Reimbursement Review

Durvalumab (Imfinzi) and Tremelimumab (Imjudo)

Sponsor: AstraZeneca Canada Inc.

Therapeutic area: Metastatic non–small cell lung cancer (NSCLC)

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; NSCLC = non–small cell lung cancer; q.3.w. = every 3 weeks; q.4.w. = every 4 weeks.

^aTremelimumab to be used in combination with durvalumab. Tremelimumab is not intended for monotherapy use.

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

Table 2: Recommended Regimen and Dosing for Durvalumab in Combination With Tremelimumab and Platinum-Based Chemotherapy for Treatment of Metastatic NSCLC

NSCLC = non–small cell lung cancer.

^aPatients with a body weight of 30 kg or less must receive weight-based dosing, until the weight increases to greater than 30 kg.

Source: Imfinzi product monograph.²

Table 3: Recommended Dosing Schedule for Durvalumab in Combination With Tremelimumab and Platinum-Based Chemotherapy for the Treatment of Metastatic NSCLC

NSCLC = non–small cell lung cancer.

^aThe dosing interval changes from every 3 weeks to every 4 weeks starting at cycle 5.

^bContinue until disease progression or unacceptable toxicity.

^cIV infusion over 60 minutes.

^dFor patients with nonsquamous NSCLC treated with pemetrexed and carboplatin or cisplatin: pemetrexed maintenance therapy from week 12 until disease progression or unacceptable toxicity is based on physician discretion.

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

Introduction

Lung cancer is the leading cause of cancer-related death in Canada; it is heterogeneous in nature, comprising several different disease subtypes categorized by histology, staging, and mutation driver status.³ Non–small cell lung cancer (NSCLC) is the most common type, representing approximately 89% of all lung cancer diagnoses in Canada.⁴ NSCLC can be categorized into histologic subtypes: squamous-cell carcinoma (17% to 27% of NSCLC cases) and non–squamous cell carcinoma (73% to 83% of NSCLC cases).^3,5 At early stages, NSCLC may be asymptomatic,³ with symptoms developing only when the cancer has become more advanced and is no longer amenable to curative-intent therapy. Presenting symptoms can sometimes be nonspecific (common manifestations include coughing, chest pain, hemoptysis, fatigue, weight loss, dyspnea, hoarseness, and recurring infections with bronchitis and pneumonia), contributing to delays in diagnosis.^3,5,6 Screening programs for NSCLC are being established throughout Canada, but many patients who would qualify for NSCLC screening live in jurisdictions where it is not yet available. Although EGFR mutations and ALK aberrations have been identified as oncogenic driver mutations,⁷ which collectively occur in approximately 25% of patients,⁸ an estimated 74.5% of NSCLC cases do not have EGFR mutations or ALK aberrations.

Nearly half (48.4%) of NSCLC cases are metastatic (stage IV).^3,8 For the majority of patients with metastatic NSCLC whose tumours lack actionable genomic alterations, treatment is selected based on tumour histology (squamous or nonsquamous); PD-L1 expression, which is predictive of response to immune checkpoint inhibitors (ICIs); and patient symptom burden (as measured by Eastern Cooperative Oncology Group Performance Status [ECOG PS]), comorbidities, and preferences.^9-11 In recent years, new treatment options include ICI-based therapies (either as monotherapy or in combination with chemotherapy), which are now the standard first-line treatment for metastatic NSCLC without targetable genomic alterations.^10,11 For patients with any or unknown tumoural PD-L1 expression, first-line options include pembrolizumab plus platinum-based chemotherapy; nivolumab and ipilimumab plus platinum-based chemotherapy; or cemiplimab plus platinum-based chemotherapy.^10,11 Monotherapy with pembrolizumab or cemiplimab may be selected as first-line treatment for patients with high PD-L1 expression (PD-L1 expression in ≥ 50% of tumour cells [TCs]).^10,11 Cemiplimab (with or without platinum-based chemotherapy) is not currently funded in Canada.¹² Platinum-based chemotherapy alone is also a potential first-line treatment option; however, the clinical experts consulted by Canada’s Drug Agency (CDA-AMC) noted that this would be reserved for patients with significant contraindications to immunotherapy for whom the perceived risks of ICI-based treatment outweigh any potential benefits. The clinical experts consulted for this review indicated that immuno-oncology treatment is considered the backbone of therapy for the first-line treatment of patients with metastatic NSCLC without targetable oncogenic aberrations. The goals of therapy for patients with metastatic NSCLC, as identified by the clinical experts, include prolonging survival, extending the time before disease progression, decreasing cancer-related symptoms, and maintaining or improving quality of life (QoL).

Tremelimumab in combination with durvalumab was previously reviewed by CDA-AMC for the first-line treatment of adult patients with unresectable hepatocellular carcinoma who require systemic therapy. The final recommendation issued in November 2023 was “reimburse with clinical criteria and/or conditions.”

The objective of this report is to review and critically appraise the evidence submitted by the sponsor on the beneficial and harmful effects of durvalumab (50 mg/mL concentrate for IV infusion) in combination with tremelimumab (20 mg/mL concentrate for IV infusion) and platinum-based chemotherapy in the first-line treatment of metastatic NSCLC with no sensitizing EGFR mutations or ALK genomic tumour aberrations.

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 our call for input and from clinical experts consulted by for the purpose of this review.

Patient Input

A joint patient group submission was received from the Canadian Cancer Survivor Network (CCSN), Lung Cancer Canada (LCC), and the Lung Health Foundation (LHF). CCSN is a national network of patients, families, survivors, friends, community partners, funders, and sponsors who have come together to promote the best standard of care (SOC), including early diagnosis, timely treatment and follow-up care, and support for patients with cancer, and regarding issues related to survivorship or quality of end-of-life care. LCC is a registered national charitable organization that serves as Canada’s leading resource for lung cancer education, patient support, research and advocacy. The LHF is a registered charity that assists and empowers people living with or caring for others with lung disease. It is a recognized leader, voice, and primary resource in the prevention and control of respiratory illness, tobacco cessation and prevention, and the effects on lung health.

CCSN, LCC, and LHF collectively produced a survey and circulated it among their networks. The survey was disseminated through the 3 organizations’ social media platforms, as well as CCSN’s monthly newsletter to gather responses from August 1, 2024, to the date of writing of the patient group submission. LCC also conducted 1 interview on October 8, 2024, with a patient who was a part of the POSEIDON trial.

Respondents from a previous survey and submission on durvalumab in combination with chemotherapy as neoadjuvant treatment, followed by durvalumab as monotherapy after surgery indicated for the treatment of patients with NSCLC with resectable tumours (≥ 4 cm and/or node-positive) and no known EGFR or ALK rearrangements shared their experiences with lung cancer and durvalumab. The following symptoms associated with lung cancer and affecting QoL were reported: fatigue; pain in chest, shoulder, back or arms; shortness of breath; infections; pneumonia and/or bronchitis; and loss of weight, appetite, hair, and teeth. Chemotherapy, immunotherapy, targeted therapy, radiation, surgical therapy, and participation in clinical trials were identified as current treatment options by respondents. The respondents emphasized that these regimens were generally well tolerated; however, they described several side effects (e.g., joint and muscle pain, fatigue, diarrhea, neuropathy, weight loss, anemia, nausea, vomiting, constipation, migraines, change in vision and hearing, and forgetfulness). One respondent reported doing well until a CT scan showed disease progression.

Fatigue was identified by the respondents as the main side effect of durvalumab. In comparison to other therapies, at least 2 respondents indicated that symptom management was much better with durvalumab; however, there was little to no difference in side effects, ease of use, and disease progression.

Respondents noted several unmet needs, such as better mental health support, availability of immunotherapy for a longer duration, access to counselling, and help with travel costs associated with therapy. The respondents highlighted the following outcomes as important: disease management, prolonged life, a cure, QoL, delayed onset of symptoms, easy-to-use medication, reduction in side effects, and access to new options for treatment.

The patient from the POSEIDON trial reported that it took a while after diagnosis before they could begin receiving treatment. The physician who treated the patient confirmed that the patient received durvalumab and tremelimumab in combination with platinum-based chemotherapy. The patient started with first-line chemotherapy and received durvalumab simultaneously. While on chemotherapy, the patient received appropriate care and dealt with minimal side effects, including mild diarrhea, bone pain, itching, and tiredness, particularly when receiving chemotherapy. While on durvalumab, the patient noted temporary side effects such as occasional diarrhea, itching, and hot flashes. The patient emphasized that they were still able to carry out regular activities during treatment.

Clinician Input

Input From Clinical Experts Consulted for This Review

The clinical experts consulted by CDA-AMC identified several goals of first-line treatment of metastatic NSCLC: improving survival, extending the time before disease progression, decreasing cancer-related symptoms, and improving QoL. The experts acknowledged that currently available first-line treatment options for metastatic NSCLC have been associated with a survival benefit compared to chemotherapy alone but noted that there remains an unmet need among patients with low or negative PD-L1 expression, including options that have lower rates of immune-related adverse events (AEs). Clinical experts recognized that there is a significant unmet need among subgroups of patients with mutations associated with a lower chance of responding to immuno-oncology therapy or with a generally poor prognosis, such as KRAS, STK11, and KEAP mutations. The clinical experts noted that durvalumab and tremelimumab plus platinum-based chemotherapy would be used as a standalone treatment for patients with previously untreated metastatic NSCLC not harbouring a targetable oncogenic aberration with an associated SOC first-line targeted therapy option (e.g., an EGFR mutation or ALK rearrangement) and that this regimen would not be expected to cause a shift in the current treatment paradigm but would be an additional first-line option combining immuno-oncology treatment and chemotherapy. According to the clinical experts, patients ideally suited for this treatment would have previously untreated metastatic NSCLC, a good ECOG PS (0 or 1), no actionable oncogenic mutations or translocations (including EGFR mutations and ALK rearrangements), any tumoural PD-L1 expression, no significant comorbidities, and the ability to report adverse effects as soon as possible. The clinical experts pointed out that, in practice, durvalumab and tremelimumab plus platinum-based chemotherapy is anticipated to be used most often in patients with PD-L1 levels of less than 50%, but they added that, because patients with any PD-L1 status were included in the POSEIDON trial, approval and funding should be in accordance with trial outcomes. The clinical experts also noted that durvalumab and tremelimumab plus platinum-based chemotherapy may be an option for patients with tumours that have biomarkers that may predict a poor or no response to currently available ICI-based therapies (e.g., those that do not express PD-L1 or have KRAS, STK11, or KEAP mutations). The clinical experts stated that the treatment strategy would be based on an informed decision by the patients in consultation with a medical oncologist, and that response to treatment would be determined through clinical and radiological assessments. Treatment and monitoring would be administered locally for patients who live close to regional cancer clinics and at community oncology networks for patients living in remote settings, where specially trained general practitioners or internists would oversee day-to-day treatment. The clinical experts acknowledged that it would be appropriate for any systemic therapy unit and lung cancer treatment team currently administering other combinations of immuno-oncology and chemotherapy to deliver treatment with durvalumab and tremelimumab plus platinum-based chemotherapy. According to the clinical experts, discontinuation of first-line treatment for metastatic NSCLC would occur when there is disease progression, unacceptable toxicity, or the patient chooses to stop treatment. One clinical expert also noted that discontinuation of treatment may occur when there is suspicion of progression and a risk of rapid clinical deterioration if treatment is continued. According to the clinical experts, early identification and management of immune-related AEs is paramount to reduce the risk of treatment-related morbidity or mortality, as is the case with any immunotherapy-containing regimen. Subspecialist consultations may be required to effectively manage immune toxicity.

Clinician Group Input

Two clinician groups, the Lung Cancer Canada Medical Advisory Committee (LCC-MAC) and the Ontario Health (Cancer Care Ontario) Lung Cancer Drug Advisory Committee (DAC) provided input for this submission. The LCC-MAC is a national charity that has been offering support and education to patients with lung cancer and their families and supporting research, and has been providing clinician input for submissions of new lung cancer drugs to the health technology assessment (HTA) process for many years. The DAC provides evidence-based clinical and health system guidance on drug-related issues to support the mandate of Cancer Care Ontario, including the Provincial Drug Reimbursement Programs and the Systemic Treatment Program.

A total of 19 clinicians from the LCC-MAC and 5 clinicians from the DAC provided input for this submission. The submission from the LCC-MAC was informed by data and information from publicly available sources, primarily published manuscripts and conference presentations, together with clinical experience of the LCC-MAC members. Information from Cancer Care Ontario was gathered via teleconference meeting and emails.

The DAC highlighted that the treatment combination of durvalumab and tremelimumab plus platinum-based chemotherapy would serve as a first-line treatment. The LCC-MAC noted that most thoracic oncologists in Canada had experience with using the treatment combination through clinical trials. It also emphasized that the treatment combination would not replace other regimens that were already approved or funded but would serve as an alternative. Chemotherapy and immunotherapy options were noted as appropriate comparators to the treatment combination by the LCC-MAC.

In terms of unmet needs, both clinician groups agreed on the need for treatments with better tolerability. The DAC noted the following additional unmet needs: not all patients respond to available treatments, patients become refractory to current treatment options, and the lack of treatments and formulations that improve compliance and that convenience.

The clinician groups highlighted treatment goals similar to those indicated by the clinical experts, in addition to tumour shrinkage. The LCC-MAC clinician group further noted that an important outcome for immunotherapy combinations is survival with the goal of improving the number of patients with durable benefits of treatment, as manifested by increases in progression-free survival (PFS) and overall survival (OS) at 5 years.

The clinician groups noted that those best suited for the treatment combination included patients without actionable driver mutations with any PD-L1 status and those with KRAS, STK11 and KEAP mutations not currently actionable in the first-line setting. The DAC also noted that patients with stage IV or incurable NSCLC considering first-line therapy would be suitable candidates for treatment.

Both groups highlighted the need for clinical assessment of symptoms and imaging (e.g., CT scans and chest X-rays) to monitor response to treatment. The DAC advised treatment response assessment every 6 weeks initially, and then less often. However, the LCC-MAC suggested that scans be made every 3 months, with symptom assessments every 3 to 4 weeks, depending on if a patient receives chemotherapy or single-drug immunotherapy during the maintenance phase.

The clinician groups agreed that disease progression is a factor when deciding whether to discontinue treatment. The LCC-MAC indicated that immunotherapy may be interrupted or discontinued because of immune-related AEs, and most low-grade toxicities were manageable with symptom management, topical or oral steroids, and potential treatment interruption. The DAC also identified intolerable side effects and patient withdrawal as factors when considering treatment discontinuation. The LCC-MAC indicated the initial patient assessment should be carried out by a medical oncologist. Both clinician groups stated that the treatment combination should be administered in facilities with expertise in managing cytotoxic anticancer therapies.

Drug Program Input

Input was obtained from the drug programs that participate in the reimbursement review process. The following were identified as key factors that could potentially affect the implementation of a recommendation for durvalumab and tremelimumab: relevant comparators; considerations for prescribing, initiation, continuation or renewal, and discontinuation of therapy; generalizability; funding algorithm; care provision; and system and economic issues. The details of the drug program input along with advice from the clinical experts consulted for this review are provided in Table 6.

Clinical Evidence

Systematic Review

Description of Studies

One phase III, randomized, open-label, global study, POSEIDON, assessed the efficacy and safety of durvalumab with or without tremelimumab for the first-line treatment of patients (total N = 1,013) with metastatic NSCLC with tumours without activating EGFR mutations and ALK fusions. Patients were randomized 1:1:1 to 1 of 3 study arms: durvalumab and tremelimumab plus SOC chemotherapy; durvalumab plus SOC chemotherapy; or SOC chemotherapy alone. Randomization was stratified according to PD-L1 expression status (TC level ≥ 50% versus < 50%), disease stage (IVA versus IVB), and histology (nonsquamous versus squamous). The primary objective of the POSEIDON trial was to assess the efficacy of durvalumab plus SOC chemotherapy compared to SOC chemotherapy alone in terms of dual primary end points of PFS and OS. The key secondary objectives were to assess the efficacy of durvalumab and tremelimumab plus SOC chemotherapy compared to SOC chemotherapy alone in terms of PFS as assessed by blinded independent central review (BICR) using Response Evaluation Criteria in Solid Tumours Version 1.1 (RECIST 1.1) and OS. Other secondary end points included objective response rate (ORR) and health-related quality of life (HRQoL), as measured by the European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30 (EORTC QLQ-C30) and European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Lung Cancer 13 (EORTC QLQ-LC13), to compare durvalumab and tremelimumab plus SOC chemotherapy versus SOC chemotherapy alone. Results from this comparison (durvalumab and tremelimumab plus SOC chemotherapy versus SOC chemotherapy alone) are presented in this report, and include results from the prespecified main analyses of the POSEIDON trial, referred to as the “final analysis” (data cut-off [DCO]: July 24, 2019, for PFS and other RECIST-related end points; March 12, 2021, for OS and safety data), and a 5-year OS analysis update (DCO: August 24, 2023). The primary end points are based on the evaluation of durvalumab plus SOC chemotherapy versus SOC chemotherapy alone, which is outside the scope of the indication approved by Health Canada and the sponsor’s reimbursement request; the primary end points are therefore not the focus of this review.

Patients eligible for participation in the POSEIDON trial were aged 18 years or older with stage IV NSCLC not amenable to curative surgery or radiation with tumours lacking activating EGFR mutations and ALK fusions. Across all study arms, the median age of patients was 64 years, and the majority were male (76.0%), white (55.9%), and either current or former smokers (78%), with an ECOG PS of 1 (66.5%). The included patients had not received prior chemotherapy or any other systemic therapy for metastatic NSCLC, including prior exposure to immune-mediated therapy. Participants were required to have a WHO performance status or ECOG PS of 0 or 1. Patients with active or prior documented autoimmune or inflammatory disorders (with some exceptions) were not eligible for the trial.

Efficacy Results

As of March 12, 2021, the median OS in the durvalumab and tremelimumab plus SOC chemotherapy group was 14.0 months (95% confidence interval [CI], 11.7 to 16.1 months) compared with 11.7 months (95% CI, 10.5 to 13.1 months) in the SOC chemotherapy alone group, with a hazard ratio (HR) of 0.77 (95% CI, 0.650 to 0.916; P = 0.00304). For the OS final analysis results, the median follow-ups were 13.63 months in the durvalumab and tremelimumab plus SOC chemotherapy group and 11.17 months in the SOC chemotherapy alone group. At the 5-year OS analysis DCO, the median OS in the durvalumab and tremelimumab plus SOC chemotherapy group was 14.0 months (95% CI, 11.7 to 16.1 months) versus 11.6 months (95% CI, 10.5 to 13.1 months) in the SOC chemotherapy alone group, with an HR of 0.76 (95% CI, 0.642 to 0.893). For the 5-year OS analysis results, the median follow-up periods were 13.63 months in the durvalumab and tremelimumab plus SOC chemotherapy group and 11.10 months in the SOC chemotherapy alone group. The difference in median OS was statistically significant in favour of durvalumab and tremelimumab plus SOC chemotherapy at the final analysis time point (statistical significance could not be determined for the 5-year time point results because these estimates were not alpha-controlled). Results for longer-term survival probability also demonstrated improvement for durvalumab and tremelimumab plus SOC chemotherapy versus SOC chemotherapy alone at 36 months (25.3% versus 13.3%, respectively), 48 months (20.9% versus 8.5%), and 60 months (15.7% versus 6.8%). The clinical experts consulted by CDA-AMC considered the differences in median OS at both time points and the differences in survival probabilities at 36 months, 48 months, and 60 months to be clinically meaningful.

As of July 24, 2019, the median PFS assessed by BICR in the durvalumab and tremelimumab plus SOC chemotherapy arm was 6.2 months (95% CI, 5.0 to 6.5 months) compared with 4.8 months (95% CI, 4.6 to 5.8 months) in the SOC chemotherapy alone arm, with an HR of 0.72 (95% CI, 0.600 to 0.860; 2-sided P = 0.00031). The difference was statistically significant in favour of durvalumab and tremelimumab plus SOC chemotherapy. The PFS rate at 12 months for patients in the durvalumab and tremelimumab plus SOC chemotherapy arm was 26.6% and for patients in the SOC chemotherapy alone arm it was 13.1%. For the PFS results, the median follow-ups for censored patients were 11.9 months in the durvalumab and tremelimumab plus SOC chemotherapy group and 5.5 months in the SOC chemotherapy alone group. The clinical experts considered the differences in median PFS and in PFS rate at 12 months to be clinically meaningful.

In terms of other secondary end points, the ORR for durvalumab and tremelimumab plus SOC chemotherapy was 46.3% compared with 33.4% for SOC chemotherapy alone, with an odds ratio (OR) of 1.72 (95% CI, 1.260 to 2.367; nominal P < 0.001). The statistical significance of this result could not be determined because the ORR end point was not included in the hierarchical statistical analysis plan. However, the clinical experts considered the difference between groups to be clinically meaningful. The median follow-up times for censored patients were 11.9 months in the durvalumab and tremelimumab plus SOC chemotherapy group and 5.5 months in the SOC chemotherapy alone group. In the assessment of HRQoL, time to deterioration in the EORTC QLQ-C30 Global Health Status (GHS)/QoL was 8.3 months (95% CI, 6.4 to 10.2) in the durvalumab and tremelimumab plus SOC chemotherapy arm compared with 5.6 months (95% CI, 4.4 to 7.5) in the SOC chemotherapy alone arm. The statistical significance of this result could not be determined; however, the clinical experts described the difference between groups as clinically meaningful. The median durations of follow-up were 13.63 months in the durvalumab and tremelimumab plus SOC chemotherapy group and 11.17 months in the SOC chemotherapy alone group.

