CADTH Reimbursement Review

Osimertinib (Tagrisso)

Sponsor: AstraZeneca Canada Inc.

Therapeutic area: Non–small cell lung cancer

This multi-part report includes:

Clinical Review

Pharmacoeconomic Review

Clinical Review

Executive Summary

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

Introduction

Lung cancer is the most commonly diagnosed cancer and a leading cause of cancer deaths in Canada, with non–small cell lung cancer (NSCLC) accounting for approximately 88% of cases.¹ Approximately half of all lung cancer cases in Canada are stage I to stage III at diagnosis,¹ and one-third of NSCLC patients have operable disease.² Overall, approximately 15% of Canadians with NSCLC have an epidermal growth factor receptor (EGFR) mutation.^3-5

The goal of treatment for patients with stage IB to stage IIIA (American Joint Committee on Cancer [AJCC] 7th edition staging system) NSCLC is to cure the disease and primarily involves surgical resection of the tumour.^2,6 After surgical resection, patients may receive adjuvant platinum-based chemotherapy.^2,6,7 Four cycles of adjuvant chemotherapy are recommended for patients with stage II to stage IIIA disease and for high-risk patients with stage IB disease (e.g., tumours > 4 cm in diameter, nodal involvement, perineural or lymphovascular invasion, or disease spread through air spaces).^2,7 Meta-analyses have estimated a 5-year overall survival (OS) benefit of approximately 4% to 5% with adjuvant platinum-based chemotherapy.^8-10 However, not all patients receive post-operative adjuvant chemotherapy.¹¹ Reasons may include that it was declined by the patient, comorbidities, complication or delay in surgery recovery, and poor performance status.¹¹ Additionally, patients who receive adjuvant chemotherapy may not finish the planned number of cycles.¹¹ Cisplatin-based chemotherapy is the preferred treatment in Canada, and carboplatin-based chemotherapy is used if there is a contraindication to cisplatin.^2,7 Adjuvant radiotherapy after surgical resection is not routinely recommended.^2,7 After adjuvant chemotherapy is complete, patients receive active surveillance, which includes CT scans every 3 months to 6 months for 2 years to 3 years, then annually thereafter.^6,7

Osimertinib is an oral EGFR tyrosine kinase inhibitor (TKI) available in 40 mg and 80 mg tablets.¹² The recommended dose is 80 mg taken once a day. Osimertinib is indicated as adjuvant therapy after tumour resection in patients with stage IB to stage IIIA (AJCC 7th edition) NSCLC whose tumours have EGFR exon 19 deletions or a sensitizing mutation in the EGFR gene with substitution of a leucine with an arginine at position 858 in exon 21 (L858R). A validated test is required to identify EGFR mutation–positive status before treatment.¹² Per the product monograph, patients in the adjuvant setting should receive treatment, until disease recurrence or unacceptable toxicity, for up to 3 years.

Osimertinib has been previously reviewed by CADTH for the first-line treatment of patients with locally advanced or metastatic NSCLC whose tumours have EGFR mutations, and for the treatment of patients with locally advanced or metastatic EGFR amino acid substitution from a threonine to a methionine at position 790 in EGFR (T790M) mutation–positive NSCLC who have progressed on or after EGFR TKI therapy.

The objective of this report is to perform a systematic review of the beneficial and harmful effects of osimertinib (80 mg dose, oral) as adjuvant therapy after tumour resection in patients with stage IB to stage IIIA NSCLC whose tumours have EGFR exon 19 deletions or exon 21 (L858R) substitution mutations.

Stakeholder Perspectives

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

Patient Input

CADTH received submissions from 4 patient groups for this review: the Canadian Cancer Survivor Network (CCSN), Lung Cancer Canada (LCC), Lung Health Foundation (LHF) (formerly Ontario Lung Association), and the CanCertainty Coalition (CanCertainty). The CCSN is a national network of patients, families, friends, community partners, and sponsors who promote the best standards of care regarding early diagnosis, timely treatment, follow-up care, support, survivorship, and quality of end-of-life care. LCC is a national charitable organization that provides resources for lung cancer education, patient support, research, and advocacy. The LHF is a charity that focuses on respiratory illness and lung health that provides programs and services to patients and health care providers, invests in lung research, and advocates for improved policies on lung health. CanCertainty is made up of more than 30 Canadian patient groups, cancer health charities, and caregiver organizations that, along with oncologists and cancer care professionals, work to improve the affordability and accessibility of cancer treatments. For their submissions, the CCSN and LCC conducted interviews with Canadian patients (n = 18 and n = 6, respectively) diagnosed with stage IB to stage IIIA NSCLC. The LHF conducted an online survey, for which they received responses from 11 patients with lung cancer and 2 family caregivers, and online focus groups that included 7 patients and 3 caregivers. CanCertainty developed its submission based on published reports relating to lung cancer statistics and Canadian drug coverage and on a past survey the group had conducted of more than 1,600 Nova Scotia residents from the general population.

The symptoms and challenges patients noted as being most significant were fatigue, shortness of breath, cough, difficulty fighting infection, and chest tightness. Other health issues that were mentioned included pain, wheezing, reduced appetite, weight loss, anxiety, and sadness. Patients reported that having lung cancer interfered with their daily lives and their ability to work, complete household chores, exercise, enjoy leisure activities, and socialize. Patients also noted the negative impact cancer had on taking day trips, thinking positively about the future, mental health, relationships with others, and time spent both managing symptoms and attending appointments.

The patients surveyed by the CCSN, LCC, and LHF identified improvements in the following outcomes as important: desire for a cure, delaying disease recurrence, limiting side effects, and maintaining quality of life. The CCSN felt that participants valued disease-free survival (DFS) and its association with improved quality of life. Patients from the LCC submission felt that new medications should not interfere with daily living and should allow individuals to maintain their independence at a level similar to what it was before having cancer. The LHF respondents also emphasized better management and reduction of symptoms along with improving quality of life—not just extension of life.

