Table 19, but it should be noted that not all covariates were used in all of the sponsor’s analyses, as described later in this report. Of the 9 factors identified as important for population adjustment, prior major surgery in the 4 weeks before index date was included as an exclusion criterion in the patient selection process instead of a prognostic factor for adjustment in the statistical models. While a sensitivity analysis of the full model including all 8 factors was conducted, the primary analysis of the base case only adjusted for the top 4 ranked factors.

Table 19: Final Ranking of Covariates and Usability

ECOG = Eastern Cooperative Oncology Group.

^aMetastatic disease time windows were anchored 30 days before to 15 days after the date of diagnosis of locally advanced or metastatic cancer.

^bAny time before the index date.

^cMajor surgery includes but is not limited to brain lobectomy, brain operation, brain tumour operation, cancer surgery, craniectomy, cranioplasty, craniotomy, lung lobectomy, lung neoplasm surgery, lung operation, lymphadenectomy, lymphadenopathy mediastinal, pleural decortication, pneumonectomy, pulmonary resection, thoracic operation, and thoracotomy.

^dThis covariate was included as an exclusion criterion in the present analysis instead of a prognostic factor for adjustment.

Source: Sponsor’s ITC Report.⁸¹

Statistical Analysis

To improve the balance of baseline confounding variables between the PAPILLON trial and real-world comparator cohorts, IPTW was used to adjust for selected baseline characteristics, following recommendations of the National Institute for Health and Care Excellence Decision Support Unit Technical Support Document.⁴⁸

IPTW with ATT weighting was used to adjust for baseline differences between the the PAPILLON trail and comparator cohorts. This method allowed the real-world cohorts to be reweighted to align with the PAPILLION cohort. It is important to note that IPTW with ATT weights may become unstable when the comparator arm sample size is small.

Propensity scores were derived with a multivariable logistic regression model using a binary treatment variable as the dependent variable and the following covariates: functional status ECOG at index date, history of brain metastases, history of liver metastases, and age at index date. These were ranked as the most important prognostic factors for adjustment. The estimated propensity scores were then used to derive weights for each patient, where patients in the PAPILLON trail group received a weight of 1, while patients in the real-world comparator groups were reweighted based on the probability of receiving treatment.

Population differences between the PAPILLON trial population and the comparator cohort populations were assessed before and after weighting, using SMDs. The sponsor considered an absolute SMD between 0 and 0.1 to be a small difference. The balance of covariates was also visually displayed by density plots of the propensity scores before and after weighting.

The ESS was calculated to reflect the impact of weighting on the available information in the individual patient data.

Outcome Analysis Methods

The comparative effectiveness of amivantamab plus carboplatin-pemetrexed versus the platinum plus IO and EGFR TKI comparator cohorts was determined for the following outcomes: OS, PFS, RW-PFS, and TTNT. Estimates of comparative effectiveness were derived for both the unadjusted comparison (i.e., amivantamab plus carboplatin-pemetrexed versus comparative treatment before IPTW) and the adjusted comparison (i.e., with IPTW).

A Cox proportional hazards model (with weights applied for the adjusted comparison) was used to derive an HR and its respective 95% CI. The variance was estimated using a robust sandwich estimator to account for the uncertainty in estimation of the weights.

Sensitivity Analysis

As summarized in Table 20, a number of sensitivity analyses were conducted using alternative weighting formulas. Average treatment effect in the control weighting and average treatment effect in the overlap (ATO) weighting were employed to improve the covariate balance and mitigate the limitations associated with the reduction in the ESS from the ATT weighting. Multivariable regression analysis using the same covariates was also conducted to assess the impact of an alternate analytical approach to adjusting for baseline differences between the 2 cohorts on the study’s results. All analyses were performed for all outcomes of interest (OS, PFS, RW-PFS, and TTNT).

IPTW with ATO weights were determined, where patients in both cohorts are reweighted to estimate the treatment effect in the overlap population between groups. ATO weights are less sensitive to extreme weights and can be considered when outliers may be a concern. ATO weighting may be beneficial when overlap between both cohorts is poor. ATO weights were derived, where patients in the PAPILLON trial group received a weight equal to the propensity score while the patients in the comparator treatment cohorts received a weight equal to 1 – propensity score. This estimated the treatment effect in the overlap population of the PAPILLON trial and comparator treatment cohorts.

A third sensitivity analysis was conducted using multivariable regression. This method requires overlap in covariate distribution, similar to IPTW. Unlike reweighting methods, however, regression models estimate the conditional average treatment effect and require a large sample size compared to the number of covariates in the model.

For each of the analytical approaches evaluated (e.g., ATT, average treatment effect in the control, ATO, and multivariable regression), adjusted comparisons were also conducted by sequentially adjusting for the remaining prognostic factors in order of importance, until the final model contained all prognostic covariates (referred to as the “full model”). The selected covariates were also adjusted for in the model to produce doubly robust results, as an additional sensitivity analysis.

Assessment of Proportional Hazards

Appropriateness of the proportional hazards assumption for survival outcomes was assessed based on visual inspection of the log-cumulative hazard plot, visual inspection of the Schoenfeld residuals plot, and performance of the Grambsch-Therneau test (with a P value less than 0.05 considered to indicate a violation of the assumption). For all analyses, patient numbers over time were reported.