No published between-group minimal important difference (MID) values were provided by the sponsor for OS, PFS, ORR, and time to deterioration in the EROTC QLQ-C30 GHS/QoL in the first-line treatment of metastatic NSCLC; as such, the thresholds used to judge the target of certainty in the Grading of Recommendations Assessment, Development and Evaluation (GRADE) assessment (Table 4) are based on input from clinical experts consulted by CDA-AMC.

Harms Results

Similar proportions of patients experience at least 1 AE in the durvalumab and tremelimumab plus SOC chemotherapy arm (97.3%) and the SOC chemotherapy alone arm (96.1%), with the most commonly reported AEs being anemia (49.7% versus 48.9%, respectively), nausea (41.5% versus 36.6%), neutropenia (30.0% versus 23.4%), decreased appetite (28.2% versus 24.6%), fatigue (24.5% versus 22.2%), diarrhea (21.5% versus 15.3%), rash (19.4% versus 6.6%), and constipation (19.1% versus 23.7%). Serious adverse events (SAEs) were reported in a higher percentage of patients in the durvalumab and tremelimumab plus SOC chemotherapy arm (44.2%) than in the SOC chemotherapy alone arm (35.1%), as were AEs leading to discontinuation of any study treatment (22.1% versus 15.3%). AEs leading to death were reported in 12.4% of patients in the durvalumab and tremelimumab plus SOC chemotherapy arm and 9.0% of patients in the SOC chemotherapy alone arm.

Adverse events of special interest (AESIs) or adverse events of potential interest (AEPIs) were reported in a higher percentage of patients in the durvalumab and tremelimumab plus SOC chemotherapy arm than in the SOC chemotherapy alone arm (█████ ██████ █████). These included grade 3 or 4 AESIs or AEPIs (█████ ██████ ████) and AESIs or AEPIs that led to study treatment discontinuation █████ ██████ ████). AESIs or AEPIs with an outcome of death were reported in | ██████ patients in the durvalumab and tremelimumab plus SOC chemotherapy arm and | ██████ patient in the SOC chemotherapy alone arms.

The percentages of patients experiencing select AEs and AESIs (identified by clinical experts as AEs of greatest clinical importance) in the durvalumab and tremelimumab plus SOC chemotherapy and SOC chemotherapy alone groups are as follows: neutropenia (█████ ██████ ███████ pneumonitis (████ ██████ ████), hepatitis (████ ██████ ██ ████████), colitis (████ ██████ ████), type 1 diabetes mellitus (0.3% versus no patients), and myocarditis (0.3% versus no patients).

For the safety analysis, the median follow-up times were 13.63 months in the durvalumab and tremelimumab plus SOC chemotherapy arm and 11.17 months in the SOC chemotherapy alone arm.

Critical Appraisal

The POSEIDON trial was a phase III, randomized, open-label, comparative study. Methods of randomization and treatment allocation were adequate. Reported baseline characteristics were generally balanced across the study arms and the clinical experts consulted by CDA-AMC did not identify any considerable differences between the durvalumab and tremelimumab plus SOC chemotherapy and SOC chemotherapy alone treatment arms that would be expected to affect the interpretation of results. The prespecified sample size was achieved and actual screening failures did not exceed the expected number. The POSEIDON trial was powered for the dual primary end points and power calculations were provided for all end points included in the multiple testing procedure (MTP), which included the key secondary end points of OS and PFS for the comparison of durvalumab and tremelimumab plus SOC chemotherapy against SOC chemotherapy alone. The study used an open-label design because of differences in administration schedule and duration of study treatments. Of the efficacy end points of interest for this review, OS is an objective end point, and PFS, ORR, best objective response (BoR), and duration of response (DoR) were assessed by BICR according to RECIST 1.1 criteria, which helped to mitigate potential bias. Patient-reported efficacy assessments (i.e., HRQoL end points) and reporting of AEs can be influenced by knowledge of the treatment assignment by patients and clinicians. Other limitations related to internal validity include the potential impact of disproportionate treatment discontinuation between study arms, higher treatment exposure to SOC chemotherapy, overall higher exposure to study treatment in the durvalumab and tremelimumab plus SOC chemotherapy arm, and the possible confounding effect of subsequently received anticancer treatments on survival. Greater study treatment exposure (to SOC chemotherapy in particular) in the durvalumab and tremelimumab plus SOC chemotherapy group may have biased efficacy results in favour of this treatment regimen, whereas the higher proportion of patients in the SOC arm who received subsequent systemic anticancer therapy may have diluted the survival benefit observed with the combination treatment compared to SOC chemotherapy alone. Additionally, non-key secondary end points were not controlled for multiplicity. Overall adherence rates to the patient-reported outcome (PRO) scales generally decreased over time and were lower in the SOC chemotherapy arm. The missing data present a challenge in evaluating effects on HRQoL.

The clinical experts consulted for this review commented that the POSEIDON trial used standard inclusion and exclusion criteria that would be expected in a study of first-line treatment of patients with metastatic NSCLC. However, the experts identified differences between the POSEIDON trial participants and patients who would receive first-line immuno-oncology plus SOC chemotherapy for metastatic NSCLC in clinical practice in Canada. For example, patients in practice would typically be older, have relative contraindications to immunotherapy (i.e., quiescent inflammatory or autoimmune conditions that would not result in significant morbidity if reactivated), would include patients with an ECOG PS of 2, and those who had received immunotherapy for earlier-stage cancer. Also, the POSEIDON trial did not include any sites in Canada. However, the clinical experts noted that, overall, they did not have concerns regarding the generalizability of the study findings to patients seen in clinical practice. The main efficacy and harms outcomes assessed in the POSEIDON trial align with outcomes of importance identified by patients and clinicians.

An important limitation of the POSEIDON trial is that durvalumab and tremelimumab plus SOC chemotherapy was compared to SOC chemotherapy alone. When the trial was conducted, platinum-based chemotherapy was the standard therapy for the indicated population; however, current first-line treatment of metastatic NSCLC without targetable genomic alterations consists of immuno-oncology treatment (on its own or in combination with chemotherapy). Concerning subsequent anticancer treatments received in the POSEIDON trial, the clinical experts noted that rechallenging patients with immunotherapy and the low rate of patients who received immunotherapy following first-line chemotherapy in the POSEIDON trial are not reflective of practice in Canada.

GRADE Summary of Findings and Certainty of the Evidence

The selection of outcomes for GRADE (Table 4) was based on the sponsor’s Summary of Clinical Evidence, consultation with the clinical experts, and input received from patient and clinician groups and public drug plans. The following list of outcomes was finalized in consultation with the expert committee members:

• OS (median and probability of survival at 36, 48, and 60 months)

• PFS (median and PFS rate at 12 months)

• ORR

• HRQoL (time to deterioration in EORTC QLQ-C30 GHS/QoL)

• AEs (AESIs or AEPIs).

Table 4: Summary of Findings for Durvalumab and Tremelimumab Plus SOC Chemotherapy vs. SOC Chemotherapy for Patients With Metastatic NSCLC With No Sensitizing EGFR Mutations or ALK Genomic Tumour Aberrations

AE = adverse event; AEPI = adverse event of potential interest; AESI = adverse event of special interest; BICR = blinded independent central review; CI = confidence interval; D = durvalumab; EORTC QLQ-C30 = European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30; GHS = Global Status of Health; HR = hazard ratio; MID = minimal important difference; NA = not applicable; NSCLC = non–small cell lung cancer; OR = odds ratio; ORR = objective response rate; OS = overall survival; PFS = progression-free survival; QoL = quality of life; RCT = randomized controlled trial; RECIST 1.1 = Response Evaluation Criteria in Solid Tumours Version 1.1; SOC = standard of care; t = tremelimumab; vs. = versus.

Note: 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.

^aFinal analysis data cut-off March 12, 2021.

^bRated down 1 level for serious imprecision. Because no MID from the literature was provided, the target of certainty appraisal was the MID of 2 to 3 months based on clinical expert input; CI for difference between groups includes the possibility of trivial effects and no difference.

^cFive-year OS update data cut-off August 24, 2023.

^dFive-year estimates and CIs are not alpha-controlled.

^eRated down 1 level for serious imprecision. Because no MID from the literature was provided, the target of certainty appraisal was the MID of 2.5 months based on clinical expert input; CI for difference between groups includes the possibility of trivial effects and no difference.

^fBecause no MID from the literature was provided, the target of certainty appraisal was the MID of 5 to 10 per 100 people based on clinical expert input; CI for difference between groups included potential for important benefit.

^gBecause no MID from the literature was provided, the target of certainty appraisal was the MID of 5 per 100 people based on clinical expert input; CI for difference between groups included potential for important benefit.

^hFinal analysis data cut-off July 24, 2019.

ⁱRated down 1 level for serious imprecision. Because no MID from the literature was provided, the target of certainty appraisal was the MID of 1 to 2 months based on clinical expert input; CI for difference between groups includes the possibility of trivial effects and no difference.

^jSecondary efficacy end point not adjusted for multiplicity (considered supportive evidence).

^kRated down 1 level for serious imprecision. Because no MID from the literature was provided, the target of certainty appraisal was the MID of 10 per 100 people based on clinical expert input; CI for difference between groups includes the possibility of trivial effects.

^lRated down 2 levels for serious risk of bias because of knowledge of treatment assignment having the potential to affect reporting or recording of HRQoL outcomes and because of missing data (adherence rates fell below 60% after 88 weeks in the durvalumab and tremelimumab plus chemotherapy arm and after 24 weeks in the chemotherapy alone arm). Rated down 1 level for serious imprecision. Because no MID from the literature was provided, the target of certainty appraisal was the MID of 1.5 months based on clinical expert input; CI for difference between groups includes the possibility of trivial effects and no difference.

^mNo statistical tests were performed. The difference in effects between groups was considered certain based on input from clinical experts.

Sources: Sponsor’s Summary of Clinical Evidence¹ and sponsor’s data on file.¹³

Long-Term Extension Studies

No long-term extension studies were submitted by the sponsor.

Indirect Comparisons

Description of Studies

The sponsor conducted indirect treatment comparisons (ITCs) to estimate the relative efficacy of durvalumab and tremelimumab plus platinum-based chemotherapy versus other approved ICI plus platinum-based chemotherapy combinations among patients with metastatic NSCLC who lack EGFR mutations and ALK aberrations.

Efficacy Results

Based on the sponsor-submitted matching adjusted indirect comparisons (MAICs), overall, the efficacy, in terms of OS and PFS, appeared comparable when durvalumab and tremelimumab plus platinum-based chemotherapy was compared with relevant comparators for the first-line treatment of patients with metastatic NSCLC and no sensitizing EGFR mutations or ALK genomic tumour aberrations.

In the comparison of durvalumab and tremelimumab plus platinum-based chemotherapy versus pembrolizumab plus chemotherapy, the OS HR (95% CI) was ████ ██████ █████. The PFS HR (95% CI) was ████ ██████ █████ in the nonsquamous population, the OS HR (95% CI) was ████ ██████ █████, and PFS HR (95% CI) was ████ ██████ █████ in the squamous population.

In the comparison of durvalumab and tremelimumab plus platinum-based chemotherapy versus nivolumab plus ipilimumab plus chemotherapy, the OS HR was ████ ██████ ████). The PFS was ████ ██████ █████ in the squamous plus nonsquamous (intention to treat [ITT]) population and the OS HR was ████ ██████ █████. The PFS HR was ████ ██████ █████ in the nonsquamous population, the OS HR was ████ ██████ ██████ and the PFS HR was ████ ██████ █████ in the squamous population.

In the comparison of durvalumab and tremelimumab plus platinum-based chemotherapy versus cemiplimab plus chemotherapy, the OS HR was ████ ██████ █████. The PFS HR was ████ ██████ █████ in the squamous plus nonsquamous (ITT) population; the OS HR was ████ ██████ ██████ The PFS HR was ████ ██████ █████ in the nonsquamous population, the OS HR was ████ ██████ ██████, and the PFS HR was ████ ██████ █████ in squamous population.

Harms Results

The sponsor-submitted ITCs did not report harms outcomes.

Critical Appraisal

Overall, the MAICs were conducted according to accepted methodological guidance. The potential key limitations of the MAICs were that the analysis did not adjust for all the important effect modifiers (such as PD-L1 status, sex, age, ECOG PS, presence of brain metastases, disease stage, and type of chemotherapy), likely leading to residual heterogeneity between studies. In addition, the MAIC method reduced the effective sample size (by up to 40%). Imprecision observed in the effect point estimates, as indicated by the wide 95% CI, precluded making conclusions about the comparative effectiveness of durvalumab and tremelimumab plus chemotherapy versus other comparators in terms of OS and PFS.

Studies Addressing Gaps in the Evidence From the Systematic Review

No studies addressing gaps in the systematic review evidence were submitted by the sponsor.

Conclusions

Based on evidence from the POSEIDON study, durvalumab and tremelimumab plus SOC chemotherapy demonstrated a benefit compared to SOC chemotherapy alone in OS, PFS, and ORR for the first-line treatment of adult patients with metastatic NSCLC with tumours that lack activating EGFR mutations and ALK fusions. The GRADE assessment suggested that durvalumab and tremelimumab plus SOC chemotherapy likely results in a clinically important increase in median OS, median PFS, and ORR compared to SOC chemotherapy, and a high certainty that durvalumab and tremelimumab plus SOC chemotherapy results in a clinically important increase in the survival rate at 36 months, 48 months, and 60 months, as well as in the PFS rate at 12 months. However, the evidence is very uncertain about the effect of durvalumab and tremelimumab plus SOC chemotherapy on HRQoL, as measured by time to deterioration in EORTC QLQ-C30 GHS/QoL when compared with SOC chemotherapy alone.

In the POSEIDON study, AEPIs and AESIs (including those that were grade 3 or 4) were experienced by a higher percentage of patients receiving durvalumab and tremelimumab plus SOC chemotherapy than those receiving SOC chemotherapy alone, with the GRADE assessment suggesting with high certainty that durvalumab and tremelimumab plus SOC chemotherapy results in a higher incidence of AEPIs and AESIs when compared to SOC chemotherapy alone. The clinical experts consulted for this review stated that the types and rates of AEs in the POSEIDON trial were not unexpected. They indicated that immune-mediated adverse events (imAEs) caused by ICI treatment are typically of low grade and reversible and that health care providers treating lung cancer have a decade of experience using ICIs in routine clinical practice. The experts also noted that some imAEs can be very serious or life-threatening and that interventions required for the treatment of imAEs can negatively impact patient QoL.

Given that SOC chemotherapy was the comparator in the POSEIDON study, the lack of direct comparative data for the efficacy and safety between durvalumab and tremelimumab plus platinum-based chemotherapy and currently used first-line treatments (i.e., immuno-oncology plus chemotherapy options) represents a key gap in the evidence.

The findings of the MAICs demonstrated a largely comparable benefit in terms of OS and PFS when comparing combinations of durvalumab and tremelimumab plus platinum-based chemotherapy with relevant comparators for first-line treatment in patients with metastatic NSCLC without EGFR mutations and ALK aberrations. Harms were not investigated in the MAICs.

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 durvalumab (50 mg/mL concentrate for IV infusion) in combination with tremelimumab (20 mg/mL concentrate for IV infusion) and platinum-based chemotherapy in the first-line treatment of metastatic NSCLC with no sensitizing EGFR mutations or ALK genomic tumour aberrations.

Disease Background

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

Lung cancer is among the most commonly diagnosed cancers¹⁴ and is the leading cause of cancer-related death in Canada.³ In Canada, the 5-year net survival rate for lung cancer (all stages and histological subtypes) is 22%.¹⁴ Lung cancer is heterogenous in nature, comprising several different disease subtypes and categorized by histology, staging, and mutation driver status.³ NSCLC is the most common type, representing approximately 89% of all lung cancer diagnoses in Canada.⁴ NSCLC can be categorized into histologic subtypes: squamous-cell carcinoma (17% to 27% of NSCLC cases), and non–squamous cell carcinoma (73% to 83% of NSCLC cases).^3,5 At early stages, NSCLC may be asymptomatic,³ with symptoms developing only when the cancer has become more advanced and is no longer amenable to curative-intent therapy. Presenting symptoms can sometimes be nonspecific (common manifestations include coughing, chest pain, hemoptysis, fatigue, weight loss, dyspnea, hoarseness, and recurring infections with bronchitis and pneumonia), contributing to delays in diagnosis.^3,5,6 Screening programs for NSCLC are being established throughout Canada, but many patients who would qualify for NSCLC screening live in jurisdictions where it is not yet available.

To determine a patient’s prognosis and treatment, a tissue biopsy is performed for histologic confirmation and the tumour is staged using the most recent American Joint Committee on Cancer staging criteria, which involve tumour-node-metastasis classification of the disease, based on assessment of the size and spread of the primary tumour, lymph node involvement, and occurrence of distance metastasis.¹⁵ Recent advances in NSCLC, including molecular profiling, allow for the identification of genomic alterations or mutations (i.e., oncogenic drivers).¹⁶ Several oncogenic driver mutations have been identified, where frequently occurring mutations include EGFR mutations and ALK aberrations,⁷ which collectively occur in approximately 25% of patients.⁸ An estimated 74.5% of NSCLC cases do not have EGFR or ALK aberrations.

A diagnosis starts with a complete medical history and physical exam, assessment of risk factor exposure (e.g., history of tobacco use or exposure to radon gas), along with standard tests, including lung function, routine hematology, renal and hepatic function, and bone biochemistry tests.^17-19 Patients also undergo radiographic staging, which may include CT scans or PET-CT scans of the chest, abdomen, and pelvis, and MRI scans of the central nervous system (CNS).^17-19 Upon receiving a confirmed diagnosis of NSCLC that includes histopathological subtyping, patients undergo testing for oncogenic drivers and PD-L1 levels via biomarker testing at an accredited laboratory. Tumours with a nonsquamous histology are tested for a wide panel of oncogenic drivers including EGFR mutations (common and uncommon), ALK rearrangements, ROS1 rearrangements, and BRAF mutations, in addition to PD-L1 immunohistochemical testing.²⁰ Tumours with squamous histology are typically tested only for PD-L1, with most jurisdictions only performing molecular testing in squamous tumours on request in specific cases (e.g., a tumour from a patient with little or no tobacco use history). Biomarker testing is performed by molecular tissue testing on a tissue sample most commonly obtained via biopsy.²⁰ Molecular tissue testing in NSCLC is well-established and a standard diagnostic workup is performed at the time of diagnoses based on current clinical practices in Canada.²⁰

Standards of Therapy

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

Although currently available ICI-based therapies can produce deep and durable responses in a small subset of patients, there is no cure for NSCLC in the metastatic stage.²¹ The goals of therapy for patients with metastatic NSCLC, as identified by the clinical experts consulted by CDA-AMC, include prolonging survival, extending the time before disease progression, decreasing cancer-related symptoms, and maintaining or improving QoL. In current practice, the treatment of metastatic NSCLC is guided by the presence or absence of oncogenic drivers, including EGFR mutations and ALK rearrangements (among others).^11,17 Nearly half (48.4%) of NSCLC cases are metastatic (stage IV).^3,8 For the majority of patients with metastatic NSCLC whose tumours lack actionable genomic alterations, treatment is selected based on tumour histology (squamous or nonsquamous); tumoural PD-L1 expression (which is predictive of response to ICIs); and patient symptom burden (as measured by ECOG PS), comorbidities, and preferences.^9-11

In the past, SOC first-line systemic therapy for patients with advanced NSCLC was treatment with a platinum doublet chemotherapy regimen.²² Advances in recent years have led to ICI-based therapies, either as monotherapy or in combination with chemotherapy, being the current standard first-line treatment for metastatic NSCLC without targetable genomic alterations.^10,11 The sponsor has described the current treatment algorithm for metastatic NSCLC without targetable genomic alterations, as presented in Figure 1. For patients with any or unknown PD-L1 expression levels, first-line options include pembrolizumab plus platinum-based chemotherapy, nivolumab and ipilimumab plus platinum-based chemotherapy, or cemiplimab plus platinum-based chemotherapy.^10,11 Monotherapy with pembrolizumab or cemiplimab may be selected as first-line treatment for patients with high PD-L1 expression (≥ 50% TCs).^10,11 Cemiplimab (with or without platinum-based chemotherapy) is not currently funded in Canada.¹² Platinum-based chemotherapy alone is also a potential first-line treatment option; however, the clinical experts noted that this would be reserved for patients with significant contraindications to immunotherapy and that few patients would receive this option. The clinical experts indicated that immuno-oncology treatment is considered the backbone of therapy for the first-line treatment of metastatic NSCLC without EGFR mutations or ALK aberrations.

Figure 1: Current Treatment Algorithm for Metastatic NSCLC Without Targetable Genomic Alterations

CDA-AMC = Canada’s Drug Agency; ICI = immune checkpoint inhibitor; NSCLC = non–small cell lung cancer; pCPA = pan-Canadian Pharmaceutical Alliance.