Clinician Input

Input From Clinical Experts Consulted by CADTH

The clinical panel reported that the current treatment for Canadian patients with surgically resected stage IB to stage IIIA (AJCC 7th edition), EGFR-mutated NSCLC is adjuvant chemotherapy followed by active surveillance or active surveillance alone. The goal of adjuvant treatment is to treat microscopic metastatic disease, reduce the risk of recurrence, and improve cure and OS rates. The clinical experts noted that while adjuvant chemotherapy is beneficial, recurrence rates are high in these patients. The clinical experts reported that when NSCLC recurs, it is typically in a setting where it is no longer curable and treatment intent is palliative. The clinical panel indicated that better treatments are needed to decrease disease recurrence and improve OS. In the absence of improved longevity, the clinical experts noted that another important outcome of adjuvant therapy may be to delay the presentation of advanced disease in a context where presenting with advanced disease has high morbidity (i.e., new brain metastasis). The clinical panel indicated that adjuvant therapy is not intended to improve symptoms.

The clinical panel reported that osimertinib would be indicated for all surgically resected stage IB to stage IIIA EGFR-mutated NSCLC patients for 3 years. The clinical panel indicated that osimertinib is not intended to replace adjuvant chemotherapy; osimertinib would be used after standard chemotherapy (if chemotherapy was indicated). The clinical panel noted that higher-risk patients who decline or are unfit to receive standard chemotherapy may be best suited to receive treatment with osimertinib. The panel also noted that patients with later-stage disease may have a larger benefit than those with earlier-stage disease. The clinical panel indicated that patients without an EGFR mutation (e.g., those with interstitial lung disease or cardiac dysfunction) or with intolerable toxicity to the drug would not be suitable for osimertinib. The clinical panel indicated that patients with stage IB disease may be less suitable for treatment because they may have a smaller benefit from adjuvant treatment and a higher cure rate from surgery, and may not want to commit to 3 years of osimertinib. The clinical panel noted there is no data on the efficacy of osimertinib in patients with resistance mutations.

The clinical panel thought that the frequency at which response to treatment is assessed should be at the clinician’s discretion. For follow-up and toxicity management, the clinical panel indicated that patients would have visits at 2 weeks and 4 weeks, blood work every 3 months, CT scans every 3 months to 6 months for the first 2 years, then CT scans annually for years 3 to 5. The members of the clinical panel indicated they would likely perform annual CT scans and visits after 5 years for additional follow-up. The clinical panel further indicated that treatment with osimertinib should be discontinued if the patient experiences disease recurrence or unacceptable toxicity.

Clinician Group Input

Input was received from 2 clinician groups: the Ontario Health (Cancer Care Ontario)’s Lung Cancer Drug Advisory Committee (OH-CCO’s L-DAC) and LCC. The OH-CCO’s L-DAC submission included input from 5 clinicians; the LCC submission included input from 16 clinicians. Input from the clinician groups was generally consistent with that of the clinical panel consulted by CADTH. The clinician groups indicated there is a need to improve DFS and OS in patients with resected, EGFR-mutated stage IB to stage IIIA NSCLC. The clinician groups agreed that osimertinib would not replace adjuvant chemotherapy. The clinician groups thought that patients who complete 3 years of adjuvant osimertinib and relapse at least 6 months following completion of therapy would be considered for re-treatment with osimertinib therapy for advanced or metastatic disease. Similar to the clinical panel, the clinician groups agreed that re-treatment with osimertinib in the advanced or metastatic setting after use in the adjuvant setting is a consideration, but data are not available to inform on re-treatment.

Drug Program Input

The drug programs noted that osimertinib has potential for drug–drug interactions, which could potentially increase pharmacy resource use. The drug programs also indicated that osimertinib adjuvant therapy may change the place in therapy of comparator drugs and the drugs reimbursed in subsequent lines. The drug plans reported that EGFR mutation testing is not reflexively completed for early-stage NSCLC across most Canadian jurisdictions; thus, expansion of testing would be required to identify eligible patients. Lastly, the drug programs expressed concerns that the budget impact may be substantial because the duration of therapy per patient is 3 years.

In response to the drug programs’ questions regarding when adjuvant chemotherapy would benefit patients that might be considered for osimertinib, the clinical panel indicated that osimertinib is not intended to replace adjuvant chemotherapy. The clinical panel thought that osimertinib would be used after standard adjuvant chemotherapy, if chemotherapy was indicated, to further reduce the risk of disease recurrence. The drug programs also asked whether patients who receive osimertinib in the adjuvant setting and experience disease relapse while off treatment would be eligible for re-treatment with osimertinib in the metastatic setting and, if so, what time frame after completion of adjuvant therapy would be appropriate to be eligible for re-treatment. The clinical experts indicated they would restart osimertinib if patients experienced disease relapse while off treatment after completing 3 years of adjuvant osimertinib. The clinical experts indicated they would rechallenge with osimertinib earlier than 12 months off treatment if a patient relapsed. The clinical experts suggested a 2-month to 3-month off-treatment interval based on clinical experience but noted that the first CT scan occurs at 6 months after treatment unless the patient experiences other symptoms indicating recurrence.

Clinical Evidence

Pivotal Studies and Protocol-Selected Studies

Description of Studies

The systematic review of osimertinib included 1 ongoing phase III randomized controlled trial (RCT). The ADAURA trial (N = 682) is an ongoing international, multi-centre, phase III, double-blind, placebo-controlled RCT to investigate the efficacy and safety of osimertinib in patients with stage IB to stage IIIA NSCLC with a centrally confirmed common-sensitizing EGFR mutation (exon 19 deletion and/or exon 21 (L858R) substitution mutations, either alone or in combination with other EGFR mutations), who have undergone complete tumour resection, with or without post-operative adjuvant chemotherapy. Patients were randomized in a 1:1 ratio to either 80 mg osimertinib orally per day (n = 339) or matching placebo (n = 343). The primary outcome of the ADAURA trial is DFS by investigator assessment. Secondary outcomes are OS and health-related quality of life (HRQoL) assessed by the Short Form (36) Health Survey version 2 (SF-36 v2). The main HRQoL outcome measures of interest were time to deterioration (TTD) of the 2 summary scores, the Physical Component Summary (PCS) and the Mental Component Summary (MCS), in the stage II to stage IIIA population. Exploratory end points included central nervous system (CNS) DFS, disease recurrence rate, progression-free survival (PFS), time to next treatment, and TTD in PCS and MCS in the overall population (i.e., stage IB to stage IIIA).