Table 20: Sensitivity Analysis

ATC = average treatment effect in the control; ATO = average treatment effect in the overlap; ATT = average treatment effect in the treated; IPTW = inverse probability of treatment weighting; OS = overall survival; PFS = progression-free survival; RW- PFS = real-world progression-free survival; TTNT = time to next treatment.

Note: “Unchanged” indicates that the analytic specification was unchanged from the main analyses.

^aECOG at index date, history of brain metastases, history of liver metastases, age at index date.

Source: Sponsor’s ITC Report.⁸¹

Results of the Nonrandomized Study

Selection of Participants

Of the 308 patients in the PAPILLON trial, the final cohort consisted of 153 patients that received amivantamab plus carboplatin-pemetrexed. Of the ██████ patients within the combined COTA and ConcertAI databases, the final RW comparator cohorts consisted of ██ ███ ██ patients who received first-line EGFR TKI or platinum plus IO treatment, respectively (Figure 3).

Figure 3: Patient Selection for the Nonrandomized Study [Redacted]

Source: Sponsor’s ITC Report⁸¹ — ITCs with individual patient data.

Treatment Regimens Received in the Real-world Cohorts

Following the application of inclusion criteria to align with the PAPILLON trial, sufficient sample size for comparison was only available for the EGFR TKI alone and platinum plus IO comparator cohorts, which were thus included as comparator groups in the analysis. A breakdown of the specific treatment regiments patients received in the EGFR TKI and platinum plus IO cohorts are included in Table 21 and Table 22.

Table 21: Specific Treatment Regimens Patients Received in the EGFR TKI Cohort

TKI = tyrosine kinase inhibitor.

Sources: Sponsor’s ITC Report⁸¹ and indirect treatment comparisons with individual patient data.

Table 22: Specific Treatment Regimens Patients Received in the Platinum Plus IO Cohort

IO = immunotherapy.

Source: Sponsor’s ITC Report.⁸¹

Summary of Baseline Population Characteristics

The population differences in baseline covariates between the PAPILLON trail and real-world comparator cohorts before and after adjustment with ATT weighting are shown in Table 23 and Table 24 for EGFR TKI and platinum + IO.

For the PAPILLON trial versus EGFR TKI comparison, before weighting, moderate differences (absolute SMD ████ ███ ████) were observed for ECOG at index date and history of smoking, and substantial differences (absolute SMD █████ seen for history of liver metastases, age, Asian race, sex, and history of other metastases. The PAPILLON trial cohort has a higher proportion of participants with ECOG PS 1, no history of liver metastases, history of other metastases, male sex, less than 65 years of age, and of Asian race. In the primary analysis, base-case ATT weighting balanced (SMD ████) all 4 covariates between cohorts, but race, sex, and history of metastases remained imbalanced. In the full model, however, all 8 of the included factors had absolute SMDs████, indicating lack of balance, with moderate differences observed for 6 factors (████ ██ █████ █████ ███████ ██ █████ ███████████ ████ ███████ ██ ████████ ████ ███████ ██ █████ ██████████), and substantial differences observed for the remaining 2 factors (███████ ██ █████ ███████████ █████ █████. The resulting ESS in the EGFR TKI cohort was ████ ███ ████ for the base-case and full model, respectively, compared to the original 18 observations.

For the PAPILLON trial versus platinum plus IO comparison, before weighting, substantial (absolute SMD ████) differences were observed for ████ ██ █████ █████ ███████ ██ █████ ███████████ ███████ ██ █████ ███████████ ████ █████ █████ ███ ███████ ██ █████ ███████████ ████ ███ ████████ ██████ ██████ █ ██████ ██████████ ██ ████████████ ████ ████ ██████ ████ ██ ███████ ██ █████ ██ █████ ███████████ ███████ ██ █████ ███████████ ████████ ████ ████ ██ █████ ██ ████ ███ ██ █████ ████. In the primary analysis, ATT weighting reduced the proportion of categories with absolute SMDs ████ ████ ████ ██████ ██ ███ ████) in the base case; however, differences remained in █████ ███████ ████████ ████ ███ ███████ ██ ██████████. The balance between populations in the full model, on the other hand, remained the same (6 of 8 factors balanced) following weighting, although moderate differences were observed for 3 factors (████ ██ █████ █████ ███████ ██ █████ ███████████ ███████ ██ ███████) and substantial differences observed for the remaining 3 factors (███████ ██ █████ ███████████ █████ █████ ███████ ██ █████ ██████████). The resulting ESS in the platinum plus IO cohort was █████ ███ ████ for the base case and full model, respectively, compared to the original ██ observations.

Table 23: Overview of Baseline Characteristics for Amivantamab Plus Carboplatin-Pemetrexed (PAPILLON) Versus EGFR TKI (COTA, ConcertAI) Before and After Adjustment With IPTW Using ATT

ATT = average treatment effect in the treated; ECOG = Eastern Cooperative Oncology Group; ESS = effective sample size; IPTW = inverse probability of treatment weighting; SMD = standardized mean difference; TKI = tyrosine kinase inhibitor.

Note: Grey cells indicate that these variables were not adjusted for in the given analysis.

Source: Sponsor’s ITC Report.⁸¹

Table 24: Overview of Baseline Characteristics for Amivantamab Plus Carboplatin-Pemetrexed (PAPILLON) Versus Platinum Plus IO (COTA, ConcertAI) Before and After Adjustment With IPTW Using ATT