Note: The sponsor noted that the indication of interest for this review is for ICI-based regimens in combination with platinum-based chemotherapy, as indicated by the blue shading in the PD-L1 Any/Unknown box. A deviation request was approved by CDA-AMC allowing for the exclusion of monotherapy with immuno-oncology drugs (i.e., those given without platinum-based chemotherapy) as comparators. Cemiplimab has a “reimburse with clinical criteria and/or conditions” CDA-AMC recommendation for patients who have a PD-L1 status of 50% or greater and in combination with chemotherapy for patients with an “Any/Unknown” PD-L1 status, but it is not currently listed on any provincial drug plan or cancer agency.^12,23

Sources: Developed by the sponsor based on the CDA-AMC provisional funding algorithm¹² and Clinical Practice Guidelines.^10,11

Drug Under Review

Key characteristics of durvalumab and tremelimumab are summarized in Table 5 with other treatments available for the treatment of metastatic NSCLC.

According to the product monograph, the recommended dosing for durvalumab in combination with tremelimumab and platinum-based chemotherapy differs based on the patients’ body weight.² The recommended dosage during chemotherapy is either durvalumab 1,500 mg in combination with tremelimumab 75 mg and platinum-based chemotherapy (body weight ≥ 30 kg) or durvalumab 20 mg/kg in combination with tremelimumab 1 mg/kg and platinum-based chemotherapy (body weight < 30 kg) every 3 weeks (21 days) for 4 cycles. After platinum-based chemotherapy, the recommended dosage of durvalumab is 1,500 mg (body weight ≥ 30 kg) or 20 mg/kg (body weight < 30kg) every 4 weeks and histology-based pemetrexed maintenance therapy every 4 weeks. A fifth dose of tremelimumab is to be given at week 16 alongside durvalumab dose 6. During cycle 1, tremelimumab is followed by durvalumab starting approximately 1 hour (maximum 2 hours) after the end of the tremelimumab infusion. Platinum-based chemotherapy infusion should start approximately 1 hour (maximum 2 hours) after the end of the durvalumab infusion. If there are no clinically significant concerns during cycle 1, then subsequent cycles of durvalumab can be given at the physician’s discretion immediately after tremelimumab, and the time between the end of the durvalumab infusion and the start of chemotherapy can be reduced to 30 minutes.²

Durvalumab is an engineered monoclonal antibody that blocks the interaction of PD-L1 with its receptors PD-1 and CD80.² Selective blockade of interactions between PD-L1 and PD-1 and between PD-L1 and CD80 lead to inhibition of immune responses and enhanced antitumour activity.² Tremelimumab is a selective, fully human immunoglobulin G2 antibody that blocks CTLA-4 interactions with CD80 and CD86, enhancing T-cell activation and proliferation, which may result in increased T-cell diversity and enhanced antitumour immune activity.²⁴

The reimbursement request is for durvalumab in combination with tremelimumab and platinum-based chemotherapy for the first-line treatment of adult patients with metastatic NSCLC with no sensitizing EGFR mutations or ALK genomic tumour aberrations, which is aligned with the Health Canada–approved indication. The HTA agency in France (Haute Autorité de Santé)²⁵ and the FDA²⁶ approved the use of the durvalumab in combination with tremelimumab and platinum-based chemotherapy for the same indication approved by Health Canada. The European Medicines Agency initially provided a positive recommendation for tremelimumab for the same indication (December 15, 2022)²⁷ and granted market authorization on February 20, 2023.²⁸ However, it withdrew the market authorization at the request of the marketing authorization holder, AstraZeneca AB, on November 6, 2023.²⁹ The HTA agency in England, the National Institute for Health and Care Excellence (NICE)³⁰ was asked to carry out a single technology appraisal of durvalumab with tremelimumab for untreated EGFR-negative and ALK-negative locally advanced and metastatic NSCLC by the Department for Health and Social Care. In August 2022, the appraisal was suspended as the company advised that it is was no longer pursuing a marketing authorization from the Medicines and Health care Products Regulatory Agency for the indication. In August 2023, appraisal was discontinued as no further information was received from the company.

Tremelimumab in combination with durvalumab has been previously reviewed by CDA-AMC for the first-line treatment of adult patients with unresectable hepatocellular carcinoma who require systemic therapy, with a recommendation to “reimburse with clinical criteria and/or conditions” (issued in November 2023).³¹ Durvalumab has been previously reviewed and received conditional positive recommendations for various therapeutic areas (unresectable NSCLC, extensive-stage small-cell lung cancer, and biliary tract cancer).^32-34

Table 5: Key Characteristics of Durvalumab in Combination With Tremelimumab; Nivolumab in Combination With Ipilimumab; Cemiplimab; and Pembrolizumab

BW = body weight; NSCLC = non–small cell lung cancer.

^aHealth Canada–approved indication.

Sources: Product monographs of durvalumab,² tremelimumab,²⁴ nivolumab,³⁵ cemiplimab,³⁶ and pembrolizumab.³⁷

Perspectives of Patients, Clinicians, and Drug Programs

The full patient and clinician group submissions received are available in the consolidated patient and clinician group input document for this review on the project website.

Patient Group Input

This section was prepared by the review team based on input provided by patient groups.

A joint patient group submission was received from CCSN, LCC, and LHF. 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 best SOC, including early diagnosis, timely treatment and follow-up care, support for patients with cancer, and regarding issues related to survivorship or quality of end-of-life care. LCC is a registered national charitable organization that serves as Canada’s leading resource for lung cancer education, patient support, research and advocacy. LHF (previously the Ontario Lung Association) is a registered charity that assists and empowers people living with or caring for others with lung disease. It is a recognized leader, voice, and primary resource in the prevention and control of respiratory illness, tobacco cessation and prevention, and its effects on lung health. The Foundation provides programs and services to patients and health care providers, invests in lung research, and advocates for improved policies in lung health.

CCSN, LCC, and LHF collectively produced a survey that was disseminated through the 3 organizations’ social media platforms, as well as CCSN’s monthly newsletter to gather responses from August 1, 2024, to the date of writing of the patient group submission. LCC also interviewed 1 patient who was a part of the POSEIDON trial, on October 8, 2024.

The experience of the patient with lung cancer who participated in the POSEIDON trial has been summarized. The patient’s diagnosis took a while before they could get started on treatments. The patient started with first-line chemotherapy and received durvalumab simultaneously. The treatment regimen of this patient was unclear to the interviewer; however, upon clarification with the physician who treated the patient, confirmation was received that the patient was treated with the drug under review (durvalumab and tremelimumab in combination with platinum-based chemotherapy). While on chemotherapy, the patient received appropriate care and dealt with minimal side effects, including mild diarrhea, bone pain, itching and tiredness, especially on the days of chemotherapy. While on durvalumab, the patient reported experiencing temporary side effects, such as occasional diarrhea, itching, and hot flashes. The patient emphasized that they were still able to carry out regular activities during treatment.

Respondents from a previous survey and submission on durvalumab in combination with chemotherapy as neoadjuvant treatment, followed by durvalumab as monotherapy after surgery indicated for the treatment of patients with NSCLC and resectable tumours (≥ 4 cm and/or node-positive) and no known EGFR or ALK rearrangements, shared their experiences with lung cancer and durvalumab. There were 5 respondents, including 2 with stage IVA lung cancer, 1 with stage IVB, and 2 others (1 with stage IV that had spread only to the brain and 1 with stage IV metastatic NSCLC). The following symptoms affecting QoL were reported: fatigue; pain in chest, shoulder, back or arms; shortness of breath; infections; pneumonia and/or bronchitis; and loss of weight, appetite, hair, and teeth. Respondents identified chemotherapy, immunotherapy, targeted therapy, radiation, surgical therapy, and participation in clinical trials as current treatment options. Respondents noted several unmet needs, including better mental health support and greater availability of immunotherapy for longer durations. In terms of accessing any therapies, 1 respondent reported not being able to acquire counselling and another mentioned burdensome travel costs associated with treatment.

On current treatment regimens, the respondents mentioned several side effects, with 4 in 6 patients reporting joint and muscle pain. Other side effects included fatigue, diarrhea, neuropathy, weight loss, anemia, nausea, vomiting, constipation, migraines, change in vision and hearing, and forgetfulness. Of the 5 respondents, 2 described their AEs as well tolerated. The respondents indicated they were managing their current treatments generally well.

The respondents rated several outcomes that they hoped to see in newer treatments to manage their disease (from most to least important based on average ranking): prolonged life, provision of a cure, QoL, delay in onset of symptoms, ease of use and reduction in side effects from medications or treatments, and access to new options for treatment.

Fatigue was noted as the main side effect for treatment with durvalumab. One of the respondents also mentioned the development of hives. For symptom management, 2 of the 5 respondents noted it was much better, and 1 of the 5 respondents noted little or no difference. For side effects, 3 of the 5 respondents noted little or no difference. For ease of use, 2 of the 5 respondents noted little or no difference. For disease progression, 1 of the 5 respondents noted it was much better, and 2 of the 5 respondents noted little or no difference.

In comparison to other therapies, at least 2 respondents indicated that symptom management was much better with durvalumab; however, little to no difference was reported for side effects, ease of use, and disease progression.

Overall, the patient group highlighted the need for treatments that improve survival (6 months or greater) regardless of the level of side effects. They also noted that patients experienced fewer AEs while receiving durvalumab and felt better versus their experience with previous lines of care.

Clinician Input

Input From Clinical Experts Consulted for This Review

All CDA-AMC review teams include at least 1 clinical specialist with expertise in the diagnosis and management of the condition for which the drug is indicated. Clinical experts are a critical part of the review team and are involved in all phases of the review process (e.g., providing guidance on the development of the review protocol, assisting in the critical appraisal of clinical evidence, interpreting the clinical relevance of the results, and providing guidance on the potential place in therapy). The following input was provided by 2 clinical specialists with expertise in the diagnosis and management of NSCLC.

Unmet Needs

The goals of first-line treatment of metastatic NSCLC identified by the clinical experts consulted by CDA-AMC were to improve survival, delay progression of cancer, decrease cancer-related symptoms, and improve QoL. The clinical experts acknowledged that currently available first-line treatment options for metastatic NSCLC have been associated with a survival benefit compared to chemotherapy alone but noted that there remains an unmet need with overall poor outcomes for the great majority of patients with advanced NSCLC, particularly among those whose tumours have low or negative tumoural PD-L1 expression, who are less likely to have a durable benefit from single-drug immuno-oncology treatment plus chemotherapy. The clinical experts also expressed the need for treatment options in patients with low or negative tumoural PD-L1 expression who have lower rates of immune-related AEs. The experts recognized that there is significant unmet need among subgroups of patients with mutations associated with a lower chance of responding to immuno-oncology therapy or with a generally poor prognosis, including KRAS, STK11, and KEAP mutations.

Place in Therapy

The clinical experts noted that durvalumab and tremelimumab plus platinum-based chemotherapy would be used as a standalone treatment for patients with previously untreated metastatic NSCLC not harbouring a targetable oncogenic aberration with an associated SOC first-line targeted therapy option (e.g., an EGFR mutation or ALK rearrangement) and without any absolute contraindications. The experts did not anticipate that durvalumab and tremelimumab plus platinum-based chemotherapy would shift the current treatment paradigm; instead, this regimen would fit into the treatment algorithm as an additional first-line option combining immuno-oncology treatment and chemotherapy.

Patient Population

In describing patients who would be best and least suited for treatment with durvalumab and tremelimumab plus chemotherapy, the clinical experts stated that there is no reliable biomarker that can be used to predict which patients will respond to this treatment. The experts noted that patients ideally suited for this treatment would have previously untreated metastatic NSCLC, a good ECOG PS (0 or 1), no actionable oncogenic mutations or translocations (including EGFR mutations and ALK rearrangements), any tumoural PD-L1 expression, no significant comorbidities (i.e., uncontrolled autoimmune or inflammatory diseases), and the ability to recognize and report potential adverse effects to their oncology care team in a timely manner. The experts stated that the eligibility criteria for the POSEIDON trial were aligned with those used in clinical trials in this patient population, but added that certain patients excluded from the trial would be considered for treatment with immuno-oncology therapy in clinical practice, including those with an ECOG PS of 2, those with autoimmune or inflammatory disorders or with infections such as HIV or hepatitis C, a subpopulation with prior exposure to immunotherapy for earlier-stage cancer, and a broader population of patients with brain metastases. The clinical experts pointed out that, in practice, durvalumab and tremelimumab plus platinum-based chemotherapy is anticipated to be used most often in patients with PD-L1 levels below 50%, but they noted that, because patients with any PD-L1 status were included in the POSEIDON trial, approval and funding should be in accordance with the trial outcomes. The experts stated that durvalumab and tremelimumab plus platinum-based chemotherapy may be an option for patients whose tumours do not express PD-L1 (based on a post hoc analysis showing that these patients may derive the most benefit from the addition of an anti–CTLA-4 antibody to an anti–PD-1 antibody), noting that reliable biomarkers to determine which patients will derive the most benefit are lacking. The experts identified durvalumab and tremelimumab plus platinum-based chemotherapy as an option for patients with KRAS, STK11, or KEAP mutations, while acknowledging that the current evidence is post hoc and hypothesis-generating. The treatment strategy for the indicated population would be based on an informed decision by the patients in consultation with a medical oncologist. According to the clinical experts, misdiagnosis of advanced NSCLC or errors in biomarker testing does not occur often because of rigorous quality control measures in place at pathology labs in Canada that perform cancer diagnoses and biomarker testing. The experts expected that all centres would have access to routine testing for KRAS, EGFR, and ALK mutations and PD-L1 status but that testing for STK11 and KEAP mutations may vary by centre.

Assessing the Response to Treatment

The clinical experts commented that patients with metastatic NSCLC receiving systemic therapy would typically be clinically assessed before each treatment (every 3 weeks) by a medical oncologist, general practitioner in oncology, or nurse practitioner with special training in medical oncology. Radiological assessment for response to treatment would typically be performed every 2 to 3 months by CT scan, and less commonly with MRI or PET-CT scan, unless symptoms suggesting disease progression prompt earlier investigations. The clinical experts noted that surveillance imaging in clinical trials is typically performed every 6 to 9 weeks initially. The experts did not expect that results would vary among physicians. Clinically meaningful end points identified by the clinical experts are improvement in survival, decrease in cancer-related symptoms, and improvement in QoL. Tumour evaluation in the POSEIDON trial was performed by CT scan or MRI at weeks 6 and 12 after randomization and every 8 weeks thereafter.³⁸

Discontinuing Treatment

The clinical experts stated that discontinuation of first-line treatment for metastatic NSCLC would be considered when there is significant disease progression or unacceptable toxicity, or if the patient chooses to stop treatment. One clinical expert added that discontinuation of treatment may be considered when there is suspicion of progression and a risk of rapid clinical deterioration if treatment is continued. Clinicians treating NSCLC already have experience diagnosing and managing immune-related adverse effects of ICIs in clinical practice, and the clinical experts stated that immuno-oncology treatment may be held temporarily but can often be safety resumed (even for high-grade immune-related adverse effects) and that permanent discontinuation may not be needed. However, they also pointed out that rechallenge with immunotherapy would typically not be undertaken if the AE was life-threatening. The clinical experts noted that, upon development of an AE, the risks and benefits of resuming treatment would be considered through consultation between the clinician and patient.

Prescribing Considerations

The clinical experts noted that first-line treatments for metastatic NSCLC would typically be initiated following consultation with a medical oncologist at a regional cancer clinic. Treatment would be administered and monitoring conducted locally for patients who live close to the regional clinic, and often at community oncology networks for patients living in remote settings, where specially trained general practitioners or internists would oversee day-to-day management of the treatment protocol. The clinical experts agreed that it would be appropriate for any systemic therapy unit and lung cancer treatment team currently administering other combinations of immuno-oncology and chemotherapy to deliver treatment with durvalumab and tremelimumab plus platinum-based chemotherapy. Clinical experts added that all patients must be closely monitored to detect early evidence of toxicity or progression, and that consultation with a subspecialist (e.g., a gastroenterologist or hepatologist in the event of liver toxicity) may be required if there is evidence of toxicity.

Additional Considerations

CDA-AMC approved a deviation request from the sponsor allowing for the exclusion of monotherapy with immuno-oncology drugs (i.e., those given without platinum-based chemotherapy) as comparators in both the pharmacoeconomic model and budget impact analysis. The sponsor’s rationale was that the proposed place in therapy for durvalumab and tremelimumab plus platinum-based chemotherapy is the same as other immuno-oncology drugs that are indicated for use in combination with platinum-based chemotherapy, and as such, the sponsor proposed that the use of durvalumab and tremelimumab plus platinum-based chemotherapy would only displace the use of other regimens that include platinum-based chemotherapy. The clinical experts described immuno-oncology monotherapy as a potential comparator for the combination of durvalumab and tremelimumab plus platinum-based chemotherapy. They also noted that, although durvalumab and tremelimumab plus platinum-based chemotherapy is anticipated to be used most often in patients with tumoural PD-L1 expression below 50% in clinical practice, approval and funding should be based on the outcomes of the POSEIDON trial, adding that patients with any PD-L1 status were included in the study. The experts stated that platinum-based chemotherapy alone does not represent a reasonable comparator to durvalumab and tremelimumab plus platinum-based chemotherapy because immuno-oncology treatment is the backbone of therapy for first-line treatment of metastatic NSCLC. Platinum-based chemotherapy alone would typically only be given to patients who are not able to or choose not to receive immuno-oncology treatment, and these patients would not be eligible for treatment with the regimen used in the POSEIDON trial. The clinical experts pointed out that platinum-based chemotherapy has been shown to be inferior to immuno-oncology treatment plus platinum-based chemotherapy or immuno-oncology monotherapy (depending on the clinical trial).

Clinician Group Input

This section was prepared by the review team based on input provided by clinician groups.

Two clinician groups, the LLC-MAC and the Cancer Care Ontario DAC, provided input for this submission. The LCC-MAC is a national charity providing support and education, supporting research, and providing clinician input for submissions to reviews of new lung cancer drugs by the HTA process for many years. The DAC provides evidence-based clinical and health system guidance on drug-related issues to support the mandate of Cancer Care Ontario, including the Provincial Drug Reimbursement Programs and the Systemic Treatment Program.

A total of 19 clinicians from the LCC-MAC and 5 clinicians from the DAC provided input for this submission. The submission from LCC-MAC was informed by data and information from publicly available sources, primarily published manuscripts and conference presentations, together with clinical experience of the LCC-MAC members. Information from the DAC was gathered via teleconference meeting and emails.

The LCC-MAC indicated that multiple treatment options are available in Canada for patients without driver mutations (e.g., EGFR mutations or ALK translocations). In addition, they noted that, for those with tumours with PD-L1 expression of any level or unknown, current treatment options include platinum doublet (cisplatin or carboplatin with pemetrexed) plus pembrolizumab (nonsquamous histology), platinum doublet (carboplatin and paclitaxel) plus pembrolizumab (squamous histology), or chemotherapy-sparing combinations of 2 cycles of platinum doublet with doublet immunotherapy in the form of nivolumab and ipilimumab (most often used in PD-L1–negative or squamous histology). Both clinician groups acknowledged the use of single-drug pembrolizumab in patients with a PD-L1 expression level of 50% or greater. The DAC noted current treatment options included chemotherapy with pembrolizumab, and chemotherapy with nivolumab and ipilimumab.

The DAC emphasized that the treatment combination of durvalumab with tremelimumab and platinum-based chemotherapy would serve as a first-line treatment. The LCC-MAC noted that most thoracic oncologists in Canada have experience with using the treatment combination through clinical trials. They also pointed out that the treatment combination would not replace other regimens that were already approved or funded but would serve as an alternative. Chemotherapy and immunotherapy options were described by the LCC-MAC as appropriate comparators to the treatment combination.

In terms of unmet needs, both clinician groups agreed that treatments with better tolerability are needed. The DAC also noted that not all patients respond to available treatments, patients become refractory to current treatment options, and treatments and formulations that improve compliance and convenience are needed. Important treatment goals identified by both clinician groups were tumour shrinkage, improvement of symptoms and QoL, and improved OS, with the LCC-MAC adding that improvement in PFS is an important goal. The clinician groups further noted that an important aspect to be looked at with immunotherapy combinations was landmark survival with the goal of improving the number of patients with durable benefits of treatment as manifested in increases in PFS and OS at 5 years.

The clinicians indicated that patients without actionable driver mutations (i.e., no EGFR or ALK alterations, no ROS and RET fusions) would be best suited for treatment. The DAC noted that patients with stage IV or incurable NSCLC considering first-line therapy would be suitable candidates for treatment. In addition, the LCC-MAC noted that patients in the POSEIDON trial with KRAS, STK11, and KEAP mutations, which are not currently actionable in the first-line setting, seem to benefit from the combination. Both groups highlighted the need for clinical assessment of symptoms and imaging (e.g., CT scans and chest X-rays) to monitor response to treatment. The DAC advised assessing treatment response every 6 weeks initially, and then less often. However, the LCC-MAC suggested that scans be performed every 3 months, with symptom assessments every 3 to 4 weeks, depending on whether a patient receives chemotherapy or single-drug immunotherapy during the maintenance phase.