The results from the ADAURA trial presented in this review are from an unplanned interim analysis with a data cut-off date of January 17, 2020. The study protocol and statistical analysis plan were amended to incorporate a multiple-testing procedure to account for the interim analysis, which controls the type I error for the end points of DFS and OS in the stage II to stage IIIA population and overall population.

Baseline characteristics were balanced between the treatment arms. In the overall population (stage IB to stage IIIA), the mean age of the patients was 62.1 years. The majority of patients had stage II to stage IIIA disease (68.3%), a WHO performance status of 0 (63.6%), adenocarcinoma histology type (96.5%), had undergone a lobectomy (95.3%), were Asian (63.6%), and female (70.1%). Most patients had received post-operative adjuvant chemotherapy (60.0%). Overall, 54.7% of patients had exon 19 deletions and 45.2% had exon 21 (L858R) substitution mutations.

Efficacy Results

As of the interim analysis, OS data were immature, per the sponsor’s assessment. Per the trial’s multiple-testing procedure, OS was formally tested in the stage II to stage IIIA population at the interim analysis. At the data cut-off date, 25 deaths had occurred in the stage II to stage IIIA patient population (5.3% maturity), comprising 8 deaths (3.4%) in the osimertinib arm and 17 deaths (7.2%) in the placebo arm. The hazard ratio (HR) was 0.40 (95% confidence interval [CI], 0.18 to 0.89; P = 0.0244), which did not reach statistical significance (P value < 0.0002 required). Since OS did not reach statistical significance in the primary stage II to stage IIIA population, OS in the overall population was not formally tested for statistical significance, per the multiple-testing procedure. A total of 29 patients (4.3%) in the overall population had died as of the interim analysis, comprising 9 patients (2.7%) in the osimertinib arm and 20 patients (5.8%) in the placebo arm. The HR was 0.48 (95% CI, 0.23 to 1.02).

The primary end point of DFS was met at the interim analysis. In the stage II to stage IIIA population, 26 patients (11.2%) in the osimertinib arm and 130 patients (54.9%) in the placebo arm had experienced a DFS event. The HR was 0.17 (95% CI, 0.12 to 0.23), which was statistically significant (P < 0.0001). In the overall population, 37 patients (10.9%) in the osimertinib arm and 159 patients (46.6%) in the placebo arm had experienced a DFS event. The HR was 0.20 (95% CI, 0.15 to 0.27), which was also statistically significant (P < 0.0001). The results of the pre-specified subgroup analyses were consistent with the primary analysis of DFS in showing a benefit of osimertinib (HR of < 0.4) for all subgroups. A post hoc exploratory analysis of DFS with disease recurrence in the CNS only suggested an improvement with osimertinib compared with placebo (HR = 0.14; 95% CI, 0.07 to 0.27).

As of the interim analysis, the disease recurrence rate was 10.9% in the osimertinib arm and 45.8% in the placebo arm. Data on PFS and time to next treatment were immature, per the sponsor’s assessment, and the comparisons for these end points were not controlled for multiple comparisons. As of the data cut-off date, 13 patients (3.8%) in the osimertinib arm and 46 patients (13.4%) in the placebo arm had experienced a PFS event. Thirty-one patients (9.1%) in the osimertinib arm and 134 patients (39.1%) in the placebo arm had experienced a first subsequent anti-cancer therapy or a death event. Of these events, 30 patients (96.8%) in the osimertinib arm and 125 patients (93.3%) in the placebo arm had received a subsequent anti-cancer treatment.

In the pre-specified TTD analyses of PCS and MCS in the stage II to stage IIIA population, comparisons were made without adjustment for multiple-comparison testing. For the PCS score, 58 patients (24.9%) in the osimertinib arm experienced confirmed deterioration by 3.1 points or greater or death compared with 39 patients (16.5%) in the placebo arm (HR = 1.43; 95% CI, 0.96 to 2.13). For the MCS score, 52 patients (22.3%) in the osimertinib arm and 52 patients (21.9) in the placebo arm experienced a confirmed deterioration by 3.8 points or greater or death (HR = 0.90; 95% CI, 0.61 to 1.33). The TTD in PCS and MCS scores in the overall population were analyzed as post hoc exploratory analyses and the results were consistent with the stage II to stage IIIA population.

Harms Results

A total of 329 patients (97.6%) in the osimertinib arm and 306 patients (89.2%) in the placebo arm experienced at least 1 treatment-emergent adverse event (AE) (any grade) as of the interim analysis. The most frequently reported AEs in the osimertinib and placebo arms were diarrhea (46.3% and 19.8%, respectively), paronychia (25.2% and 1.5%, respectively), dry skin (23.4% and 6.4%, respectively), pruritis (19.3% and 8.7%, respectively), and cough (18.4% and 16.6%, respectively).

In the ADAURA trial, 54 patients (16.0%) in the osimertinib arm and 42 patients (12.2%) in the placebo arm experienced a serious adverse event (SAE) as of the interim analysis. The most frequently reported SAEs in the osimertinib and placebo arms were pneumonia (1.5% and 1.2%, respectively), cataracts (0.9% and 0%, respectively), diarrhea (0.6% and 0%, respectively), acute kidney injury (0.6% and 0%, respectively), ureterolithiasis (0.6% and 0%, respectively), and femur fracture (0.6% and 0.3%, respectively).