The LCC-MAC also noted that this treatment regimen would be an appropriate choice for patients without an actionable driver mutation with any PD-L1 status. The groups agreed that disease progression is a factor when deciding whether to discontinue treatment. The LCC-MAC indicated that immunotherapy may be interrupted or discontinued because of imAEs. They added that most low-grade toxicities were manageable with symptom management, topical or oral steroids, and potential treatment interruption. The DAC also identified intolerable side effects and patient withdrawal as factors when considering treatment discontinuation.

The LCC-MAC indicated that initial patient assessments should be conducted by a medical oncologist. They noted that, with high rates of lung cancer, Canada has an established network of experts, including family physicians and general practitioners in oncology who can assist in the assessment of these patients on treatment, particularly in remote areas. Both clinician groups agreed that the treatment combination should be administered in facilities with expertise in managing cytotoxic anticancer therapies.

Drug Program Input

The drug programs provide input on each drug being reviewed through the Reimbursement Review processes by identifying issues that may affect their ability to implement a recommendation. The implementation questions and corresponding responses from the clinical experts consulted by for this review are summarized in Table 6.

Table 6: Summary of Drug Plan Input and Clinical Expert Response

AE = adverse event; CDA-AMC = Canada’s Drug Agency; ECOG PS = Eastern Cooperative Oncology Group Performance Status; NSCLC = non–small cell lung cancer; PBC = platinum-based chemotherapy; pERC = pan-Canadian Oncology Drug Review Expert Review Committee.

Clinical Evidence

The objective of this Clinical Review report is to review and critically appraise the clinical evidence submitted by the sponsor on the beneficial and harmful effects of durvalumab (50 mg/mL, concentrate for IV infusion) in combination with tremelimumab (20 mg/mL, concentrate for IV infusion) and platinum-based chemotherapy in the first-line treatment of adult patients with metastatic NSCLC but no sensitizing EGFR mutations or ALK genomic tumour aberrations. The focus is on comparing durvalumab in combination with tremelimumab and platinum-based chemotherapy to relevant comparators and identifying gaps in the current evidence.

A summary of the clinical evidence included by the sponsor in the review of durvalumab in combination with tremelimumab and platinum-based chemotherapy 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 randomized controlled trials (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 third section includes indirect evidence from the sponsor. The sponsor did not include long-term extension studies (second section) or additional studies addressing important gaps in the pivotal and RCT evidence (fourth section).

Included Studies

Clinical evidence from 1 pivotal study identified in systematic review and 1 ITC is included in the review and appraised in this document.

Systematic Review

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

Description of Studies

Characteristics of the included study are summarized in Table 7.

Table 7: Details of Studies Included in the Systematic Review

AE = adverse event; AJCC = American Joint Committee on Cancer; ALT = alanine transaminase; AST = aspartate transaminase; AUC = area under the curve; BICR = blinded independent central review; bTMB = blood tumour mutational burden; CrCL = creatinine clearance; ECOG PS = Eastern Cooperative Oncology Group Performance Status; EORTC QLQ-C30 = European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30; EORTC QLQ-LC13 = European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Lung Cancer 13; HBsAg = hepatitis B surface antigen; HBV = hepatitis B virus; HCV = hepatitis C virus; IASLC = International Association for the Study of Lung Cancer; ICF = informed consent form; ILD = interstitial lung disease; NSCLC = non–small cell lung cancer; OS = overall survival; PFS = progression-free survival; PFS2: time from randomization to second progression; q.3.w. = every 3 weeks; q.4.w = every 4 weeks; RECIST 1.1 = Response Evaluation Criteria in Solid Tumours Version 1.1; ULN = upper limit normal; SAE = serious adverse event; vs. = versus.

^aEfficacy data for the durvalumab plus chemotherapy arm of the POSEIDON trial will not be presented in the systematic review because this combination is not seeking approval in Canada and therefore is not aligned with the Health Canada product monograph.

^bThe IASLC is the group responsible for conducting research to inform future versions of the AJCC. Version 8 of the IASLC Staging Manual in Thoracic Oncology was endorsed and published by the AJCC as the AJCC 8th edition.^15,43,44 The 2 are considered interchangeable.^45,46

^cPatients whose weight fell to 30 kg or below received weight-based dosing of 20 mg/kg of durvalumab and 1 mg/kg of tremelimumab q.3.w. (in the combination stage) or q.4.w. (in the maintenance stage) until their weight improved to greater than 30 kg, at which point the patient started receiving the fixed dosing of 1,500 mg durvalumab at 75 mg tremelimumab.

^dPatients were treated until clinical or radiological progression unless there was unacceptable toxicity, consent withdrawal, or another discontinuation criterion was met.

^ePatients with nonsquamous tumour histology who received carboplatin or cisplatin plus pemetrexed and who had not progressed after 4 cycles of carboplatin or cisplatin plus pemetrexed could receive pemetrexed maintenance therapy, unless contraindicated as determined by the investigator.

^fPatients whose weight fell to 30 kg or below received weight-based dosing of 20 mg/kg of durvalumab q.3.w. (in the combination stage) or q.4.w. (in the maintenance stage) until the weight improved to greater than 30 kg, at which point the patient started receiving the fixed dosing of durvalumab at 1,500 mg.

Sources: Sponsor’s Summary of Clinical Evidence,¹ POSEIDON Clinical Study Report,³⁸ POSEIDON Clinical Study Protocol,⁴⁷ and POSEIDON Statistical Analysis Plan.⁴⁸

The POSEIDON trial was a phase III, randomized, open-label, multicentre, global study investigating the efficacy and safety of durvalumab and tremelimumab plus SOC chemotherapy versus SOC chemotherapy alone or durvalumab monotherapy plus SOC chemotherapy versus SOC chemotherapy alone for the first-line treatment in patients (total N = 1,013) with metastatic NSCLC with tumours without activating EGFR mutations and ALK fusions.^1,38,47 There were no study sites in Canada.^1,38

Patients were randomized 1:1:1 by interactive voice response system (IVRS) or interactive web response system (IWRS) over a 15-month period to 1 of 3 study arms: durvalumab and tremelimumab plus SOC chemotherapy (arm 1; n = 338); durvalumab plus SOC chemotherapy (arm 2; n = 338); or SOC chemotherapy alone (arm 3; n = 337). SOC chemotherapy consisted of the investigator’s choice of 1 of the following regimens: nab-paclitaxel plus carboplatin (patients with squamous and nonsquamous tumour histology); pemetrexed plus cisplatin or carboplatin (patients with nonsquamous tumour histology only); or gemcitabine plus cisplatin or carboplatin (patients with squamous tumour histology only). Randomization was stratified according to PD-L1 tumour-expression status (TCs ≥ 50% versus < 50%), disease stage (IVA versus IVB), and histology (nonsquamous versus squamous).^1,38 The study design for the POSEIDON trial is presented in Figure 2.

The primary objective of the POSEIDON trial was to assess the efficacy of durvalumab plus SOC chemotherapy compared to SOC chemotherapy alone in terms of dual primary end points of PFS and OS in all patients. The key secondary objectives were to assess the efficacy of durvalumab and tremelimumab plus SOC chemotherapy compared to SOC chemotherapy alone in terms of PFS and OS in all patients.^1,38

The study arm that evaluated durvalumab plus SOC chemotherapy is not aligned with the indication approved by Health Canada and therefore results for this arm are not reported in this CDA-AMC review. Only results from the comparison of durvalumab and tremelimumab plus SOC chemotherapy versus SOC chemotherapy alone are presented in this review and include results from the final analysis (DCO: July 24, 2019, for PFS and other RECIST 1.1–related end points; DCO: March 12, 2021, for OS, safety, and all other data). Results from an updated 5-year OS analysis (DCO: August 24, 2023) have also been reported for this comparison.^1,38 The POSEIDON trial was supported by AstraZeneca.³⁹

The POSEIDON trial included a 4-week screening period to complete eligibility assessments, during which informed consent was obtained, patient data and medical history were obtained, and various laboratory and baseline assessments were performed.⁴⁷ The choice of SOC chemotherapy was patient-specific based on the treatment most appropriate for the patient and was selected before receiving a randomization number in the IVRS or IWRS. The trial comprised 2 treatment periods following randomization, referred to as the combination (during chemotherapy) stage and the maintenance (post−chemotherapy) stage. In the combination stage, patients in arm 1 received treatment with 4 cycles of durvalumab and tremelimumab plus SOC chemotherapy and patients in arm 3 received 4 cycles of SOC chemotherapy. In the maintenance stage, patients in arm 1 received treatment with durvalumab until disease progression as well a single additional dose of tremelimumab; patients with nonsquamous histology may have also received pemetrexed maintenance therapy until disease progression. In the maintenance stage, patients in arm 3 could receive an additional 2 cycles of SOC chemotherapy, if clinically indicated, before entering follow-up; patients with nonsquamous histology could also receive pemetrexed maintenance therapy. Dose reductions of durvalumab and tremelimumab were not permitted.^1,38 Crossover was not permitted as part of the study.³⁸ Continuation of durvalumab monotherapy beyond disease progression was permitted for patients who continued to receive a benefit and met the criteria to remain on treatment. Patients who had received 5 cycles of durvalumab plus tremelimumab and then experienced disease progression during durvalumab monotherapy were permitted to receive re-treatment once with up to 4 additional cycles of tremelimumab alongside durvalumab.^1,39,47 Tumour evaluation for patients in all treatment arms was by CT scan or MRI, with scans performed at screening (as baseline), week 6 (± 1 week) and week 12 (± 1 week) from the date of randomization, and then every 8 weeks (± 1 week) until radiological progression.^1,38,47

Figure 2: Study Design for the POSEIDON Trial

ECOG PS = Eastern Cooperative Oncology Group Performance Status; NSCLC = non–small cell lung cancer; q3w = every 3 weeks; q4w = every 4 weeks; TC = tumour cell.

* Chemotherapy options comprised nab-paclitaxel plus carboplatin (patients with either squamous or nonsquamous histology); gemcitabine plus cisplatin or carboplatin (patients with squamous histology); and pemetrexed plus cisplatin or carboplatin (patients with nonsquamous histology). The choice of chemotherapy was at the discretion of the investigator.

^† Patients received an additional dose of tremelimumab postchemotherapy (total of 5 durvalumab plus tremelimumab combination doses).

^‡ Patients with nonsquamous histology who received pemetrexed-platinum doublet and who had not progressed after 4 to 6 cycles received pemetrexed maintenance therapy unless contraindicated by the investigator.

^§ Patients continued treatment until disease progression, unacceptable toxicity, or withdrawal of consent. In the durvalumab plus tremelimumab and chemotherapy arm and the durvalumab plus chemotherapy arm, continuation of durvalumab monotherapy after objective radiological progression was permitted at the investigator’s discretion if there was evidence of clinical benefit and the patient met the criteria for treatment in the setting of progressive disease. Patients allocated to durvalumab plus tremelimumab and chemotherapy who had completed 5 dosing cycles of durvalumab plus tremelimumab combination therapy with clinical benefit per investigator judgment, but subsequently had radiological progression during durvalumab monotherapy, could receive re-treatment with up to 4 additional cycles of tremelimumab alongside durvalumab.

^¶ Anticancer therapy utilized after discontinuation of study treatment was recorded as part of the follow-up schedule.

Source: Sponsor’s Summary of Clinical Evidence.¹

Populations

Inclusion and Exclusion Criteria

Detailed key inclusion and exclusion criteria are provided in Table 7. Patients eligible for participation in the POSEIDON trial were aged 18 years or older with histologically or cytologically documented stage IV NSCLC (according to Version 8 of the IASLC Staging Manual in Thoracic Oncology [2016]⁴⁴) not amenable to curative surgery or radiation with tumours that lacked activating EGFR mutations and ALK fusions. Patients had not received prior chemotherapy or any other systemic therapy for metastatic NSCLC. Patients who had received prior platinum-containing adjuvant, neoadjuvant, or definitive chemoradiation for advanced disease were eligible if progression had occurred more than 12 months after completing therapy. Patients had no prior exposure to immune-mediated therapy (including other anti–CTLA-4, anti–PD-1, anti–PD-L1, and anti–PD-L2 antibodies). Participants were required to have a WHO performance status or ECOG PS of 0 or 1. Patients who had suspected brain metastases at screening were required to have an IV contrast-enhanced MRI or CT scan of the brain before study entry. Those with brain metastases were required to be treated before randomization and were eligible if a stable condition was confirmed after the intervention, there was a return neurologically to baseline, and associated steroid treatment had been completed at least 5 days before randomization. Patients with active or prior documented autoimmune or inflammatory disorders (e.g., colitis, Crohn disease, systemic lupus erythematosus, rheumatoid arthritis), with some exceptions (e.g., vitiligo, alopecia, and chronic skin conditions not requiring systemic treatment), were not eligible for the trial.^1,38

Interventions

In the POSEIDON trial, patients were randomized 1:1:1, to 1 of 3 treatment arms:^1,38

• arm 1: Durvalumab and tremelimumab plus SOC chemotherapy (n = 338)

• arm 2: Durvalumab plus SOC chemotherapy (n = 338)

• arm 3: SOC chemotherapy alone (n = 337).

Arm 1: Durvalumab and Tremelimumab Plus SOC Chemotherapy

Combination (During Chemotherapy) Stage^1,38

• Durvalumab, 1,500 mg by 60-minute IV infusion, every 3 weeks at weeks 0, 3, 6, and 9

• Tremelimumab, 75 mg by 60-minute IV infusion, every 3 weeks at weeks 0, 3, 6, and 9

• SOC chemotherapy for 4 cycles:

  • nab-paclitaxel (patients with squamous and nonsquamous tumour histology) (100 mg/m²) IV infusion on days 1, 8, and 15 of each 21-day cycle plus carboplatin (area under the plasma drug concentration-time curve [AUC] 5 or 6) IV infusion on day 1 of each 21-day cycle, for 4 cycles; or

  • gemcitabine (patients with squamous tumour histology only) (1,000 mg/m² or 1,250 mg/m²) IV infusion on days 1 and 8 of each 21-day cycle plus cisplatin (75 mg/m²) or carboplatin (AUC 5 or 6) IV infusion on day 1 of each 21-day cycle, for 4 cycles; or

  • pemetrexed (patients with nonsquamous tumour histology only) (500 mg/m²) IV infusion plus cisplatin (75 mg/m²) or carboplatin (AUC 5 or 6) IV infusion on day 1 of each 21-day cycle, for 4 cycles.

If a dosing delay occurred during chemotherapy while on the every 3 weeks schedule, all future dosing days were delayed to ensure that the intervals between study treatment doses were always a minimum of 21 days.^1,38

Maintenance (Postchemotherapy) Stage^1,38

• Durvalumab, 1,500 mg by 60-minute IV infusion, every 4 weeks from week 12 to disease progression

  • Patients were treated until clinical or radiological progression unless there was unacceptable toxicity, consent withdrawal, or another criterion for discontinuation was met.

• Tremelimumab, 75 mg by 60-minute IV infusion, 1 dose at week 16

  • In the case of dose delay(s), more than 1 tremelimumab plus durvalumab combination dose could be administered at and after week 16 postchemotherapy to ensure that up to 5 combination doses were given. If patients received fewer than 4 cycles of platinum doublet chemotherapy, the remaining cycles of combined durvalumab and tremelimumab (up to a total of 5) were to be given after a combination of platinum doublet chemotherapy (with maintenance pemetrexed, if applicable).

• Pemetrexed (patients with nonsquamous tumour histology only), 500 mg/m² by IV infusion, every 4 weeks from week 12 to disease progression

  • Patients with nonsquamous tumour histology who received pemetrexed plus carboplatin or cisplatin and who had not progressed after 4 cycles of pemetrexed plus carboplatin or cisplatin could receive maintenance therapy with pemetrexed, unless contraindicated as determined by the investigator.

  • Patients were treated until clinical progression or RECIST 1.1–defined radiological progression unless there was unacceptable toxicity or consent withdrawal or another criterion for discontinuation was met.

In both stages, weight-based doses were received by patients whose weight fell to 30 kg or less, equivalent to 20 mg/kg of durvalumab and 1 mg/kg of tremelimumab (every 3 weeks in the combination stage, every 4 weeks in the maintenance stage) until the weight improved to greater than 30 kg, at which point the fixed dosing of durvalumab at 1,500 mg and tremelimumab at 75 mg was started.³⁸ Dose reductions of durvalumab and tremelimumab were not permitted.^1,38

Arm 2: Durvalumab Plus SOC Chemotherapy

Combination (During Chemotherapy) Stage^1,38

• Durvalumab, 1,500 mg by 60-minute IV infusion, every 3 weeks at weeks 0, 3, 6, and 9

• SOC chemotherapy for 4 cycles

  • nab-paclitaxel (patients with squamous and nonsquamous tumour histology) (100 mg/m²) IV infusion on days 1, 8, and 15 of each 21-day cycle plus carboplatin (AUC 5 or 6) IV infusion on day 1 of each 21-day cycle, for 4 cycles; or

  • gemcitabine (patients with squamous tumour histology only) (1,000 mg/m² or 1,250 mg/m²) IV infusion on days 1 and 8 of each 21-day cycle plus cisplatin (75 mg/m²) or carboplatin (AUC 5 or 6) IV infusion on day 1 of each 21-day cycle, for 4 cycles; or

  • pemetrexed (patients with nonsquamous tumour histology only) (500 mg/m²) IV infusion plus cisplatin (75 mg/m²) or carboplatin (AUC 5 or 6) IV infusion on day 1 of each 21-day cycle, for 4 cycles.

If a dosing delay occurred during chemotherapy while on the every 3 weeks schedule, all future dosing days were delayed to ensure that the intervals between dosing study treatment were always a minimum of 21 days.^1,47

Maintenance (Postchemotherapy) Stage^1,38

• Durvalumab, 1,500 mg by 60-minute IV infusion, every 4 weeks from week 12 to disease progression

  • Patients were treated until clinical or radiological progression unless there was unacceptable toxicity, consent withdrawal, or another criterion for discontinuation was met.

• Pemetrexed (patients with nonsquamous tumour histology only), 500 mg/m² by IV infusion, every 4 weeks from week 12 to disease progression

  • Patients with nonsquamous tumour histology who received pemetrexed plus carboplatin or cisplatin and who had not progressed after 4 cycles of pemetrexed plus carboplatin or cisplatin could receive maintenance therapy with pemetrexed, unless contraindicated per the investigator.

  • Patients were treated until clinical progression or RECIST 1.1–defined radiological progression unless there was unacceptable toxicity or consent withdrawal or another criterion for discontinuation was met.

In both stages, weight-based dosing was received by patients whose weight fell to 30 kg or less, equivalent to 20 mg/kg of durvalumab every 3 weeks in the combination stage and every 4 weeks in the maintenance stage, until the weight improved to greater than 30 kg, at which point the fixed dosing of durvalumab at 1,500 mg was started.³⁸ Dose reductions of durvalumab were not permitted.^1,38

Results from the durvalumab plus SOC chemotherapy arm are not included in the systematic review because this regimen is not part of the Health Canada–approved indication and reimbursement request under review by CDA-AMC.

Arm 3: SOC Chemotherapy^1,38

• Nab-paclitaxel (patients with squamous and nonsquamous tumour histology) (100 mg/m²) IV infusion on days 1, 8, and 15 of each 21-day cycle plus carboplatin (AUC 5 or 6) IV infusion on day 1 of each 21-day cycle, for 4 cycles

• Gemcitabine (patients with squamous tumour histology only) (1,000 mg/m² or 1,250 mg/m²) IV infusion on days 1 and 8 of each 21-day cycle plus cisplatin (75 mg/m²) or carboplatin (AUC 5 or 6) IV infusion on day 1 of each 21-day cycle, for 4 cycles

• Pemetrexed (patients with nonsquamous tumour histology only) (500 mg/m²) IV infusion plus cisplatin (75 mg/m²) or carboplatin (AUC 5 or 6) IV infusion on day 1 of each 21-day cycle, for 4 cycles

An additional 2 doses of SOC chemotherapy (weeks 12 and 15) could have been given at the investigator’s discretion, if clinically indicated.

Patients with nonsquamous tumour histology who received pemetrexed plus carboplatin or cisplatin and who had not progressed after 4 to 6 cycles of pemetrexed plus carboplatin or cisplatin could receive maintenance therapy with pemetrexed given every 3 weeks or every 4 weeks.

If a dosing delay occurred during chemotherapy while on the every 3 weeks schedule, all future dosing days were delayed to ensure that the intervals between dosing study treatment were always a minimum of 21 days.³⁸

A summary of the overall dosing scheme for POSEIDON trial is provided in Table 8.

Table 8: POSEIDON Trial Dosing Scheme

D = durvalumab; peme = pemetrexed; q.3.w. = every 3 weeks; q.4.w. = every 4 weeks; t = tremelimumab.

Note: Response Evaluation Criteria in Solid Tumours Version 1.1 assessment was performed at week 12 ± 1 week from the date of randomization, and every 8 weeks ± 1 week thereafter until radiological progression (regardless of whether q.3.w. or q.4.w. regimen was chosen). Weight-based dosing was received by patients whose weight fell to 30 kg or less, equivalent to durvalumab 20 mg/kg and tremelimumab 1 mg/kg until the weight improved to greater than 30 kg, at which point the fixed dosing of durvalumab 1,500 mg and tremelimumab 75 mg was started. The durvalumab dose was 1,500 mg during chemotherapy and 1,500 mg postchemotherapy; the tremelimumab dose was 75 mg.