Withdrawals specifically due to AEs were not reported. As of the data cut-off, a total of 33 patients (4.8%) had withdrawn from the ADAURA trial: 19 (5.6%) in the osimertinib arm and 14 (4.1%) in the placebo arm. Thirty-six patients (10.7%) in the osimertinib arm and 10 patients (2.9%) in the placebo arm had discontinued study treatment due to AEs. The most common AEs leading to treatment discontinuation in the osimertinib arm were interstitial lung disease (n = 8; 2.4%), diarrhea (n = 3; 0.9%), and decreased appetite (n = 3; 0.9%). The most common AE leading to treatment discontinuation in the placebo arm was decreased ejection fraction (n = 3; 0.9%).

A total of 29 patients (4.3%) had died as of the interim analysis: 9 patients (2.7%) in the osimertinib arm and 20 patients (5.8%) in the placebo arm.

Regarding notable harms, 8 patients (2.4%) experienced interstitial lung disease and 2 patients (0.6%) experienced pneumonitis in the osimertinib arm as of the interim analysis. No patients in the placebo arm experienced interstitial lung disease or pneumonitis. The frequency of cardiac disorder AEs was greater in the osimertinib arm compared with the placebo arm (11.0% versus 5.2%, respectively). In the osimertinib arm, 6.5% patients experienced QT interval prolongation compared with 1.2% in the placebo arm. Four patients (1.2%) in the osimertinib arm experienced congestive heart failure, cardiac failure, or left ventricular dysfunction compared with zero in the placebo arm. Four patients (1.2%) in the osimertinib arm experienced atrial fibrillation compared with 1 (0.3%) in the placebo arm; 6 patients (1.8%) in the osimertinib arm experienced an arrhythmia other than atrial fibrillation compared with zero in the placebo arm. Overall, 3 patients (0.4%) experienced keratitis: 2 (0.6%) in the osimertinib arm and 1 (0.3%) in the placebo arm. In the osimertinib arm, 70.6% experienced a skin or subcutaneous tissue disorder compared with 35.6% of patients in the placebo arm. The most common skin disorders in the osimertinib and placebo arms were paronychia (25.2% and 1.5%, respectively), dry skin (23.4% and 6.4%, respectively), pruritis (19.3% and 8.7%, respectively), and dermatitis acneiform (11.0% and 4.7%, respectively).

Critical Appraisal

The ADAURA trial was a double-blind RCT to minimize bias. Baseline characteristics were balanced between treatment arms and few randomized patients had been lost to follow-up as of the data cut-off. The interim analysis was not planned, and the trial is ongoing. A multiple-testing procedure was employed to control overall type I error at the 5% 2-sided level for the end points of DFS and OS, which was modified to account for the unplanned interim analyses. The primary end point was met at the interim analysis since the log-rank test for DFS in patients with stage II to stage IIIA disease met the pre-specified threshold for statistical significance. The log-rank test for DFS in the overall population also met statistical significance for this analysis. Due to early reporting of the study, data maturity is lower than planned at the interim analysis, per the sponsor. At the data cut-off date, the sponsor assessed the OS data to be immature. Furthermore, the comparison of OS in the stage II to stage IIIA population was not statistically significantly different between the treatment groups. Thus, at the time of this review, we cannot conclude that osimertinib confers an OS benefit compared with placebo. In addition, the data on time to next treatment and PFS are considered by the sponsor to be of limited clinical significance at the data cut-off of the interim analysis due to the immaturity of the data on patients who experienced a disease recurrence event, and comparisons for these end points were not controlled for multiple comparisons. The results did not support conclusions for an effect of osimertinib on PFS and time to next treatment, and any potential clinical benefit for these outcomes is associated with uncertainty due to the immaturity of the data and lack of control for type I error. Conclusions could not be drawn for the effect of osimertinib on HRQoL end points, as these end points were not adjusted for multiple comparisons. In addition, the majority of patients had not had the opportunity to receive the study treatment for the planned duration of 3 years.

The osimertinib dose and treatment regimen used in the ADAURA trial aligns with the Health Canada indication. The ADAURA trial included patients who had received standard-of-care adjuvant chemotherapy, which is commonly used in Canadian practice. This also aligns with the intended use of osimertinib in Canada, per the clinical experts consulted by CADTH and the clinician groups that provided input, both of which indicated that osimertinib is not intended to replace adjuvant chemotherapy. The clinical experts consulted by CADTH thought that the inclusion and exclusion criteria used in the trial were appropriate and generally reflected the characteristics of the intended patient population in Canada. However, the trial limited enrolment to patients with a WHO performance status of 0 to 1 and the clinical experts reported that many patients in their practice with resected stage IA to stage IIIB NSCLC have a performance status of 2. The clinical experts did not think that exclusion of patients with worse performance status limits the generalizability of the trial results. The proportion of Asian patients in the trial was higher than in the Canadian NSCLC population, per the clinical experts consulted by CADTH. In addition, the ADAURA trial reported a higher EGFR mutation–positive rate than currently seen in Canada, where EGFR genetic testing is routinely offered to patients with locally advanced disease that is not amenable to curative-intent therapy or who have metastatic NSCLC only.

Indirect Comparisons

No indirect treatment comparisons were included in the sponsor’s submission to CADTH or identified in the literature search.

Other Relevant Evidence

No long-term extension studies or other relevant studies were included in the sponsor’s submission to CADTH or identified in the literature search.

Conclusions

One ongoing phase III, double-blind, placebo-controlled RCT (ADAURA) provided direct evidence regarding the safety and efficacy of osimertinib adjuvant therapy in adult patients with resected stage IB to stage IIIA (AJCC 7th edition) NSCLC. The trial included patients regardless of whether they had received standard adjuvant chemotherapy. The ADAURA trial met its primary end point at an unplanned interim analysis that showed a statistically significant difference in DFS in both the stage II to stage IIIA population and the overall population. This DFS benefit was consistently observed in all pre-specified subgroups. The results did not support conclusions for an effect of osimertinib on OS, PFS, or time to next treatment, and any potential clinical benefit for these outcomes is associated with uncertainty due to the immaturity of the data. Conclusions could not be drawn for the effect of osimertinib on HRQoL end points, as these end points were not adjusted for multiple comparisons. The majority of study participants reported treatment-emergent AEs. A greater proportion of patients in the osimertinib arm experienced an AE compared with the placebo arm. Interstitial lung disease, pneumonitis, cardiac disorders, and skin and subcutaneous tissue disorders were reported more frequently in the osimertinib arm. These AEs are consistent with the known safety profile of osimertinib. Keratitis was uncommon in both treatment arms. A greater number of patients in the osimertinib arm discontinued study treatment due to AEs and experienced an SAE. Overall, 4% of study participants had died as of the interim analysis.

Introduction

Disease Background

Lung cancer is the most commonly diagnosed cancer and the leading cause of cancer deaths in Canada.¹ Survival from lung cancer of all stages and histologies is poor, with an overall 5-year net survival rate of 19%.¹ In 2020, it was estimated there would be 29,800 new cases of lung cancer diagnosed and 21,200 deaths from lung cancer that year.¹ It is estimated that 1 in 17 Canadians will die of lung cancer.¹⁵

Lung cancer is classified as either NSCLC or small cell lung cancer, with NSCLC accounting for approximately 88% of cases in Canada.¹ NSCLC is further classified into 3 main histologic subtypes: adenocarcinoma, squamous cell carcinoma, and large-cell carcinoma. To determine a patient’s prognosis and treatment, NSCLC is staged using the AJCC staging criteria, which involves TNM (tumour, node, metastasis) classification of the disease based on the size and spread of the primary tumour (T), lymph node involvement (N) and occurrence of metastasis (M).¹⁶ Approximately half of all lung cancer cases in Canada are stage I to stage III at diagnosis.¹ Early-stage (i.e., stage I to stage IIIA per the AJCC 7th edition) NSCLC is often asymptomatic.^16,17 If patients do present with symptoms, they are often unspecific and difficult to directly attribute to a lung cancer diagnosis.¹⁷ The most common symptoms include unspecific cough, chest and shoulder pain, hemoptysis, weight loss, dyspnea, hoarseness, bone pain, and fever.¹⁷ Diagnostic procedures include imaging of the lungs, sputum cytology, and tissue biopsy.² Approximately one-third of NSCLC patients have operable disease.²

Approximately 15% of Canadians with NSCLC have an EGFR–activating mutation in the region encoding the tyrosine kinase domain.^3-5 EGFR mutations are more frequently observed in those who have never smoked, people of Asian ethnicity, patients with adenocarcinoma, and females.^3,18 The most common EGFR mutations are exon 19 deletions and the exon 21 codon 858–point mutation (L858R).^4,5

Standards of Therapy

The goal of treatment for patients with stage IB to stage IIIA NSCLC (per the AJCC 7th edition; the equivalent stages using the AJCC 8th edition are stage IIA to stage IIIB) is to cure the disease and primarily involves surgical resection of the tumour.^2,6 After surgical resection, the majority of patients will receive 4 cycles of adjuvant platinum-based chemotherapy.^2,6,7 Meta-analyses have estimated a 5-year OS benefit of approximately 4% to 5% with adjuvant platinum-based chemotherapy.^8-10 Adjuvant chemotherapy is recommended for patients with stage II to stage IIIA disease and stage IB patients considered at high risk of relapse (e.g., tumours > 4 cm in diameter, nodal involvement, perineural or lymphovascular invasion, or disease spread through air spaces).^2,7 However, not all patients receive post-operative adjuvant chemotherapy. A retrospective real-world study conducted in Europe, which included 831 patients with stage IB to stage IIIA NSCLC, found that 52% of patients did not receive adjuvant chemotherapy (15.1% of those with stage IB, 55.1% with stage II, and 71.4% of patients with stage IIIA received adjuvant chemotherapy).¹¹ The most common reasons for not receiving adjuvant chemotherapy were that it was declined by the patient (12.6%), comorbidities (11.9%), complication or delay in surgery recovery (8.4%), and poor performance status (7.0%).¹¹Additionally, approximately one-third of patients who received chemotherapy did not finish the planned number of cycles.¹¹ In Canada, cisplatin-based chemotherapy is the preferred systemic treatment and carboplatin-based chemotherapy is used if there is a contraindication to cisplatin.^2,7 Adjuvant radiotherapy after complete surgical resection is not routinely recommended.^2,7 After adjuvant chemotherapy is complete, patients receive active surveillance, also known as “watch and wait,” which includes CT scans every 3 months to 6 months for 2 years to 3 years then annually thereafter until year 5.^6,7

Drug

Osimertinib (Tagrisso) is an oral EGFR TKI available in 40 mg and 80 mg tablets.¹² The recommended dose is 80 mg taken once a day. Osimertinib is a selective irreversible inhibitor of EGFR-sensitizing mutations and the T790M resistance mutation that has limited activity against wild-type EGFR (mutation-negative) NSCLC. Osimertinib is indicated as adjuvant therapy after tumour resection in patients with stage IB to stage IIIA (AJCC 7th edition) NSCLC whose tumours have EGFR exon 19 deletions or exon 21 (L858R) substitution mutations. A validated test is required to identify EGFR mutation–positive status before treatment.¹² Per the product monograph, patients in the adjuvant setting should receive treatment, until disease recurrence or unacceptable toxicity, for up to 3 years. Osimertinib underwent Health Canada priority review and was part of Project ORBIS.¹⁴ Health Canada issued a Notice of Compliance on January 18, 2021. The sponsor’s reimbursement request is per the Health Canada indication.