^aPemetrexed maintenance therapy was for patients with nonsquamous NSCLC who received treatment with pemetrexed and carboplatin or cisplatin during chemotherapy and did not progress after 4 to 6 cycles, unless contraindicated as determined by the investigator. Pemetrexed maintenance therapy could have been given q.3.w. or q.4.w. (i.e., q.4.w. for arms 1 and 2; for arm 3, pemetrexed maintenance therapy could have been given either q.3.w. or q.4.w. depending on investigator decision and local standards, unless contraindicated per the investigator).

^bFor patients in arm 1, an additional dose of durvalumab and tremelimumab was given at week 16 postchemotherapy. In the case of dose delay(s), more than 1 durvalumab and tremelimumab combination dose could have been given at and after week 16 postchemotherapy to ensure that up to 5 combination doses were administered in arm 1. If patients received fewer than 4 cycles of platinum doublet chemotherapy, the remaining cycles of combined durvalumab and tremelimumab (up to a total of 5) were to be given after combination of platinum doublet chemotherapy (with maintenance pemetrexed, if applicable).

^cThe durvalumab plus chemotherapy arm is not included in the systematic review because this regimen is not approved by Health Canada.

^dIn arm 3, chemotherapy could have been given for an additional 2 cycles q.3.w. on weeks 12 and 15 (i.e., a total of 6 cycles postrandomization) if clinically indicated, at the investigator’s discretion before patients entered follow-up. This did not alter the planned scan schedule of every 8 weeks starting at week 12 for patients in arm 3.

Sources: Sponsor’s Summary of Clinical Evidence¹ and POSEIDON Clinical Study Report.³⁸

Duration of Treatment

Treatment with chemotherapy in arms 1 and 2 was limited to 4 cycles administered at 3-week intervals subsequent to randomization. Patients in arm 3 (SOC chemotherapy alone) could receive an additional 2 doses of SOC chemotherapy at weeks 12 and 15 (a total of 6 doses postrandomization), as clinically indicated, at the investigator’s discretion. Treatment with immunotherapy plus SOC chemotherapy in arms 1 and 2, as well as treatment with SOC chemotherapy alone in arm 3, was administered beginning on cycle 1 day 1.^1,38,47

For patients randomized to arms 1 and 2, treatment with durvalumab monotherapy continued until clinical or radiological progression, unless there was unacceptable toxicity or consent withdrawal, or another criterion for discontinuation was met. Patients in arm 1 received an additional dose of durvalumab and tremelimumab at week 16 postchemotherapy. If dose delay(s) occurred, more than 1 dose of the durvalumab and tremelimumab combination could have been given at and after week 16 postchemotherapy to ensure that up to 5 combination doses were administered in arm 1. If patients received fewer than 4 cycles of platinum doublet chemotherapy, the remaining cycles of combined durvalumab and tremelimumab (up to a total of 5) were to be given after a combination of platinum doublet chemotherapy (with maintenance pemetrexed if applicable). All patients with nonsquamous tumour histology who received a pemetrexed doublet in the initial part of the study were to receive maintenance therapy with pemetrexed in the postchemotherapy phase of the study, unless contraindicated as determined by the investigator.^1,38,47

In arms 1 and 2, when SOC chemotherapy was discontinued because of treatment-related toxicity, patients could continue to receive durvalumab monotherapy or durvalumab plus tremelimumab at the investigator’s discretion when toxicity resolved to at least grade 2 or less.^1,38,47

Treatment Through Progression (Arms 1 and 2)

In arms 1 and 2, patients with objective radiological progression who continued to receive a benefit from their assigned treatment (in the investigator’s opinion) and who met the criteria for treatment in the setting of disease progression could continue to receive durvalumab monotherapy for as long as they were gaining clinical benefit.^1,38,47

Patients with rapid tumour progression or with symptomatic progression requiring urgent medical intervention (e.g., CNS metastasis, respiratory failure because of tumour compression, spinal cord compression) were not eligible to continue any study treatment.^1,38,47

In arms 1 and 2, the investigator ensured that all patients who were treated through progression did not have any significant, unacceptable, or irreversible toxicities that indicated continuing treatment would not further benefit the patient. Patients could continue receiving their assigned treatment in the setting of unconfirmed disease progression, at the investigator’s discretion, until disease progression was confirmed with a subsequent scan.^1,38,47

Re-Treatment (Arm 1)

Patients in arm 1 (durvalumab and tremelimumab plus SOC chemotherapy) with radiological progression who, in the investigator’s opinion, continued to receive a benefit from their assigned treatment and who met the criteria for re-treatment in the setting of disease progression, could have re-treatment with durvalumab and tremelimumab combination therapy.^1,38 Patients continued durvalumab dosing at 1,500 mg every 4 weeks with tremelimumab at 75 mg every 4 weeks for 4 doses per cycle each and then continued with durvalumab monotherapy at 1,500 mg every 4 weeks (with weight-based dosing used for patients whose weight fell to 30 kg or below).⁴⁷

Patients with rapid tumour progression or with symptomatic progression that required urgent medical intervention (e.g., CNS metastasis, respiratory failure because of tumour compression, or spinal cord compression) were not eligible to begin re-treatment with durvalumab and tremelimumab combination therapy.^1,38

In arm 1, the investigator ensured that all patients who began re-treatment did not have any significant, unacceptable, or irreversible toxicities that indicated restarting treatment would not further benefit the patient. Patients who met the re-treatment criteria followed the same treatment guidelines followed during the original, every 4 weeks, postchemotherapy-maintenance treatment period. Those who met the re-treatment criteria were permitted to receive re-treatment only once.^1,38

Treatment Through Progression and Re-Treatment (Arm 3)

In arm 3, treatment through progression and re-treatment was not permitted. Crossing over to durvalumab and tremelimumab plus SOC chemotherapy or durvalumab plus SOC chemotherapy was not permitted following objective disease progression. The investigator could offer other therapies following objective disease progression, and patients were still followed for survival if another treatment for NSCLC was initiated.^1,38,47

Discontinuation of Study Treatment

Reasons for discontinuation of any further study treatment (i.e., durvalumab and tremelimumab combination therapy, durvalumab monotherapy, or SOC chemotherapy) included clinical progression (investigator determination that the patient is no longer benefiting from treatment in arms 1, 2, and 3), radiological progressive disease (based on RECIST 1.1 criteria and investigator determination that the patient is no longer benefiting from treatment in arms 1 and 2), RECIST 1.1-defined radiological progression (arm 3), AE meeting discontinuation criteria, consent withdrawal, and other reasons (e.g., nonadherence to study protocol or initiation of an alternative anticancer therapy).^38,47 Patients who discontinued study treatment for reasons other than objective disease progression continued to receive RECIST 1.1 scans until RECIST 1.1–defined radiological progression plus an additional subsequent scan, or death (whichever came first). Survival status was assessed for all patients until the end of the study, including those who initiated alternative treatment for their NSCLC.³⁸

Concomitant Medications

Concomitant medications that were permitted included:⁴⁷

• medications necessary for prophylactic or supportive care (e.g., acetaminophen, diphenhydramine), except those identified as prohibited

• best supportive care (e.g., antibiotics, optimal symptom control, pain management [including palliative radiotherapy to nontarget lesions])

• inactivated viruses (e.g., influenza vaccine).

Concomitant medications that were prohibited for all treatment arms included:⁴⁷

• other investigational anticancer therapies

• other monoclonal antibodies against CTLA4, PD-1, or PD-L1

• other concurrent chemotherapy, radiotherapy, immunotherapy, or biologic or hormonal therapy for cancer treatment (local treatment of isolated lesions excluding target lesions for palliative intent was acceptable by radiotherapy or surgically)

• live attenuated vaccines (not to be given until 30 days following the last dose of study treatment).

Concomitant medications that were prohibited for arms 1 and 2 included:

• immunosuppressive medications, with some exceptions (e.g., management of study treatment-related AEs; short-term premedication as required by prescribing information for those receiving combination agents SOC; inhaled, topical, or intranasal corticosteroids; temporary management of non–immunotherapy-related events)

• medications with laxative properties, herbal or natural remedies for constipation (use with caution up until 90 days following the last dose of tremelimumab)

• sunitinib (not to be used concomitantly or up until 90 days following the last dose of tremelimumab)

• epidermal growth factor receptor tyrosine kinase inhibitors (not to be used concomitantly and to be used with caution in the 90 days following the last dose of durvalumab)

• herbal or natural remedies with the potential for immune-modulating effects.

Outcomes

A list of efficacy and safety end points included in this Clinical Review report is provided in Table 9, 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 by the clinical experts consulted for this review and input from patient and clinician groups and public drug plans. Using the same considerations, we selected end points considered to be most relevant to inform expert committee deliberations and finalized this list of end points in consultation with members of the expert committee. Select efficacy end points and notable harms outcomes considered important for informing expert committee deliberations were assessed using GRADE.

Table 9: Outcomes Summarized From the Study Included in the Systematic Review

AE = adverse event; BICR = blinded independent central review; BoR = best objective response; D = durvalumab; DoR = duration of response; EORTC QLQ-C30 = European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30; EORTC QLQ-LC13 = European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Lung Cancer 13; ORR = objective response rate; OS = overall survival; PFS = progression-free survival; RECIST 1.1 = Response Evaluation Criteria in Solid Tumours Version 1.1; SAE = serious adverse event; SOC = standard of care; t = tremelimumab; vs. = versus.

^aThe primary end point was a dual end point of OS and PFS for the durvalumab plus SOC chemotherapy vs. SOC chemotherapy comparison. Because this combination is not aligned with the Health Canada product monograph indication for durvalumab, results for the combination of durvalumab plus SOC chemotherapy are not presented in this review.

^bStatistical testing for these end points was adjusted for multiple comparisons (hierarchal testing).

^cThe prespecified secondary outcome of ORR (based on BICR) using unconfirmed responses was defined as the number (percentage) of patients with at least 1 visit response of complete response or partial response. Post hoc sensitivity analysis of ORR (based on BICR) was conducted using confirmed responses; confirmed ORR was defined as the number (percentage) of patients with at least 1 visit response of complete response or partial response and a confirmatory scan no sooner than 4 weeks after the initial complete response or partial response.

Sources: Sponsor’s Summary of Clinical Evidence,¹ POSEIDON Clinical Study Report,³⁸ POSEIDON Study Protocol,⁴⁷ and the sponsor’s Summary of Clinical Evidence.

Survival Assessments

In the POSEIDON trial, survival assessments were made for patients who completed or discontinued treatment at months 2, 3, and 4, and then every 2 months (8 weeks ± 2 weeks) following the last dose of study treatment. Patients on treatment or in survival follow-up were also contacted to provide survival data following the DCO for the primary and all subsequent survival analyses.⁴⁷

Tumour Assessments

Tumour evaluation was performed at screening (as baseline), at week 6 (± 1 week) and week 12 (± 1 week) from the date of randomization, and then every 8 weeks (± 1 week) until radiological disease progression. RECIST 1.1 assessments were performed on images from CT scans (preferred) or MRI scans.^1,38,47 Patients who discontinued study treatment because of toxicity without objective disease progression continued with scheduled tumour assessments until disease progression or death.^47,48 End points of PFS, ORR, DoR, and BoR were derived using a BICR of all radiologic scans, according to RECIST 1.1.^47,48

Primary End Point

The primary objective of the POSEIDON trial was to assess the efficacy of durvalumab plus SOC chemotherapy compared to SOC chemotherapy alone in terms of dual primary end points of PFS and OS.^1,38,48 The primary end points are not the focus of this review because this objective is outside the scope of the approved Health Canada indication and the sponsor’s reimbursement request.

Key Secondary End Points

Key secondary end points of PFS and OS were assessed for the comparison of durvalumab and tremelimumab plus SOC chemotherapy versus SOC chemotherapy alone.³⁸ These end points were alpha-controlled and included in the hierarchical statistical analysis plan.¹

Overall Survival

OS was defined as the time from randomization to death because of any cause.^1,47 Any patients not known to have died at the time of analysis were censored at the date last known to be alive. If confirmation of a patient being alive or having died occurred after the DCO date, patients were censored at the DCO date.⁴⁷ The sponsor stated that there is no reported MID for OS in the first-line treatment of metastatic NSCLC.¹

Progression-Free Survival

PFS was defined as the time from the date of randomization until the date of objective disease progression or death (because of any cause in the absence of disease progression), regardless of whether the patient withdraws from study treatment or receives another anticancer treatment before progression. PFS was evaluated by BICR according to RECIST 1.1,^1,47 which defines progressive disease as a 20% or greater increase in the sum of diameters of target lesions in addition to an absolute increase of 5 mm or more (with reference to the smallest sum of diameters since initiation of treatment, including the baseline sum if this is the smallest sum on study). Progressive disease is also characterized by fulfillment of progression criteria for nontarget lesions or by the appearance of 1 or more new lesions.^47,48 Patients who did not have disease progression or death at the time of analysis were censored at the time of the latest date of assessment from their last evaluable RECIST 1.1 assessment. Patients who experienced disease progression or death following 2 or more missed visits were censored at the time of the latest evaluable RECIST 1.1 assessment.^47,48 The sponsor stated that there is no reported MID for PFS in the first-line treatment of metastatic NSCLC.¹

Other Secondary End Points

Objective Response Rate

ORR was defined as the number (percentage) of patients with at least 1 visit response of complete response (CR) or partial response (PR). Data obtained up until progression, or the last evaluable assessment in the absence of progression, were included in the assessment of ORR.^1,47 Patients who discontinued study treatment without disease progression, received subsequent (nonradiotherapy) anticancer treatment, and then responded were not included as responders in the ORR.^47,48 The prespecified secondary outcome of ORR (based on BICR) using unconfirmed responses was defined as the number (percentage) of patients with at least 1 visit response of CR or PR.^1,38,47 Post hoc sensitivity analysis of ORR (based on BICR) was conducted using confirmed responses; confirmed ORR was defined as the number (percentage) of patients with at least 1 visit response of CR or PR and a confirmatory scan no sooner than 4 weeks after the initial CR or PR.^1,38 The sponsor stated that there is no reported MID for ORR in the first-line treatment of metastatic NSCLC.¹

Duration of Response

DoR was defined as the time from the date of first documented response until the first date of documented progression or death in the absence of disease progression.^1,47 For patients who did not progress following a response, PFS censoring time was used for their DoR. For patients who did not have a documented response, DoR was not defined.⁴⁷ The sponsor stated that there is no reported MID for DoR in the first-line treatment of metastatic NSCLC.¹

Best Objective Response

BoR was defined as the best response a patient had during the study but before starting any subsequent cancer treatment, and up until RECIST 1.1 progression or the last evaluable assessment in the absence of RECIST 1.1 progression, as determined by BICR. BoR categorization was based on RECIST 1.1 using CR, PR, stable disease, progressive disease, and not-evaluable response categories.^1,48 The sponsor stated that there is no reported MID for BoR in the first-line treatment of metastatic NSCLC.¹

HRQoL End Points

HRQoL was assessed using the EORTC QLQ-C30 and EORTC QLQ-LC13 questionnaires.^1,49,50 PRO assessments were completed by patients during clinic visits using an electronic tablet.⁴⁷

Version 3 of the EORTC QLQ-C30 is a self-administered measure used to assess HRQoL in response to treatment in clinical trials.^1,47,48,51 It consists of 30 questions that can be combined to produce:^1,47,48

• 5 functional scales (physical, role, cognitive, emotional, social)

• 3 symptom scales (fatigue, pain, nausea and vomiting)

• 6 individual items, including 5 symptoms (dyspnea, insomnia, appetite loss, constipation, diarrhea) and 1 item assessing financial difficulties

• 2-item global measure of health status.

For each of the 15 domains, final scores are converted to a standardized score that ranges from 0 to 100 using a linear transformation, with a higher score indicating better functioning on the functional scales, higher (i.e., worse) symptoms on the symptom scales, and better HRQoL on the GHS scale.^1,47-49

The EORTC QLQ-LC13 is a 13-item, lung cancer–specific, self-administered questionnaire. The module consists of both multi-item and single-item measures of lung cancer–associated symptoms (i.e., coughing, hemoptysis, dyspnea, and pain) and side effects from conventional chemotherapy and radiotherapy (i.e., hair loss, neuropathy, sore mouth, and dysphagia).^47,48,50 The scoring approach is similar to that used for the symptom scales and items of the EORTC QLQ-C30.^47,48

In the POSEIDON trial, a clinically meaningful change for the EORTC QLQ-C30 and for the EORTC QLQ-LC13 was defined as an absolute change in the score from baseline of 10 or more points. For each postbaseline assessment, the change from baseline for the scales and items were categorized as improvement, no change, or deterioration.^47,48 This threshold for clinically meaningful change was based on a study by Osoba et al. (1998)⁵² evaluating the significance to patients of changes in EORTC QLQ-C30 scores. Patients who received chemotherapy for either breast cancer or small-cell lung cancer completed a subjective significance questionnaire and the EORTC QLQ-C30, with “a little” change in the subjective significance questionnaire corresponding to changes in the EORTC QLQ-C30 of 5 to 10 points, “moderate” change corresponding with changes of about 10 to 20 points, and “very much” change corresponding with changes of more than 20 points.^1,52

The POSEIDON protocol predefined the following primary PRO symptoms (key symptoms): fatigue, appetite loss (from the EORTC QLQ-C30), coughing, dyspnea, and chest pain (from the EORTC QLQ-LC13). The GHS/QoL and physical functioning domains of the EORTC QLQ-C30 were also prespecified PRO domains of interest.^1,47 PROs were not included in the MTP; however, the overall type 1 (5% [2-sided]) error was controlled across the 5 key symptoms.^1,47,48

Time to deterioration: Time to deterioration was derived for all EORTC QLQ-C30 items except financial difficulties and for all lung cancer symptoms scale items in the EORTC QLQ-LC13. Time to deterioration is defined as the time from the date of randomization until the date of the first clinically meaningful deterioration that is confirmed at a subsequent visit or death (by any cause) in the absence of a clinically meaningful deterioration, regardless of whether the patient withdraws from study treatment or receives another anticancer therapy before HRQoL or symptom deterioration.^1,48

HRQoL/function improvement rate: The function improvement rate was calculated for the EORTC QLQ-C30 and is defined as the number (percentage) of patients with 2 consecutive assessments at least 14 days apart that show a clinically meaningful improvement in that function scale from baseline.^1,48

Symptom improvement rate: Symptom improvement rate was calculated for the EORTC QLQ-LC13 and is defined as the number (percentage) of patients with 2 consecutive assessments at least 14 days apart that show a clinically meaningful improvement in that symptom from baseline.^1,47

Table 10: Summary of Outcome Measures and Their Measurement Properties

ECOG PS = Eastern Cooperative Oncology Group Performance Status; EORTC QLQ-C30 = European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30; EORTC QLQ-LC13 = European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Lung Cancer 13; HRQoL = health-related quality of life; HADS = Hospital Anxiety and Depression Scale; MID = minimal important difference; NSCLC = non–small cell lung cancer; QoL = quality of life; SD = standard deviation; SSQ = subjective significance questionnaire; vs. = versus.

Harms

AEs included any AE with a date of onset on or after the date of the first dose or pretreatment AEs that increased in severity on or after the date of first dose and up to and including the earlier of 90 days following the date of the last dose of study treatment or the date of initiation of the first subsequent anticancer treatment (including radiotherapy, with the exception of palliative radiotherapy) following discontinuation of study treatment (whichever occurred first).^1,48

Coding of AEs was based on the Medical Dictionary for Regulatory Activities dictionary (version 23.1) and grading for severity was according to Common Terminology Criteria for Adverse Events (version 4.03).^1,38,47 AE categories included: all AEs (including those causally related to study treatment), AEs of Common Terminology Criteria for Adverse Events grade 3 or 4 (including those causally related to study treatment), AEs leading to dose modification, AEs leading to hospitalization, AEs with an outcome of death (including those causally related to study treatment), SAEs (including those causally related to study treatment), AEs leading to discontinuation (including those causality related to study treatment), and infusion-related AEs.^1,38

Per the POSEIDON study protocol, for durvalumab with or without tremelimumab, AESIs included events with a potential immune-mediated or inflammatory mechanism and that may require more frequent monitoring and/or interventions (e.g., steroids, immunosuppressants, and/or hormone replacement therapy).^1,47,48

Per the POSEIDON statistical analysis plan, the following were considered AESIs:^1,48

• diarrhea or colitis

• pneumonitis or interstitial lung disease

• increases in alanine transaminase or aspartate transaminase, hepatitis, and hepatotoxicity

• neuropathy or neuromuscular toxicity (e.g., Guillain-Barré, myasthenia gravis)

• endocrinopathies (i.e., events of hypophysitis, hypopituitarism, adrenal insufficiency, diabetes insipidus, hyper- and hypothyroidism, type I diabetes mellitus)

• rash or dermatitis

• nephritis or blood creatinine increases

• pancreatitis (or laboratory results suggestive of pancreatitis)

• other inflammatory responses that are rare with a potential immune-mediated etiology include, but are not limited to, myocarditis, pericarditis, uveitis, vasculitis, noninfectious meningitis, and noninfectious encephalitis

• infusion-related reactions and hypersensitivity or anaphylactic reactions with a different underlying pharmacological etiology.