Osimertinib is also indicated for the first-line treatment of patients with locally advanced (not amenable to curative therapies) or metastatic NSCLC whose tumours have EGFR exon 19 deletions or exon 21 (L858R) substitution mutations (either alone or in combination with other EGFR mutations).¹² In addition, osimertinib is indicated for the treatment of patients with locally advanced or metastatic EGFR T790M mutation–positive NSCLC whose disease has progressed while on or after TKI therapy.¹² For this latter indication, marketing authorization was based on results from a randomized phase III trial (AURA3) demonstrating that osimertinib was superior to platinum-doublet chemotherapy in prolonging PFS as assessed by investigator using Response Evaluation Criteria in Solid Tumors Version 1.1 (RECIST 1.1) guidelines.¹² Osimertinib has been previously reviewed by CADTH for the first-line treatment of patients with locally advanced or metastatic NSCLC whose tumours have EGFR mutations and for the treatment of patients with locally advanced or metastatic EGFR T790M mutation–positive NSCLC whose disease has progressed while on or after EGFR TKI therapy.

Stakeholder Perspectives

Patient Group Input

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

About the Patient Groups and Information Gathered

CADTH received 4 submissions from patient groups for the reimbursement review of osimertinib from the CCSN, LCC, LHF (formerly Ontario Lung Association), and CanCertainty. CCSN is a national network of patients, families, friends, community partners, and sponsors who promote the best standards of care regarding early diagnosis, timely treatment, follow-up care, support, survivorship, and quality of end-of-life care. LCC is a national charitable organization that provides resources for lung cancer education, patient support, research, and advocacy. LHF is a charity that focuses on respiratory illness and lung health that provides programs and services to patients and health care providers, invests in lung research, and advocates for improved policies on lung health. CanCertainty is made up of more than 30 Canadian patient groups, cancer health charities, and caregiver organizations who, along with oncologists and cancer care professionals, work to improve the affordability and accessibility of cancer treatments.

For its submission, the CCSN contracted help from Broadstreet Health Economics and Outcomes Research (an independent consultancy specializing in epidemiology and health economics) to conduct interviews with Canadian English-speaking patients who had been diagnosed with stage IB, stage II, or stage IIIA lung cancer (33%, 17%, and 50%, respectively). Interviews were approximately 45 minutes long, held over Zoom (teleconferencing platform), and took place between January 13 and February 8, 2021. Interview participants were found through the organization’s Right2Survive mailing list. The majority of the 18 participants were female (83%), ranged in age from 51 years to 85 years, and were from across the country: Atlantic region (22%), prairies (22%), central (50%), and west coast (6%).

LCC conducted interviews with 6 Canadian patients who had stage IB to stage IIIA EGFR mutation–positive NSCLC. Females made up 83% of respondents and ages ranged from 45 years to 69 years.

In February 2021, LHF released an online survey to which they received responses from 11 patients with lung cancer and 2 family caregivers. In April 2021, 3 focus groups were held over Zoom that were made up of 7 patients and 3 caregivers. Participants were from British Columbia, Alberta, Manitoba, Ontario, Quebec, New Brunswick, and Nova Scotia. A certified respiratory educator who works with LHF also provided input and reviewed sections related to disease experience, treatment, and outcomes.

CanCertainty developed its submission based on published reports relating to lung cancer statistics and Canadian drug coverage as well as a past survey (2017) the group had conducted of more than 1,600 randomly selected individuals from the general population of Nova Scotia.

Disease Experience

Patients who were interviewed for the CCSN submission expressed fear and concern over the low survival rate and the possibility that their tumour may be inoperable. One interviewee reflected, “I guess I was kind of desperate to think I was going to be okay and asked how things looked. And, she told me at the time that I had a 15 percent chance of surviving five years… it becomes such a psychological burden that, you know, I was looking at 15 people out of a 100 survive for five years. What’s the chance that I would be one of those 15?”

From the 23 LHF respondents, the symptoms and challenges patients noted as being most significant were fatigue (64%), shortness of breath (64%), cough (27%), difficulty fighting infection (27%), and chest tightness (18%). Other health issues that were mentioned include pain, wheezing, reduced appetite, weight loss, anxiety, and sadness. Patients responded that having lung cancer interfered with their daily lives and their ability to work (54%), complete household chores (40%), exercise (40%), enjoy leisure activities (36%), and socialize (29%). The submission also noted the negative impact cancer had on taking day trips, thinking positively about the future, mental health, relationships with others, and time spent both managing symptoms and attending appointments. To further illustrate this, the following quotes describe some of their experiences, “You never stop thinking about it,” “It can be so overwhelming,” and “I have a lot of anxiety and sadness now.” The LHF focus group discussed challenges that patients faced, such as a lack of support groups for those living with lung cancer compared with other cancers, feeling that there are limited treatment options, and not being able to access services close to home.

Experiences With Currently Available Treatments

Patients from the CCSN submission had previously received surgery (89%) and chemotherapy (56%) as cancer treatments. Although most reported that their surgery was successful, 1 patient responded that the cancer had metastasized before surgery while another was devastated with how their lifestyle changed due to reduced lung function as a result of the surgery. One participant interviewed also explained that they declined chemotherapy since it was not a guaranteed cure.

LCC described that the patient experience and burden of care with chemotherapy cited by patients in the current submission is consistent with that of other lung cancer treatment settings that have been discussed in previous LCC submissions to CADTH. Financial burden and additional strain on caregivers were other factors that patients considered when deciding on treatments. They also noted that patients who receive adjuvant chemotherapy may relapse, which demonstrates the need for therapies that are effective at preventing disease recurrence.

The LHF respondents reported trying the following treatments: gefitinib, entrectinib, Anoro Ellipta, Ventolin, Trelegy, Onbrez, Alvesco, amlodipine, Lyrica, and Breo. Respondents indicated these treatments were associated with side effects that included fatigue, diarrhea, nausea, appetite loss, weight loss, heart palpitations, face blistering or rash, headaches, and difficulty sleeping.

CanCertainty noted in its submission that reimbursement of oral cancer drugs is not equal across jurisdictions in Canada. As a result, patients who do not have adequate insurance may have to pay out-of-pocket for medication and/or apply to funding-assistance programs, which can take time and delay access to treatment. The group also considered that patients who do have private insurance they may still have co-pays, deductibles, and annual or lifetime caps that increase the financial burden on them and their families. CanCertainty suggested that, due to the differential drug coverage plans among provinces and territories, patients can face discrimination based on age, income, geography, and cancer type or treatment. From their survey of Nova Scotia residents, 60% of respondents stated they would consider leaving the province if they had to pay for cancer treatments and only 7% felt they could afford monthly drug costs exceeding $200. Furthermore, CanCertainty investigated the impact of “financial toxicity” that can result from out-of-pocket costs, lost income, and treatment-related expenses. CanCertainty indicated that those at greater risk of “financial toxicity” are patients who are young, have limited income, are uninsured or unemployed, have been recently diagnosed, or have more severe cancer. To mitigate the financial burden, CanCertainty noted that patients may delay or skip treatment, opt for less expensive alternatives, or avoid appointments.

Improved Outcomes

The patients surveyed by the CCSN, LCC, and LHF identified improvements in the following outcomes as important: desire for a cure, delaying disease recurrence, limiting side effects, and maintaining quality of life.

The CCSN felt that the participants valued DFS and its association with improved quality of life, and 72% of participants indicated they would be willing to take a medication that improved DFS without data on OS. One individual stated, “You can live a decent life without being cured of cancer if you can keep it stable or under control,” while another shared, “I would go for the quality over the quantity.” Patients also felt that being involved in the treatment decision-making process was very important along with having access to comprehensive information, either through their own research or their health care team.

In addition to the improved outcomes stated earlier, patients contributing to the LCC submission felt that new medications should not interfere with daily living and should allow individuals to maintain their independence at a level similar to what it was before having cancer. Further discussion around quality of life highlighted the value that patients and families placed on being able to return to work and regular life and on being productive.

The LHF respondents also emphasized better symptom reduction and management along with improving quality of life and not just extension of life. Other topics that came up were having access to information about lung cancer and to community-based support.

Experience With Drug Under Review

LCC interviewed 1 patient who had not yet started treatment, 4 who were currently receiving osimertinib, and 1 who had past experience with it; all of these patients were described as having stage IB to stage IIIA cancer. The setting (i.e., adjuvant or metastatic) where these patients received osimertinib was not reported in the submission. Patients were able to receive osimertinib through a special access program (n = 3), insurance (n = 2), or clinical trial (n = 1). Of the 5 who received the drug, the duration of treatment ranged from 1 month to 4 months. The single patient who was part of a clinical trial had to stop after 3 months due to pneumonitis. In general, patients were described as tolerating the medication well, allowing them to return to work full time and continue with their daily lives. Patients preferred an oral pill to receiving treatment intravenously and osimertinib was described as being “much better than chemotherapy.” LCC also noted that osimertinib could reduce the burden on caregivers, the need to travel for treatment, the time spent receiving treatment, and the burden on hospitals because treatment does not require occupying a chemotherapy chair. Patients described the following side effects while receiving treatment with osimertinib: rash on face and/or chest, pneumonitis (which resulted in discontinuing the drug), skin pustules (which were treated with cortisol and resolved), and mild nausea. According to LCC, if not for assistance enabling them to access the medication (e.g., special access programs, insurance, clinical trial), costs would be prohibitive for these patients. Furthermore, the patient group indicated that a diagnosis of later-stage early lung cancer appeared to influence patients to accept osimertinib more readily.

No patients interviewed or surveyed by the CCSN or LHF had experience with osimertinib at the time of the submission.

Companion Diagnostic Test

LCC noted that EGFR genetic testing is routine for later-stage lung cancer in many cancer centres across Canada and suggested it could be expanded to include earlier-stage cancers.

Additional Information

The CCSN highlighted the value of having treatments approved quickly by decision-makers, for instance, when only DFS information is known, and OS data are not yet available. The patients emphasized that treatment approval and improved access are both key to potentially offering the best chances of survival. The group also acknowledged that differential approval across jurisdictions is another hurdle to accessing medications.

LCC noted in its submission that early treatment could not only improve patients’ long-term outcomes and prevent progression to later-stage disease, but it may also alleviate pressures on patients, families, and the health care system.

CanCertainty emphasized the importance of equitable access to cancer medications for all Canadians. The group also raised concerns over the safety of take-home cancer drugs compared with those administered in clinics by trained health care professionals, particularly with regard to dosing, handling, limited monitoring or nonadherence, and toxicity.

Clinician Input

Input From Clinical Experts Consulted by CADTH

All CADTH review teams include at least 1 clinical specialist with expertise in the diagnosis and management of the condition for which the drug is indicated. Clinical experts are a critical part of the review team and are involved in all phases of the review process (e.g., providing guidance on the development of the review protocol, assisting in the critical appraisal of clinical evidence, interpreting the clinical relevance of the results, and providing guidance on the potential place in therapy). In addition, as part of the osimertinib review, a panel of 4 clinical experts from across Canada was convened to characterize unmet therapeutic needs, assist in identifying and communicating situations where there are gaps in the evidence that could be addressed through the collection of additional data, promote the early identification of potential implementation challenges, gain further insight into the clinical management of patients living with a condition, and explore the potential place in therapy of the drug (e.g., potential reimbursement conditions). A summary of this panel discussion is presented subsequently.

Unmet Needs

The clinical panel reported that current treatment for Canadian patients with surgically resected stage IB to stage IIIA (AJCC 7th edition), EGFR-mutated NSCLC is adjuvant chemotherapy followed by active surveillance or active surveillance alone. Adjuvant chemotherapy consists of 4 cycles of platinum-doublet chemotherapy (e.g., cisplatin plus vinorelbine, cisplatin plus pemetrexed). Adjuvant chemotherapy is offered to patients with good performance status and without other comorbidities that would contraindicate chemotherapy (e.g., recent myocardial infarction, neuropathy). In patients with stage IA (AJCC 7th edition) disease, active surveillance is offered to patients with a tumour size of less than 4 cm, whereas adjuvant chemotherapy is offered to those with a tumour size of 4 cm or greater. The clinical experts reported that cisplatin-based adjuvant chemotherapy can be challenging to administer because it requires adequate performance status and renal function. Cisplatin-based adjuvant chemotherapy also has substantial toxicity (e.g., chronic renal failure, future blood dyscrasias, neuropathy, hearing loss) and some patients are not candidates for it.