While the POSEIDON study was being conducted, the process for identifying imAEs was redefined, which resulted in revision to the criteria for defining AESIs and AEPIs. The definitions for AESIs, AEPIs, and imAEs in the study were:³⁸

• AESIs: AEs with a likely inflammatory or immune-mediated pathophysiological basis resulting from the mechanism of action of durvalumab and requiring more frequent monitoring and/or interventions (e.g., corticosteroids, immunosuppressants, endocrine therapy)

• AEPIs: AEs that could have a potential inflammatory or immune-mediated pathophysiological basis resulting from the mechanism of action of durvalumab but are more likely to have occurred because of other pathophysiological mechanisms; the likelihood of the event being inflammatory or immune-mediated in nature is not high and/or is most often or usually explained by the other causes.

Statistical Analysis

A summary of the statistical analyses relevant to the efficacy end points of interest from the POSEIDON trial is presented in the following section and in Table 11.

Sample Size and Power Calculation

The POSEIDON trial planned to enrol approximately 2,000 patients and randomize approximately 1,000 patients in a 1:1:1 ratio (approximately 333 patients in each treatment arm) to durvalumab and tremelimumab plus SOC chemotherapy, durvalumab plus SOC chemotherapy, or SOC chemotherapy alone, including at least 250 patients in each treatment arm with PD-L1 TC level of less than 50%. Participants were randomized in a stratified manner according to PD-L1 tumour-expression status (PD-L1 TC ≥ 50% versus < 50%), disease stage (IVA versus IVB), and histology (nonsquamous versus squamous). The study was powered for dual primary end points to characterize PFS and OS effects for the comparison of durvalumab plus SOC chemotherapy versus SOC chemotherapy alone in the ITT population.^1,38,48 A statistically significant benefit in either PFS or OS for this comparison had to be demonstrated for the study to be declared positive.¹

One interim analysis and 1 final analysis of the primary and key secondary PFS end points were performed and 3 interim analyses and 1 final analysis were planned to be performed for the primary and key secondary OS end points.^1,48

The final PFS (primary end point) analysis for superiority was prespecified to be performed when approximately 497 BICR PFS events from the global cohort had occurred across the durvalumab plus SOC chemotherapy and SOC chemotherapy alone arms (75% maturity). The final OS (primary end point) analysis for superiority was prespecified to be performed when approximately 532 OS events had occurred across the durvalumab plus SOC chemotherapy and SOC chemotherapy alone arms (80% maturity).^1,38,48

Key Secondary End Points

Durvalumab and Tremelimumab Plus SOC Chemotherapy Versus SOC Chemotherapy (PFS in ITT Population)

Assuming a true PFS HR of 0.51 and median PFS of 6 months in the SOC chemotherapy arm, 465 PFS events from the global cohort (70% maturity) would provide greater than 90% power to demonstrate statistical significance at the 2-sided alpha level of 0.9% (with an overall alpha for PFS of 1%), allowing for 1 interim analysis conducted at approximately 80% of the target events (information fraction). The smallest statistically significant treatment difference would be an HR of 0.78. This analysis was anticipated to be 25 months from first patient in (assuming a recruitment period of 16 months).^1,38,48

Durvalumab and Tremelimumab Plus SOC Chemotherapy Versus SOC Chemotherapy (OS in ITT Population)

Assuming a true OS HR of 0.7 and median OS of 12.9 months in the SOC chemotherapy arm, 532 OS events (80% maturity) would provide greater than 90% power to demonstrate statistical significance at the 2-sided alpha level of 3.3% (with an overall alpha for OS of 4%), allowing for 3 interim analyses conducted at approximately 45%, 61%, and 84% of the target events (information fraction). The smallest statistically significant treatment difference would be an HR of 0.83. This analysis was anticipated to be 46 months from first patient in (assuming a recruitment period of 16 months).^1,38,48

Statistical Test or Model

Statistical analysis of efficacy end points for the POSEIDON trial are presented in Table 11. Efficacy analyses were based on the ITT population (full analysis set [FAS]), and safety analyses were based on the safety analysis set (SAF) (Table 12).

Multiple Testing Procedure

To control the type I error at 5% (2-sided), an MTP with a gatekeeping strategy was used across the following end points included in the MTP:^1,38

• Dual primary end points: PFS (using BICR assessments according to RECIST 1.1) and OS for the durvalumab plus SOC chemotherapy versus SOC chemotherapy comparison in the ITT population

• Key secondary end points: PFS (using BICR assessments according to RECIST 1.1) and OS for the durvalumab and tremelimumab plus SOC chemotherapy versus SOC chemotherapy comparison in the ITT population

• Secondary end point: OS for the durvalumab and tremelimumab plus SOC chemotherapy versus SOC chemotherapy comparison in the populations of patients of with a blood tumour mutational burden of 20 mutations per megabase or greater (bTMB20 high), 16 mutations per megabase or greater (bTMB16 high), and 12 mutations per megabase or greater (bTMB12 high).

Hypotheses were tested using an MTP with an alpha-exhaustive recycling strategy and were tested in a predefined order (Figure 3).⁴⁸ According to alpha (test mass) splitting and alpha recycling, if the higher-level hypothesis in the MTP was rejected for superiority, the next lower-level hypothesis was then tested. The test mass that became available after each rejected hypothesis was recycled to the lower-level hypotheses not yet rejected. This testing procedure stopped when the entire test mass was allocated to nonrejected hypotheses.^1,38,48

The alpha level allocated to the interim or final analyses was determined by the Lan-DeMets spending function that approximates an O’Brien-Fleming approach. The alpha level applied at the interims was set to the proportion of information available; that is the actual observed number of PFS or OS events and percent of maturity at the time of interim analyses. A separate Lan-DeMets (O’Brien-Fleming) spending function accounting for interim and final analysis was applied for OS and PFS end points.^1,38,48

In the MTP, the overall 5% type I error was first split between the dual primary end points of PFS and OS; an alpha level of 1% was allocated to the primary PFS analysis and an alpha of 4% was allocated to the primary OS analysis (durvalumab plus SOC chemotherapy versus SOC chemotherapy) (Figure 3). The MTP involved the following strategy:^1,38,48

• If the primary PFS analysis was significant, the 1% alpha level could be recycled to the key secondary PFS comparison (durvalumab and tremelimumab plus SOC chemotherapy versus SOC chemotherapy).

• If the primary OS analysis was significant, the 4% alpha level could be recycled to the key secondary OS comparison (durvalumab and tremelimumab plus SOC chemotherapy versus SOC chemotherapy).

• If the key secondary PFS comparison between durvalumab and tremelimumab plus SOC chemotherapy versus SOC chemotherapy was significant, the 1% alpha level could be recycled to the OS comparison between durvalumab and tremelimumab plus SOC chemotherapy versus SOC chemotherapy.

• If the key secondary OS comparison between durvalumab and tremelimumab plus SOC chemotherapy versus SOC chemotherapy was significant, the 4% alpha level could be recycled to the key secondary PFS comparison between durvalumab and tremelimumab plus SOC chemotherapy versus SOC chemotherapy.

• If both key secondary PFS and OS comparisons of durvalumab and tremelimumab plus SOC chemotherapy versus SOC chemotherapy were significant, the available alpha level could be recycled to the OS comparison between durvalumab and tremelimumab plus SOC chemotherapy versus SOC chemotherapy alone in the bTMB20 high population. If the comparison was significant, the alpha level could be recycled to the OS comparison between durvalumab and tremelimumab plus SOC chemotherapy versus SOC chemotherapy in the bTMB16 high population. If the comparison was significant, the alpha level could be recycled to the OS comparison between durvalumab and tremelimumab plus SOC chemotherapy versus SOC chemotherapy in the bTMB12 high population.

Figure 3: Multiple Testing Procedures for Controlling the Type I Error Rate

bTMB12 = blood tumour mutational burden of 12 mutations per megabase or greater; bTMB16 = blood tumour mutational burden of 16 mutations per megabase or greater; bTMB20 = blood tumour mutational burden of 20 mutations per megabase or greater; combo = durvalumab and tremelimumab plus chemotherapy; ITT = intention to treat; mono = durvalumab plus chemotherapy; OS = overall survival; PFS = progression-free survival; SoC = standard of care chemotherapy.

Sources: Sponsor’s Summary of Clinical Evidence¹ and POSEIDON Clinical Study Report.³⁸

One interim analysis of PFS was planned for when approximately 80% of the target PFS events (approximately 497) had occurred across the durvalumab plus SOC chemotherapy and SOC chemotherapy arms (397 out of 497 events).^1,38,48

Three interim analyses of OS were planned:^1,38,48

• the first at the time of the interim PFS analysis (approximately 45% of the target OS events [approximately 532] in the durvalumab plus SOC chemotherapy and SOC chemotherapy arms [243 out of 532 events])

• the second at the time of the final PFS analysis (approximately 61% of the target OS events in the durvalumab plus SOC chemotherapy and SOC chemotherapy arms [328 out of 532 events])

• the third when approximately 84% of the target OS events had occurred in the durvalumab plus SOC chemotherapy and SOC chemotherapy arms (449 out of 532 events).

The interim analyses were performed for the analyses specified in MTP. Global recruitment (excluding China) was completed before the results of the interim analyses becoming available (data presented in this review report are for the global cohort excluding China).^1,38

The DCO for the final analysis of the POSEIDON trial occurred on July 24, 2019, for PFS and all RECIST 1.1–related end points, and on March 12, 2021, for OS, safety, and all other data.^1,38

A prespecified long-term survival follow-up analysis of the POSEIDON study was scheduled to occur at least 3 years after the last patient was randomized and has been completed. This is referred to as the 4-year OS analysis and the DCO date of this analysis was March 11, 2022.¹

A subsequent long-term survival follow-up analysis of the POSEIDON trial was scheduled to occur at least 4 years after the last patient was randomized and has been completed. This is referred to as the 5-year OS analysis and the DCO date of this analysis was August 24, 2023.¹

Data from the final analysis and 5-year OS analysis (but not the 4-year OS analysis) are presented in this Clinical Review.

The 5-year OS analysis included all 3 treatment arms and was limited to an updated summary of OS (including sensitivity and subgroup analyses) and a summary of key safety data occurring since the final analysis DCO date (as well as other updated data). All safety and efficacy end points were evaluated using the same statistical methodology as previously defined for the final analysis.⁵⁸ The updated OS estimates and CIs were not alpha-controlled.¹

PRO symptoms and HRQoL end points were not part of the MTP. However, the overall type I error (5% [2-sided]) was controlled across the 5 predefined primary PRO measures (key symptoms) of fatigue and appetite loss (from EORTC QLQ-C30) and coughing, dyspnea, and pain in chest (from EORTC QLQ-LC13) using the Bonferroni-Holm procedure. The EORTC QLQ-C30 overall health status and physical functioning domains were prespecified end points of interest.^1,48

Data Imputation Methods

Missing efficacy and safety data were generally not imputed.⁴⁸

Subgroup Analyses

The following prespecified subgroups analyses were performed comparing OS and PFS between treatments of the primary and key secondary end points to evaluate the consistency of treatment effect across expected prognostic and/or predictive factors:^1,48

• sex (male versus female)

• age at randomization (< 65 years versus ≥ 65 years)

• PD-L1 status (PD-L1 TC ≥ 50% versus TC < 50%)

• histology (squamous versus nonsquamous)

• chemotherapy (nab-paclitaxel doublet versus pemetrexed doublet versus gemcitabine doublet)

• smoking (current smoker, former smoker, or never smoker)

• race (Asian versus non-Asian)

• PD-L1 using cut points of 1% and 25% tumour expression (< 1% versus ≥ 1%, and < 25% versus ≥ 25%)

• performance status (normal activity [0] versus restricted activity [1])

• brain metastases (yes versus no)

• disease stage (IVA versus IVB).

The study was not powered for any of the individual subgroup evaluations and no adjustments were made for multiple testing subgroup analyses.^1,38,48

The HR and 95% CI was calculated from an unstratified Cox proportional hazards model with treatment as the only covariate for each subgroup level of a factor.⁴⁸

Interactions between treatment and stratification factors were to be tested to rule out any qualitative interaction using the Gail and Simon approach.^48,59 This test was to be performed separately for the treatment comparisons of the dual primary and key secondary end points based on the FAS.⁴⁸

Sensitivity Analyses

The POSEIDON study conducted sensitivity analyses for the key secondary end points of OS and PFS in the comparison of durvalumab and tremelimumab plus SOC chemotherapy versus SOC chemotherapy alone, adjusting for further prespecified covariates of age at randomization, sex, smoking history, and race (Table 11). Additional sensitivity analyses were conducted for OS to rule out attrition bias, for PFS to assess the robustness of the PFS effect to potential sources of bias, and for OS (post hoc) using confirmed responses (Table 11).

Table 11: Statistical Analysis of Efficacy End Points for POSEIDON Study

BICR = blinded independent central review; CI = confidence interval; D = durvalumab; EORTC QLQ-C30 = European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30; EORTC QLQ-LC13 = European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Lung Cancer 13; HR = hazard ratio; HRQoL = health-related quality of life; ITT = intention to treat; NA = not available; OR = odds ratio; OS = overall survival; PFS = progression-free survival; PRO = patient-reported outcomes; RECIST 1.1 = Response Evaluation Criteria in Solid Tumours Version 1.1; t = tremelimumab; TC = tumour cell; vs. = versus.

Sources: Sponsor’s Summary of Clinical Evidence,¹ POSEIDON Clinical Study Report,³⁸ and POSEIDON Statistical Analysis Plan.⁴⁸

Analysis Populations

Analysis populations relevant to the outcomes reported in this Clinical Review are presented in Table 12.

Table 12: Analysis Populations of the POSEIDON Trial

AE = adverse event; BoR = best objective response; D = durvalumab; DoR = duration of response; FAS = full analysis set; GHS = Global Health Status; HRQoL = health-related quality of life; ITT = intention to treat; ORR = objective response rate; OS = overall survival; PFS = progression-free survival; PRO = patient-reported outcome; SAF = safety analysis set; t = tremelimumab.

^aPatients who are evaluable for the analysis of ORR are those with measurable disease at baseline.

^bPatients who are evaluable for the analysis of DoR are those who responded in the ORR analysis.

^cThe population for analysis of time to GHS/HRQoL or function deterioration will include a subset of the FAS population who have baseline scores of 10 or higher; the population for analysis of time to symptom deterioration will include a subset of the FAS population who have baseline scores of 90 or lower. For symptom improvement rate, the denominator will consist of a subset of the FAS who have a baseline symptom score of 10 or higher. For HRQoL/function improvement rate, the denominator will consist of a subset of the FAS who have a baseline HRQoL/function score of 90 or lower.

Sources: Sponsor’s Summary of Clinical Evidence,¹ POSEIDON Clinical Study Report,³⁸ and POSEIDON Statistical Analysis Plan.⁴⁸

Outcomes of the MTP

The outcomes of the MTP at the final analysis (DCO: July 24, 2019, for PFS; March 12, 2021, for OS) are presented Figure 4. At level 1, the dual primary end point of PFS and OS for the comparison of durvalumab plus SOC chemotherapy versus SOC chemotherapy alone was tested and at level 2, the key secondary end points of PFS and OS for the comparison of durvalumab and tremelimumab plus SOC chemotherapy versus SOC chemotherapy alone were tested. At level 1, the HR for PFS favoured durvalumab plus SOC chemotherapy and was statistically significant at the 1% alpha level; the 1% alpha was therefore recycled to test PFS for the durvalumab and tremelimumab plus SOC chemotherapy versus SOC chemotherapy alone comparison at level 2. The HR for PFS favoured durvalumab and tremelimumab plus SOC chemotherapy and was statistically significant at the 1% alpha level. At level 1, the HR for OS numerically favoured durvalumab plus chemotherapy but did not cross the prespecified statistical significance threshold at the 4% alpha level. For this reason, at level 2, the 1% alpha level from the PFS for durvalumab and tremelimumab plus SOC chemotherapy versus SOC chemotherapy alone comparison was recycled to test OS for this same comparison. The HR for OS favoured durvalumab and tremelimumab plus SOC chemotherapy and was statistically significant at the 1% alpha level. At level 3, OS in the bTMB20 population for the comparison between durvalumab and tremelimumab plus SOC chemotherapy versus SOC chemotherapy alone was tested at the recycled 1% alpha level. The HR for OS favoured durvalumab and tremelimumab plus SOC chemotherapy but did not cross the prespecified statistical significance threshold; bTMB16 and bTMB12 populations were therefore not tested.^1,38

Efficacy results for the comparison between durvalumab and tremelimumab plus SOC chemotherapy versus SOC chemotherapy alone are presented in this Clinical Review, because this treatment arm aligns with the indication approved by Health Canada. Results for the dual primary end points (OS and PFS) for the comparison between durvalumab and chemotherapy versus SOC chemotherapy alone are presented in Appendix 1.

Figure 4: Actual Multiple Testing Procedure at Final Analysis in the POSEIDON Trial

bTMB = blood tumour mutational burden; ITT = intention to treat; mut/Mb = mutations per megabase; OS = overall survival; PFS = progression-free survival.

Source: Sponsor’s Summary of Clinical Evidence.¹

Results

Patient Disposition

Patient disposition in the POSEIDON trial is summarized in Table 13.

A total of 1,807 patients were enrolled into the study, of whom 1,013 were randomized to 1 of the 3 study arms (arm 1: N = 338; arm 2: N = 338; arm 3: N = 337).

In the FAS for treatment arm 1 and arm 3, 7 patients (2.1%) and 6 patients (1.8%), respectively, were randomized but did not receive treatment.³⁸ Of the patients who received treatment in arm 1 (n = 331), 294 (88.8%) had discontinued durvalumab treatment at the final analysis DCO (March 12, 2021), and 311 (94.0%) had discontinued durvalumab treatment at the 5-year OS analysis DCO (August 24, 2023).

The most common reason for discontinuation of durvalumab treatment was the patient’s “condition under investigation worsened” (216 [65.3%] and 219 [66.2%] at the final analysis and 5-year OS analysis DCO dates, respectively). Of the patients who received treatment in arm 1, at both DCO dates, 117 patients (35.3%) had discontinued tremelimumab treatment and 213 patients (64.4%) had completed tremelimumab treatment. The most common reason for discontinuation of tremelimumab was that the patient’s “condition under investigation worsened” (69 [20.8%] for both DCO dates).

Of the patients who received treatment in arm 1, 311 (94.0%) had discontinued chemotherapy at the final analysis DCO and 320 (96.7%) had discontinued chemotherapy at the 5-year OS analysis DCO. The most common reason for discontinuation of chemotherapy in arm 1 was that the patient’s “condition under investigation worsened” (138 [41.7%] at both DCO dates); 95 patients (28.7%) at both DCO dates discontinued because the maximum number of chemotherapy cycles had been reached.

Of the patients who received treatment in arm 3 (n = 331), 326 (98.5%) had discontinued chemotherapy at the final analysis DCO and 330 (99.7%) had discontinued chemotherapy at the 5-year OS analysis DCO. The most common reason for discontinuation of chemotherapy was also that the patient’s “condition under investigation worsened” (183 [55.3%] and 185 [55.9%] at the final analysis and 5-year OS analysis DCO dates, respectively); 73 patients (22.1%) at the final analysis and 74 patients (22.4%) at the 5-year analysis discontinued because the maximum number of chemotherapy cycles had been reached. In the full analysis set (FAS), a higher proportion of patients in arm 3 had terminated the study at the final analysis and 5-year OS analysis DCOs (88.1% and 94.4%, respectively) than in arm 1 (76.3% and 85.2%, respectively), with death the most common reason for study termination in both groups.

Table 13: Patient Disposition in the POSEIDON Trial

5Y = 5-year; CDA-AMC = Canada’s Drug Agency’ D = durvalumab; DCO = data cut-off; eCRF = electronic case report form; FA = final analysis; NA = not available; OS = overall survival; SOC = standard of care; t = tremelimumab.

Notes: Results from the durvalumab plus SOC chemotherapy arm are not relevant to the systematic review because this regimen is not part of the Health Canada–approved indication and reimbursement request under review by CDA-AMC.

For patients who underwent re-treatment with T, definitions were as follows:⁶⁰ If a patient discontinues T re-treatment for the reason “maximum cycle of immunotherapy reached,” they are classed as “completed.” If a patient discontinues T re-treatment for any reason other than “maximum cycle of immunotherapy reached,” they are classed as “discontinued.” If a patient started T re-treatment, but has either not discontinued T re-treatment or has not filled out the DOSDISC8 form, they are classed as “ongoing.” DCO dates: March 12, 2021 (final analysis), and August 24, 2023 (5-year OS analysis).

^aThe median duration of follow-up in censored patients was 63.44 months (range, 0 to 73.9 months) at the 5-year DCO date.

^bPercentages are calculated from number of patients enrolled.

^cPercentages are calculated from number of patients in the FAS.

^dPercentages are calculated from number of patients who received treatment.

^ePatients who received tremelimumab re-treatment were not included.

^fPatients who completed tremelimumab have “maximum cycle of immunotherapy reached” reported for tremelimumab on the eCRF. Patients who completed SOC chemotherapy have “maximum cycle of chemotherapy reached” reported for any SOC chemotherapy molecule on the eCRF.

^gA patient is considered as having discontinued SOC chemotherapy when all molecules are discontinued. If different reasons for discontinuation are collected, the last discontinuation reason by date is selected.

^hA patient is considered to be ongoing treatment when receiving at least 1 study treatment at DCO. Treatment includes D, T, CT, and T re-treatment where applicable.

ⁱA patient is considered as having discontinued treatment when the patient has discontinued all study treatments. Treatment includes D, T, CT, and T re-treatment where applicable.

^jRepresents patients who were alive and continuing in the study at the DCO date of the FA.

^kPatients who completed the study includes patients who are alive and remain in the study at 5-year DCO. One patient was marked as completed (alive) at OS final analysis (March 12, 2021), and updated information is not available. One patient was marked as completed (alive) at the 4-year OS analysis (March 11, 2022).

Sources: Sponsor’s Summary of Clinical Evidence,¹ POSEIDON Clinical Study Report,³⁸ and POSEIDON Clinical Study Report 5-year OS Addendum.⁵⁸

Baseline Characteristics

Demographic and baseline tumour characteristics of patients in the POSEIDON trial (Table 14) were largely similar between arm 1 and arm 3 of the POSEIDON trial. Across all 3 treatment arms (FAS), the median age of patients was 64 years (range, 27 to 87), with approximately 47% of patients aged 65 years or older and the majority of patients being male (76.0%) and white (55.9%). The durvalumab and tremelimumab plus SOC chemotherapy group (arm 1) had a lower percentage of female patients compared to the SOC chemotherapy group (arm 3) (20.4% versus 26.4%) as well as a lower percentage of Asian patients compared to the chemotherapy group (29.3% versus 38.0%). Overall, most patients were either current or former smokers (78%) and 21.9% of patients reported having never smoked, with a lower proportion of patients in the durvalumab and tremelimumab plus SOC chemotherapy arm being never smokers compared with the SOC chemotherapy arm (17.5% versus 23.4%). Overall, the majority of patients had an ECOG PS of 1 (66.5%), nonsquamous histology (62.9%), a PD-L1 TC level of less than 50% (71.1%), and 10.5% of patients had brain and/or CNS metastases at study entry. Between arm 1 and arm 3, histology type, PD-L1 status, and receipt of prior anticancer therapy (cytotoxic chemotherapy or radiotherapy) were generally well balanced.^1,38

Table 14: Summary of Baseline Characteristics of the POSEIDON Trial (Full Analysis Set)

AJCC = American Joint Committee on Cancer; BMI = body mass index; CDA-AMC = Canada’s Drug Agency; D = durvalumab; ECOG = Eastern Cooperative Oncology Group; eCRF = electronic case report form; IASLC = The International Association for the Study of Lung Cancer; SD = standard deviation; t = tremelimumab; TC = tumour cell.

Notes: Results from the durvalumab plus SOC chemotherapy arm are not relevant to the systematic review because this regimen is not part of the Health Canada–approved indication and reimbursement request under review by CDA-AMC.

Percentages are calculated from number of patients in the full analysis set in that treatment group.

^aAge at randomization.

^bECOG performance status at baseline, where baseline is defined as the last evaluable assessment before randomization.

^cThe IASLC is the group responsible for conducting research to inform future versions of the AJCC. Version 8 of the IASLC Staging Manual in Thoracic Oncology was endorsed and published by the AJCC as the AJCC 8th edition.^15,43 The 2 are considered interchangeable.^45,46

^dMetastatic disease: patient has any metastatic site of disease.

^eLocally advanced: patient has only locally advanced sites of disease. Two patients in the D + T + CT arm and 1 in the durvalumab plus chemotherapy arm were incorrectly randomly assigned with stage III disease; these were reported as protocol deviations.

^fStratification factor recorded on an eCRF. PD-L1 tumour-expression status is summarized based on laboratory data outside of the eCRF.

Sources: Sponsor’s Summary of Clinical Evidence¹ and POSEIDON Clinical Study Report.³⁸

Exposure to Study Treatments

Patient exposure to study treatments in arm 1 and arm 3 of the POSEIDON trial is summarized in Table 15. In the SAF, the mean and median duration of exposure to durvalumab and tremelimumab plus SOC chemotherapy was longer than that to SOC chemotherapy for both DCO dates. For the final analysis DCO (March 12, 2021), the median extents of exposure were 29.93 weeks (range, 1.10 to 189.6) and 18.00 weeks (range, 0.7 to 184.4) in arms 1 and 3, respectively. For the same DCO, the mean extents of exposure were 49.62 weeks (standard deviation [SD] = 48.151) and 25.83 weeks (SD = 29.004) in arms 1 and 3, respectively. For the 5-year OS analysis DCO (August 24, 2023), the median extents of exposure were 29.93 weeks (range, 1.1 to 317.4) and 18.00 weeks (range, 0.7 to 284.1) in arms 1 and 3, respectively. For the 5-year DCO, the mean extent of exposure was 60.37 weeks (SD = 75.397) and 27.26 weeks (SD = 36.968) in arms 1 and 3, respectively. At the final analysis, the total durations of treatment across all patients were 313.8 years in the durvalumab and tremelimumab plus SOC chemotherapy arm and 164.9 years in the SOC chemotherapy alone arm. At the 5-year OS update analysis the total duration of treatment across all patients had increased to 381.8 years in the durvalumab and tremelimumab plus SOC chemotherapy arm and 174.0 years in the SOC chemotherapy alone arm.

Patient exposure for durvalumab, tremelimumab, and SOC chemotherapy treatments in the durvalumab and tremelimumab plus SOC chemotherapy arm (arm 1, SAF) is summarized in Table 16. The median number of infusions and median duration of exposure to durvalumab and to SOC chemotherapy was the same for both the final analysis and 5-year OS DCOs. With the longer duration of follow-up, the maximum number of infusions and the maximum duration of exposure (in weeks) increased from 49 to 80, and 189.6 to 317.4, respectively for both durvalumab and SOC chemotherapy between the 2 DCOs. In arm 1 at the 5-year OS DCO, there are an additional 65.5 patient-years of exposure to durvalumab and an additional 35 patient-years for chemotherapy (with longer exposure to chemotherapy because of patients who received pemetrexed maintenance therapy). There was no change in the total number of infusions of tremelimumab or duration of exposure in arm 1 between the 2 DCOs.

Patient exposure to chemotherapy in the SOC chemotherapy arm (arm 3, SAF) is summarized in Table 17. The median numbers of infusions and median duration of exposure to chemotherapy were the same for both the final analysis and 5-year OS DCOs. The maximum number of chemotherapy infusions and the maximum duration of exposure (in weeks) increased from 57 to 94, and 184.4 to 284.1, respectively between the 2 DCOs. In arm 3 at the 5-year OS DCO, there are 9.1 additional patient-years of exposure to SOC chemotherapy because of patients who received pemetrexed maintenance therapy.

Among patients who received SOC chemotherapy, patients with nonsquamous histology most commonly received the pemetrexed-platinum regimen (598 of 626 [95.5%]) and patients with squamous histology most commonly received the gemcitabine platinum regimen (326 of 369 [88.3%]).³⁹

Table 15: Summary of Patient Exposure From the POSEIDON Trial (Safety Analysis Set)

D = durvalumab; DCO = data cut-off; OS = overall survival; SD = standard deviation; t = tremelimumab.

Note: Percentages are calculated from number of patients in the safety analysis set in that treatment group. Chemotherapy for a patient who received it during re-treatment is also included in this table. DCO dates: March 12, 2021 (final analysis), and August 24, 2023 (5-year OS analysis).

^aTotal treatment duration = minimum of (last infusion or dose date of the last cycle + 20 days [if last infusion or dose date was during chemotherapy] or last infusion or dose date of the last cycle + 27 days [if last infusion or dose date was postchemotherapy], date of death, date of DCO) – first infusion or dose date of first cycle + 1.

Sources: Sponsor’s Summary of Clinical Evidence,¹ POSEIDON Clinical Study Report,³⁸ and POSEIDON Clinical Study Report 5-year OS Addendum.⁵⁸

Table 16: Patient Exposure in the Durvalumab and Tremelimumab Plus Chemotherapy Arm (N = 330) of the POSEIDON Trial (Safety Analysis Set)

DCO = data cut-off; OS = overall survival; SD = standard deviation.

Note: DCO dates: March 12, 2021 (final analysis), and August 24, 2023 (5-year OS analysis).

^aThe median duration of follow-up in censored patients was 63.44 months (range, 0 to 73.9 months) at the 5-year data cut-off date.

^bTotal treatment duration = minimum of (last infusion or dose date of the last cycle + 20 days [if last infusion or dose date was during combination] or last infusion or dose date of the last cycle + 27 days [if last infusion or dose date was in maintenance or tremelimumab re-treatment], date of death, date of DCO) – first infusion or dose date of first cycle + 1.

Sources: Sponsor’s Summary of Clinical Evidence,¹ POSEIDON Clinical Study Report,³⁸ and POSEIDON Clinical Study Report 5-year OS Addendum.⁵⁸

Table 17: Patient Exposure in the Chemotherapy Arm (N = 333) of the POSEIDON Trial (SAF)

DCO = data cut-off OS = overall survival; SAF = safety analysis set; SD = standard deviation.

Note: DCO dates: March 12, 2021 (final analysis), and August 24, 2023 (5-year OS analysis).

^aThe median duration of follow-up in censored patients was 63.44 months (range, 0 to 73.9 months) at the 5-year data cut-off date.

^bTotal treatment duration = minimum of (last infusion or dose date of the last cycle + 20 days [if last infusion or dose date was during combination] or last infusion or dose date of the last cycle + 27 days [if last infusion or dose date was in maintenance], date of death, date of DCO) – first infusion or dose date of first cycle + 1.

Sources: Sponsor’s Summary of Clinical Evidence,¹ POSEIDON Clinical Study Report,³⁸ and POSEIDON Clinical Study Report 5-Year OS Addendum.⁵⁸

Durvalumab Treatment Through Progression

Patients in arm 1 (durvalumab and tremelimumab plus SOC chemotherapy) who were clinically stable at the initial RECIST 1.1–defined progressive disease assessment continued to receive durvalumab monotherapy, at the discretion of the investigator and patient, until confirmation of progression at the next subsequent scan (no sooner than 4 weeks and preferably at the next scheduled imaging visit). Patients who were considered by the investigator to be still benefiting from durvalumab monotherapy following confirmed progression could continue receiving durvalumab monotherapy as long as criteria for treatment in the setting of objective radiological progression were fulfilled.

The durvalumab and tremelimumab plus SOC chemotherapy arm included 132 patients (40.0%) with a RECIST 1.1–defined disease progression followed by a subsequent scan confirming RECIST 1.1–defined disease progression. In the durvalumab and tremelimumab plus SOC chemotherapy arm, 36 patients (10.9%) received treatment with durvalumab following the date of disease progression confirmation and 18 patients (5.5%) continued to receive durvalumab 8 weeks (56 days) after progression confirmation, with a median duration of treatment postprogression confirmation of 8.4 weeks. Data for patients receiving durvalumab after the date of disease progression confirmation in the durvalumab and tremelimumab plus SOC chemotherapy arm are presented in Table 18.⁶⁰

Tremelimumab Re-Treatment

Patients in the durvalumab and tremelimumab plus SOC chemotherapy arm with radiological progression who, in the investigator’s opinion, continued to receive benefit from their assigned treatment and who met the criteria for re-treatment in the setting of disease progression, could be re-treated with 4 doses per cycle of tremelimumab alongside durvalumab followed by continued durvalumab monotherapy.^1,38,47

At the final analysis, 11 patients (3.3%) in the durvalumab and tremelimumab plus SOC chemotherapy arm received 6 or more cycles of tremelimumab (i.e., patients receiving tremelimumab re-treatment).³⁸ Table 13 reports a total of 10 patients as having received tremelimumab re-treatment (presented in the rows of ongoing, completed, and discontinued tremelimumab re-treatment) as the data in this table align with patient disposition data from the 5-year follow-up Clinical Study Report. The difference in the 1 additional patient being reported was because of incorrect data entry.⁶⁰

Table 18: Summary of Patients Who Received Durvalumab Through Confirmed Progression (Based on Site-Investigator Assessments According to RECIST 1.1) (Safety Analysis Set)

RECIST 1.1 = Response Evaluation Criteria in Solid Tumours Version 1.1; SOC = standard of care.

Note: The date of progression confirmation is defined as the date of the second progression assessment of the first case of 2 consecutive assessments of progression. Duration of treatment postprogression is defined as the time (weeks) from progression confirmation until the date of last durvalumab dose + 1 day. There were 6 patients in the durvalumab and tremelimumab plus SOC chemotherapy arm (N = 330) in whom the durvalumab component was not the last to be discontinued. Data cut-off: March 12, 2021.

^aProgression based on site-investigator assessments according to RECIST 1.1.

Source: Sponsor’s data on file.⁶⁰

Concomitant Medications and Cointerventions

A summary of allowed concomitant medications that were taken during study treatment (by ≥ 20% of patients in any treatment arm) in the POSEIDON trial is presented in Table 19. At the final analysis DCO (March 12, 2021), almost all patients in the durvalumab and tremelimumab plus SOC chemotherapy and SOC chemotherapy alone arms ██████ ███ █████, respectively) received allowed concomitant medications during study treatment. The most commonly received concomitant medications were serotonin (5HT₃) antagonists and glucocorticoids. Concomitant medications with a 5% or greater difference between the durvalumab and tremelimumab plus SOC chemotherapy and SOC chemotherapy alone arms, respectively, were serotonin antagonists (█████ ██ █████); anilides ██████ ██ ██████; combinations of penicillin, including beta-lactamase inhibitors ██████ ██ ██████; and colony-stimulating factors ██████ ██ ██████.³⁸

Table 19: Allowed Concomitant Medications During Study Treatment Received by 20% or More of Patients in Any Treatment Arm (Full Analysis Set)

ATC = Anatomic Therapeutic Chemical; D = durvalumab; H2 = histamine H2; t = tremelimumab.

Note: A patient can have 1 or more generic terms reported under a given ATC classification. Includes medications that began before randomization and were ongoing after randomization, or medications that started on or after date of randomization and before the date of last dose of study drug plus 90 days. WHO Drug Dictionary version 202009. Percentages are calculated from number of patients in the full analysis set in that treatment arm. Data cut-off dates: March 12, 2021 (final analysis).

Sources: Sponsor’s Summary of Clinical Evidence¹ and POSEIDON Clinical Study Report.³⁸

Subsequent Treatment

A summary of subsequent anticancer treatments received by patients in the POSEIDON trial following discontinuation of study treatment is presented in Table 20. At both the final analysis and 5-year OS analysis DCO dates, a smaller proportion of patients in the durvalumab and tremelimumab plus SOC chemotherapy group (arm 1) received subsequent anticancer therapy compared to patients in the SOC chemotherapy group (arm 3). At the final analysis DCO date (March 12, 2021), 40.8% and 60.2% of patients in arms 1 and 3, respectively, had received subsequent anticancer treatment; at the 5-year OS analysis (August 24, 2023), these percentages were 42.9% and 60.8%, respectively. Systemic therapy was used more commonly than radiotherapy in both treatment arms. At the 5-year OS DCO (August 24, 2023), 37.9% of patients in the durvalumab and tremelimumab plus SOC chemotherapy arm and 57.9% of patients in the SOC chemotherapy alone arm had received subsequent systemic therapy. At the same DCO, generally similar proportions of patients in arms 1 and 3 (33.4% and 36.2%, respectively) had received cytotoxic chemotherapy; however, a smaller proportion of patients in the durvalumab and tremelimumab plus SOC chemotherapy arm had received subsequent immunotherapy compared to those in the SOC chemotherapy alone arm (7.4% versus 33.2%).

Table 20: Subsequent Anticancer Treatments in the POSEIDON Trial (Full Analysis Set)

CDA-AMC = Canada’s Drug Agency; D = durvalumab; DCO = data cut-off; IO = immuno-oncology; OS = overall survival; SOC = standard of care; t = tremelimumab.

Note: Results from the durvalumab plus SOC chemotherapy arm are not relevant to the systematic review because this regimen is not part of the Health Canada–approved indication and reimbursement request under review by CDA-AMC.

Patients with therapies in more than 1 category are counted once in each of those categories. Percentages are calculated from the number of patients in the FAS in that treatment arm. DCO dates: March 12, 2021 (final analysis), and August 24, 2023 (5-year OS analysis).

^aTherapies after discontinuation of study treatment.

^bThe duration of follow-up in censored patients was 63.44 months (range, 0 to 73.9 months) at the 5-year DCO date.

^cOne patient had their subsequent therapy updated from “IO + other” at the FA DCO to “IO + IO” at the 5-year DCO.

Sources: Sponsor’s Summary of Clinical Evidence,¹ POSEIDON Clinical Study Report,³⁸ and POSEIDON Clinical Study Report 5-year OS Addendum.⁵⁸

Efficacy

Summary of Efficacy Outcomes

Efficacy results presented are for the final analysis based on DCO dates of July 24, 2019, for all RECIST 1.1–related end points and March 12, 2021, for all other data. Results from a long-term survival follow-up analysis are also presented for OS (5-year OS DCO: August 24, 2023).^1,38,58

A summary of results for key efficacy outcomes for the durvalumab and tremelimumab plus SOC chemotherapy arm (arm 1) and SOC chemotherapy alone arm (arm 3) of the POSEIDON trial is presented in Table 21.

Table 21: Summary of Key Efficacy Results From the POSEIDON Trial

CI = confidence interval; D = durvalumab; DCO = data cut-off; EORTC QLQ-C30 = European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30; EORTC QLQ-LC13 = European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Lung Cancer 13; HR = hazard ratio; HRQoL; health-related quality of life; NA = not available; OS = overall survival; QoL = quality of life; t = tremelimumab; vs. = versus.

Note: DCO dates: July 24, 2019, and March 12, 2021 (final analysis), and August 24, 2023 (5-year OS analysis).

^aThe HR and CI were estimated from a stratified Cox proportional hazards model with the Efron method to control for ties, the stratification factors PD-L1 tumour expression (≥ 50% vs. < 50%), histology (squamous vs. nonsquamous), and disease stage (IVA vs. IVB) in the strata statement, and the CI calculated using a profile likelihood approach.

^bAn HR of less than 1 favours D + T + CT to be associated with a longer OS than CT alone.

^cP value has been adjusted for multiple testing.

^dCalculated using the Kaplan-Meier technique.

^eThe median duration of follow-up in censored patients was 63.44 months (range, 0 to 73.9 months) at the 5-year DCO date.

^fThe 5-year estimates and CIs are not alpha-controlled.

^gBased on a Lan-DeMets alpha spending function with an O’Brien-Fleming boundary with the actual number of events observed, the boundaries for declaring statistical significance are 0.00735 for a 1% overall alpha. Corresponding CIs are shown.

^hP values were generated using the stratified log-rank test adjusting for PD-L1 (≥ 50% vs. < 50%), histology (squamous vs. nonsquamous), and disease stage (IVA vs. IVB) and using the Breslow approach for handling ties.

ⁱAn odds ratio greater than 1 favours D + T + CT compared to CT alone.

^jPatients who have not shown a clinically meaningful deterioration, or who showed clinically meaningful deterioration or died after 2 or more missed visits, will be censored at the time of the latest evaluable PRO assessment (or at the latest evaluable PRO assessment before the 2 missed visits), or at the date of randomization if there are no evaluable visits or no baseline data.

Sources: Sponsor’s Summary of Clinical Evidence,¹ POSEIDON Clinical Study Report,³⁸ POSEIDON Clinical Study Report 5-year OS update,⁵⁸ and sponsor’s data on file.¹³

Overall Survival

As of the March 12, 2021, final analysis DCO, there were 251 deaths in the durvalumab and tremelimumab plus SOC chemotherapy arm and 285 deaths in the SOC chemotherapy alone arm for a total of 536 deaths across the 2 groups (79.4% maturity for OS overall).^1,38 The final analysis of OS met the prespecified boundary for declaring statistical significance between the durvalumab and tremelimumab plus SOC chemotherapy versus SOC chemotherapy alone arms (2-sided P value boundary of 0.00797 for a 1% overall alpha recycled from the significant PFS test of durvalumab and tremelimumab plus SOC chemotherapy versus SOC chemotherapy alone in the FAS). The median OS was greater in the durvalumab and tremelimumab plus SOC chemotherapy group (14.0 months; 95% CI, 11.7 to 16.1 months) than in the SOC chemotherapy alone group (11.7 months; 95% CI, 10.5 to 13.1 months), with an HR of 0.77 (95% CI, 0.650 to 0.916; P = 0.00304), as presented in Table 21.

As of the August 24, 2023, DCO for 5-year OS, there were 279 deaths in the durvalumab and tremelimumab plus SOC chemotherapy arm and 304 deaths in the SOC chemotherapy alone arm for a total of 583 deaths across the 2 groups (86.4% maturity for OS overall, representing an approximate 7% increase in data maturity from the final analysis DCO).^1,38 The median OS was greater in the durvalumab and tremelimumab plus SOC chemotherapy group (14.0 months; 95% CI, 11.7 to 16.1 months) than in the SOC chemotherapy alone group (11.6 months; 95% CI, 10.5 to 13.1 months), with an HR of 0.76 (95% CI, 0.642 to 0.893), as presented in Table 21. The 5-year estimates and CIs are not alpha-controlled.

The median duration of follow-up for the final analysis DCO of March 12, 2021, in censored patients in the durvalumab and tremelimumab plus SOC chemotherapy arm was 35.09 months (range, 1.5 to 43.9) and in the SOC chemotherapy alone arm was 34.18 months (range, 0.0 to 43.9). Overall, the median duration of follow-up in censored patients across the 3 treatment arms was 34.89 months (range, 0.0 to 44.5). In all patients, the median duration of follow-up in the durvalumab and tremelimumab plus SOC chemotherapy arm was 13.63 months (range 0.3 to 43.9) and in the SOC chemotherapy alone arm it was 11.17 months (range, 0.0 to 43.9); the median duration of follow-up in all patients across the 3 treatment arms was 12.52 months (range, 0.0 to 44.5).³⁸

The median duration of follow-up for the 5-year OS DCO of August 24, 2023, in censored patients was 63.44 months (range, 0 to 73.9). In the durvalumab and tremelimumab plus SOC chemotherapy arm the median duration of follow-up in censored patients was 64.30 months (range, 1.5 to 73.3) and in the SOC chemotherapy alone arm it was 61.04 months (range, 0 to 73.3). In all patients, the median duration of follow-up in the durvalumab and tremelimumab plus SOC chemotherapy arm was 13.63 months (range, 0.3 to 73.3) and in the SOC chemotherapy alone arm it was 11.10 months (range, 0 to 73.3).⁵⁸

Kaplan-Meier plots of OS for the comparison of durvalumab and tremelimumab plus SOC chemotherapy versus SOC chemotherapy alone at the final analysis and the 5-year OS analysis are presented in Figure 5 and Figure 6, respectively.

Figure 5: Kaplan-Meier Plot of OS — Durvalumab and Tremelimumab Plus Chemotherapy vs. Chemotherapy Alone, Final Analysis (FAS)

CI = confidence interval; Durva = durvalumab; FAS = full analysis set; OS = overall survival; SoC = standard of care chemotherapy; Treme = tremelimumab; vs. = versus.

Notes: One patient died 1 day before randomization and was censored at day 1. Data cut-off date: March 12, 2021.

The + symbol on the curves represents censored observations.

Source: POSEIDON Clinical Study Report.⁶¹

Figure 6: Kaplan-Meier Plot of OS — Durvalumab and Tremelimumab Plus Chemotherapy vs. Chemotherapy Alone, 5-Year OS Update (FAS)

CI = confidence interval; CT = standard-of-care chemotherapy; D = durvalumab; FAS = full analysis set; HR = hazard ratio; mOS = median overall survival; OS = overall survival; t = tremelimumab; vs. = versus.

Source: Sponsor’s Summary of Clinical Evidence.¹

Sensitivity Analyses of OS

Sensitivity analysis for OS adjusting for electronic case report form–derived stratification variables of PD-L1 status (≥ 50% versus < 50%), histology (squamous versus nonsquamous), and disease stage (IVA versus IVB) resulted in HRs of 0.78 (95% CI, 0.663 to 0.922) at the 5-year OS analysis and 0.79 (95% CI, 0.667 to 0.940, 2-sided P value = 0.008) at the previous final analysis.^1,58,61

Two multivariate Cox models were used: model 1 adjusted for treatment and the stratification factors only and model 2 adjusted for treatment, stratification factors, and additional prespecified covariates. The HR results were 0.78 (95% CI, 0.658 to 0.925) from model 1 and 0.76 (95% CI, 0.635 to 0.903) from model 2.^1,38

The impact of nonproportional hazards was assessed in a sensitivity analysis using a 3-component, stratified max-combo test with the same stratification factors as the primary analysis; a restricted mean survival time of an AUC approach was also performed. These analyses were generally consistent with the primary analyses.¹

Subgroup Analyses of OS

At the final analysis DCO (March 12, 2021), the survival benefit of durvalumab and tremelimumab plus SOC chemotherapy compared to SOC chemotherapy alone was consistent across all subgroups, except for the subgroup of never smokers (HR = 1.15; 95% CI, 0.792 to 1.665). An OS benefit was observed in the subgroup of patients with nonsquamous histology for durvalumab and tremelimumab plus SOC chemotherapy compared to SOC chemotherapy alone (HR = 0.70; 95% CI, 0.558 to 0.870).^1,38,61

At the 5-year OS analysis DCO (August 24, 2023), the survival benefit of durvalumab and tremelimumab plus SOC chemotherapy compared to SOC chemotherapy alone was also consistent across subgroups (estimated HRs were < 1), except for the subgroup of never smokers (HR = 1.17; 95% CI, 0.820 to 1.661), as shown in Figure 7. An OS benefit of durvalumab and tremelimumab plus SOC chemotherapy compared to SOC chemotherapy alone was also observed in patients with nonsquamous histology (HR = 0.69; 95% CI, 0.56 to 0.85) at the 5-year analysis, with OS rates of 20.5% in the durvalumab and tremelimumab plus SOC chemotherapy arm and 9.1% in the SOC chemotherapy alone arm, as presented in Figure 8.^1,58

The POSEIDON study was not powered for any individual subgroup evaluations, adjustments were not made for multiple testing, and there were low numbers of patients and events across individual subgroups.^1,38

Post Hoc Exploratory Subgroup Analysis of OS in KRAS and STK11 Mutation Subgroups

Post hoc exploratory subgroup analyses were performed in patients with nonsquamous histology for subgroups with KRAS mutations and those with STK11 mutations. At the 5-year OS analysis DCO (August 24, 2023), an OS benefit of durvalumab and tremelimumab plus SOC chemotherapy compared to SOC chemotherapy alone was observed in both subgroups.¹ For the KRAS mutation subgroup, the median OS in the durvalumab and tremelimumab plus SOC chemotherapy group (n = 60) was 25.7 months (95% CI, 9.9 to 36.7) and in the SOC chemotherapy alone group (n = 53) it was 10.4 months (95% CI, 7.3 to 12.6) with an HR of 0.55 (95% CI, 0.36 to 0.83). For the STK11 mutation subgroup, the median OS in the durvalumab and tremelimumab plus SOC chemotherapy group (n = 31) was 15.0 months (95% CI, 8.2 to 23.8) and in the SOC chemotherapy alone group (n = 22) it was 10.7 months (95% CI, 6.0 to 14.9), with an HR of 0.57 (95% CI, 0.32 to 1.04).^1,62

The POSEIDON study was not powered for these post hoc subgroup evaluations and there were low numbers of patients and events across the subgroups.¹

Figure 7: Overall Survival in Patient Subgroups With Durvalumab and Tremelimumab Plus Chemotherapy vs. Chemotherapy Alone, 5-Year OS Analysis

AJCC = American Joint Committee on Cancer; CT = standard-of-care chemotherapy; D = durvalumab; DCO = data cut-off; ECOG = Eastern Cooperative Oncology Group Performance Status; HR = hazard ratio; NSQ = nonsquamous; OS = overall survival; SQ = squamous; t = tremelimumab; TC = tumour cell; vs. = versus.

Note: HRs and 95% confidence intervals in subgroups were estimated using an unstratified Cox proportional hazards model. DCO: August 24, 2023 (5-year OS analysis).

Source: Sponsor’s Summary of Clinical Evidence.¹

Figure 8: Kaplan-Meier Plot of Overall Survival — Durvalumab and Tremelimumab Plus Chemotherapy vs. Chemotherapy Alone (Nonsquamous Histology Subgroup), 5-Year OS Analysis

CI = confidence interval; CT = standard-of-care chemotherapy; D = durvalumab; DCO = data cut-off; HR = hazard ratio; mOS = median overall survival; OS = overall survival; t = tremelimumab; vs. = versus.

Note: DCO: 24 August 2023 (5-year OS analysis).

^a HRs and 95% CIs in subgroups were estimated using an unstratified Cox proportional hazards model.

Source: Sponsor’s Summary of Clinical Evidence.¹

Progression-Free Survival

At the July 24, 2019, final analysis DCO, 238 PFS events had occurred in the durvalumab and tremelimumab plus SOC chemotherapy arm and 258 PFS events had occurred in the SOC chemotherapy alone arm, for a total of 496 PFS events across the 2 groups (73.5% maturity for PFS overall). The final analysis of PFS met the prespecified boundary for declaring statistical significance between the durvalumab and tremelimumab plus chemotherapy versus chemotherapy alone arms (2-sided P value boundary of 0.00735 for a 1% overall alpha recycled from the significant PFS test of durvalumab plus chemotherapy versus chemotherapy alone in the FAS).^1,38

The median PFS assessed by BICR in the durvalumab and tremelimumab plus SOC chemotherapy arm (6.2 months; 95% CI, 5.0 to 6.5 months) was greater than that in the SOC chemotherapy alone arm (4.8 months; 95% CI, 4.6 to 5.8 months), with an HR of 0.72 (95% CI 0.600 to 0.860; 2-sided P = 0.00031), as presented in Table 21.^1,38 The PFS rate at 12 months for patients in the durvalumab and tremelimumab plus SOC chemotherapy arm was 26.6% (95% CI, 21.7% to 31.7%) and for patients in the SOC chemotherapy alone arm it was 13.1% (95% CI, 9.3% to 17.6%).

At the final analysis DCO, 24.6% and 12.8% of randomized patients were still progression-free at the time of analysis in the durvalumab and tremelimumab plus SOC chemotherapy and SOC chemotherapy alone arms, respectively. In total, 18.9% and 16.6% of patients had a progression event because of death in the absence of RECIST progression in the durvalumab and tremelimumab plus SOC chemotherapy and SOC chemotherapy alone arms, respectively.

The median durations of follow-up in censored patients for the final analysis DCO of July 24, 2019, were 11.99 months (range, 0.0 to 23.1) in the durvalumab and tremelimumab plus SOC chemotherapy arm and 5.5 months (range, 0.0 to 22.8) in the SOC chemotherapy alone arm. Overall, the median duration of follow-up in censored patients across the 3 treatment arms was 10.33 months (range, 0.0 to 23.1).³⁸

A summary of PFS in the durvalumab and tremelimumab plus SOC chemotherapy arm versus SOC chemotherapy alone arm is presented in Table 22. A Kaplan-Meier plot of PFS for the comparison of durvalumab and tremelimumab plus chemotherapy versus chemotherapy alone is presented in Figure 9.

Table 22: Progression-Free Survival for Durvalumab and Tremelimumab Plus SOC Chemotherapy vs. SOC Chemotherapy Alone Based on BICR According to RECIST 1.1 (FAS)

BICR = blinded independent central review; CI = confidence interval; D = durvalumab; FAS = full analysis set; HR = hazard ratio; PFS = progression-free survival; RECIST 1.1 = Response Evaluation Criteria in Solid Tumours Version 1.1; SOC = standard of care; t = tremelimumab; vs. = versus.

Note: One patient died 1 day before randomization and was censored on day 1. Data cut-off: July 24, 2019.

^aPatients who did not progress or die, or who progressed or died after 2 or more missed visits, were censored at the latest evaluable RECIST assessment or at day 1 if there were no evaluable visits or no baseline data and patient did not die within 2 visits of baseline.

^bRECIST 1.1 progression event occurred after 2 or more missed visits or within 2 visits of baseline without any evaluable visits or baseline data.

^cDeath occurred after 2 or more missed visits in the absence of progression.

^dCalculated using the Kaplan-Meier technique.

^eThe HR and CI were estimated from a stratified Cox proportional hazards model with the Efron method to control for ties, the stratification factors PD-L1 (≥ 50% vs. < 50%), histology (squamous vs. nonsquamous), and disease stage (IVA vs. IVB) in the strata statement, and the CI was calculated using a profile likelihood approach.

^fAn HR of less than 1 favours T + D + chemotherapy to be associated with a longer PFS than chemotherapy alone.

^gBased on a Lan-DeMets alpha spending function with O’Brien-Fleming type boundary with the actual number of events observed, the boundaries for declaring statistical significance are 0.00735 for a 1% overall alpha. Corresponding CIs are shown.

^hP values were generated using the stratified log-rank test adjusting for PD-L1 (≥ 50% vs. < 50%), histology (squamous vs. nonsquamous), and disease stage (IVA vs. IVB) and using the Breslow approach for handling ties.

Sources: Sponsor’s Summary of Clinical Evidence,¹ POSEIDON Clinical Study Report,³⁸ and the sponsor’s Summary of Clinical Evidence.

Figure 9: Kaplan-Meier Plot of PFS for Durvalumab and Tremelimumab Plus Chemotherapy vs. Chemotherapy Alone (FAS)

CI = confidence interval; CT = standard-of-care chemotherapy; D = durvalumab; FAS = full analysis set; HR = hazard ratio; PFS = progression-free survival; t = tremelimumab; vs. = versus.

Note: Data cut-off: July 24, 2019.

Source: Sponsor’s Summary of Clinical Evidence.¹

Sensitivity Analyses of PFS

Robustness of the PFS effect to possible sources of bias in the measurement of PFS were assessed via sensitivity analyses. HRs for sensitivity analyses ranged from 0.66 to 0.74.^38,61

Censoring for PFS occurred for 100 patients (29.6%) in the durvalumab and tremelimumab plus SOC chemotherapy arm and 79 patients (23.4%) in the SOC chemotherapy alone arm. A Kaplan-Meier plot of time to censoring that was used to further assess any potential attrition bias showed that some censoring occurred earlier in the SOC chemotherapy alone arm in comparison to the durvalumab and tremelimumab plus SOC chemotherapy arm.³⁸

Two multivariate Cox models were used: model 1 adjusted for treatment and the stratification factors only, and model 2 adjusted for treatment, stratification factors, and additional prespecified covariates. The HRs were 0.71 (95% CI, 0.598 to 0.854) from model 1 and 0.70 (95% CI, 0.585 to 0.845) from model 2.³⁸

Subgroup Analyses of PFS

Prespecified subgroup analyses to investigate the consistency of treatment effect across prespecified stratification factors, demographics, planned chemotherapy administered, and baseline disease characteristics showed that the PFS benefit of durvalumab and tremelimumab plus SOC chemotherapy versus SOC chemotherapy alone was observed across all prespecified subgroups (Figure 10). Estimated HRs for all subgroups were in favour of durvalumab and tremelimumab plus SOC chemotherapy (HR < 1) and consistent with the overall HR for PFS. A PFS benefit of durvalumab and tremelimumab plus SOC chemotherapy compared to SOC chemotherapy alone was observed in patients with nonsquamous histology (HR = 0.66; 95% CI, 0.52 to 0.84).¹

No significant qualitative interaction was found when the stratification factors of PD-L1 status (TC ≥ 50% versus < 50%), histology (squamous versus nonsquamous) and disease stage (IVA versus IVB) were tested for any interaction with treatment (P = 0.500 for all factors).^1,38

The POSEIDON study was not powered for any individual subgroup evaluations, adjustments were not made for multiple testing, and there were low numbers of patients and events across individual subgroups.^1,38

Figure 10: PFS in Patient Subgroups With Durvalumab and Tremelimumab Plus SOC Chemotherapy vs. SOC Chemotherapy Alone

AJCC = American Joint Committee on Cancer; CI = confidence interval; ECOG = Eastern Cooperative Oncology Group; HR = hazard ratio; SOC = standard of care; TC = tumour cell; vs. = versus.

Note: Data cut-off: July 24, 2019. HRs and 95% CIs in the intention-to-treat population were estimated using a Cox proportional hazards model stratified by PD-L1 expression status, histology, and disease stage. HRs and 95% CIs in subgroups were estimated using an unstratified Cox proportional hazards model.

Source: Sponsor’s Summary of Clinical Evidence.¹

Objective Response Rate

At the July 24, 2019, final analysis DCO, the prespecified end point of ORR using unconfirmed responses based on BICR was greater in the durvalumab and tremelimumab plus SOC chemotherapy arm than in the SOC chemotherapy alone arm (46.3% versus 33.4%, respectively), with an OR for durvalumab and tremelimumab plus SOC chemotherapy versus SOC chemotherapy alone of 1.72 (95% CI, 1.260 to 2.367; nominal P < 0.001), as presented in Table 21.^1,38

Post hoc analysis of ORR using confirmed responses based on BICR showed that the ORR was greater in the durvalumab and tremelimumab plus SOC chemotherapy arm than in the SOC chemotherapy alone arm (38.8% versus 24.4%), with an OR for durvalumab and tremelimumab plus SOC chemotherapy versus SOC chemotherapy alone of 2.00 (95% CI, 1.428 to 2.807; nominal P < 0.001), as presented in Table 23.^1,38

Table 23: Objective Response Rate (Confirmed and Unconfirmed) Based on BICR Assessments According to RECIST 1.1 (FAS, Patients With Measurable Disease at Baseline) for Durvalumab and Tremelimumab Plus Chemotherapy vs. Chemotherapy Alone

BICR = blinded independent central review; CI = confidence interval; D = durvalumab; FAS = full analysis set; ORR = objective response rate; RECIST 1.1 = Response Evaluation Criteria in Solid Tumours Version 1.1; t = tremelimumab; vs. = versus.

Note: The prespecified secondary outcome of ORR (based on BICR) using unconfirmed responses was the number (percentage) of patients with at least 1 visit response of complete response or partial response.^1,38,47 Post hoc sensitivity analysis of ORR (based on BICR) was conducted using confirmed responses, which were defined as the number (percentage) of patients with at least 1 visit response of complete response or partial response and a confirmatory scan no sooner than 4 weeks after the initial complete response or partial response.^1,38 The analysis was performed using logistic regression adjusting for PD-L1 (≥ 50% vs. < 50%), histology (squamous vs. nonsquamous), and disease stage (IVA vs. IVB), with the CI calculated using a profile likelihood approach and the P value calculated based on twice the change in log-likelihood resulting from the addition of a treatment factor to the model. One patient died 1 day before randomization and was censored at day 1. Data cut-off date: July 24, 2019.

^aAn odds ratio greater than 1 favours T + D + chemotherapy compared to chemotherapy alone.

Sources: Sponsor’s Summary of Clinical Evidence¹ and POSEIDON Clinical Study Report.³⁸

Best Objective Response

BoR results in the durvalumab and tremelimumab plus SOC chemotherapy and SOC chemotherapy alone arms at the July 24, 2019, DCO are presented in Table 24. The BoR was a CR for 2 patients (0.6%) in the durvalumab and tremelimumab plus SOC chemotherapy arm whereas no patients in the SOC chemotherapy alone arm had a BoR of a CR. BoR was a PR for 153 patients (45.7%) in the durvalumab and tremelimumab plus SOC chemotherapy arm and 111 patients (33.4%) in the SOC chemotherapy alone arm. In the durvalumab and tremelimumab plus SOC chemotherapy and SOC chemotherapy alone arms, 120 (35.8%) and 150 (45.2%) of the patients, respectively, had a BoR of stable disease of 6 weeks or longer; 26 patients (7.8%) and 36 patients (10.8%), respectively, had a BoR of RECIST 1.1–based progression, and 22 (6.6%) and 25 (7.5%), respectively, died.^1,38

Table 24: Best Objective Response Based on BICR Assessments According to RECIST 1.1 (Unconfirmed and Confirmed Responses) (FAS, Patients With Measurable Disease at Baseline) for Durvalumab and Tremelimumab plus SOC Chemotherapy vs. SOC Chemotherapy Alone

BICR = blinded independent central review; D = durvalumab; FAS = full analysis set; RECIST 1.1 = Response Evaluation Criteria in Solid Tumours Version 1.1; SOC = standard of care; t = tremelimumab; vs. = versus.

Note: Data cut-off date: July 24, 2019.

^aResponse did not require confirmation.

^bIn practice, 5 weeks was considered the threshold to allow for the 1-week permitted time window.

Sources: Sponsor’s Summary of Clinical Evidence¹ and POSEIDON Clinical Study Report.³⁸

Duration of Response

At the July 24, 2019, final analysis DCO, in the prespecified analysis of DoR using unconfirmed responses based on BICR, the durvalumab and tremelimumab plus SOC chemotherapy and SOC chemotherapy alone arms both had a median time to response of 1.5 months (Table 25). Median DoR was longer in the durvalumab and tremelimumab plus SOC chemotherapy arm (7.4 months) than in the SOC chemotherapy alone arm (4.2 months), as presented in Table 25. The proportion of responders with an estimated DoR of 12 months or longer was greater in the durvalumab and tremelimumab plus SOC chemotherapy arm (42.5%) than in the SOC chemotherapy alone arm (16.4%) (Figure 11).^1,38

Post hoc analysis of DoR using confirmed ORR assessed by BICR also demonstrated a similar median time to response in the durvalumab and tremelimumab plus SOC chemotherapy and SOC chemotherapy alone arms (1.5 months and 1.4 months, respectively), as presented in Table 25. In this analysis, the median DoR was also longer in the durvalumab and tremelimumab plus SOC chemotherapy arm (9.5 months) than in the SOC chemotherapy alone arm (5.1 months), as presented in Table 25. The proportion of responders with an estimated DoR of 12 months or longer was greater in the durvalumab and tremelimumab plus SOC chemotherapy arm (49.7%) than in the SOC chemotherapy alone arm (21.4%).^1,38 Curves for the proportion of patients remaining in response over time using confirmed responses³⁹ follow a similar pattern as those presented using unconfirmed responses.

Table 25: Duration and Onset of Unconfirmed and Confirmed Objective Response Based on BICR Assessments According to RECIST 1.1 (FAS, Patients With Measurable Disease at Baseline) for Durvalumab and Tremelimumab Plus SOC Chemotherapy vs. SOC Chemotherapy Alone