The clinical panel indicated that the goal of adjuvant treatment is to treat microscopic metastatic disease, reduce the risk of recurrence, improve cure rates, and improve OS. The clinicians noted they are not able to identify patients who are cured with surgery alone versus those who have microscopic residual disease; therefore, adjuvant chemotherapy is offered to all eligible patients. The clinical experts reported that adjuvant chemotherapy improves cure by approximately 5% to 15%, depending on disease stage. The members of the clinical panel indicated that, in their experience, approximately one-quarter to one-third of patients decline adjuvant chemotherapy because the patients consider this small clinical benefit to be not worth the toxicity associated with adjuvant chemotherapy treatment. Other patients are unfit for 4 cycles of chemotherapy.

The clinical experts noted that while adjuvant chemotherapy is beneficial, recurrence rates remain high in these patients. The clinical panel reported that many patients with resected lung cancer eventually relapse despite receiving adjuvant chemotherapy. The clinical experts reported that when NSCLC recurs, it is typically in a setting where the disease is no longer curable and treatment intent is considered palliative. The clinical experts estimated that greater than 50% to 60% of patients with resected stage IIIA NSCLC relapse with incurable disease. The clinical panel indicated that better treatments are needed to decrease disease recurrence and improve OS.

In the absence of improved survival, the clinical experts noted that other important outcomes of adjuvant therapy may be to delay the presentation of advanced disease, in the context where presenting with advanced disease has high morbidity (e.g., new brain metastasis). The clinical experts also noted that some patients may be cured with surgery; therefore, an ideal adjuvant therapy would have a minimal side effect profile and minimal risk from long-term complications. The clinical panel indicated that adjuvant therapy is not intended to improve symptoms.

Place in Therapy

The clinical panel thought that osimertinib would be indicated for 3 years for all patients with surgically resected stage IB to stage IIIA (AJCC 7th edition) EGFR-mutated NSCLC. The clinical experts estimated this would include approximately 15% to 20% of patients with lung cancer in Canada. The clinical panel indicated that osimertinib is not intended to replace adjuvant chemotherapy; rather, osimertinib would be used after standard chemotherapy (if chemotherapy was indicated). As such, the clinical panel thought that osimertinib likely would not cause a shift in the current treatment paradigm, except in patients with EGFR-mutated stage IB disease (tumour < 4 cm) where often only active surveillance is required after tumour resection. In this setting, the clinical panel thought that some clinicians might offer 3 years of adjuvant osimertinib only. In addition, osimertinib could be considered for patients who are unfit for adjuvant chemotherapy.

The clinical experts reported that, in their experience, most patients (approximately two-thirds to three-quarters) who are offered adjuvant chemotherapy will complete 4 cycles. However, the clinical experts thought that there is no reason that patients would need to complete all 4 cycles of chemotherapy before starting osimertinib, and that osimertinib could be offered to patients who receive fewer than 4 cycles (i.e., the clinical experts thought there is no required minimum number of cycles). The clinical panel also noted there is a small group of patients who would be eligible for osimertinib and who would not be offered chemotherapy (e.g., stage IB, unfit). The clinical experts reiterated that osimertinib is not a substitute for chemotherapy.

The clinical experts predicted that the use of osimertinib in the adjuvant setting would necessitate a change in the timing for EGFR genetic testing in NSCLC patients. In Canada, EGFR genetic testing is routinely done in patients with late-stage NSCLC. EGFR genetic testing often is not performed in patients with early-stage NSCLC due to the cost and the fact that it likely would not impact early care. The clinical panel noted that EGFR genetic testing availability varies by jurisdiction. The clinical panel noted there is a potential added cost to treatment if EGFR genetic testing occurs earlier in the treatment course. They noted that a subset of early-stage patients (e.g., stage IB) would be cured with resection and adjuvant chemotherapy and, thus, would never need testing for metastatic disease, which is when most testing occurs. The clinical panel noted that osimertinib is currently available in some jurisdictions through special access programs in the adjuvant setting for patients with stage IB to stage IIIA resected EGFR-mutated NSCLC.

Patient Population

The clinical panel indicated that patients who meet the eligibility criteria for the ADAURA trial (i.e., those with EGFR-mutated, resected, stage IB to stage IIIA NSCLC) would be suitable for treatment with osimertinib. The clinical panel thought that higher-risk patients who decline or are unfit to receive standard chemotherapy may be best suited to receive treatment with osimertinib. The panel also thought that patients with later-stage disease may receive a larger benefit with adjuvant osimertinib than those with earlier-stage disease.

The clinical panel indicated that patients without an EGFR mutation and those who have intolerable toxicity to the drug (e.g., interstitial lung disease or cardiac failure) would not be suitable for osimertinib. The clinical panel thought that patients with stage IB disease may be less suitable for treatment because they receive less clinical benefit from adjuvant treatment, experience a higher cure rate from surgery, and may not want to commit to 3 years of osimertinib. However, the clinical experts noted that, with the current treatment options, recurrence is approximately 25% in patients with stage IB disease. In addition, the clinical panel thought that patients with resistance mutations could be less suitable for treatment with osimertinib, but there are no data in that setting.

Assessing Response to Treatment

The clinical panel identified the following outcomes of interest when assessing response to osimertinib treatment: no disease recurrence, OS, DFS, and prevention of CNS metastases.

The clinical panel thought that the frequency at which treatment response is assessed should be at the clinician’s discretion. For follow-up and toxicity management, the members of the clinical panel indicated they would require visits at 2 weeks and 4 weeks, blood work every 3 months, CT scans every 3 months to 6 months for the first 2 years, and then CT scans annually for years 3 to 5. The clinical panel indicated they would likely perform annual CT scans and require visits after 5 years for additional follow-up.

Discontinuing Treatment

The clinical panel indicated that treatment with osimertinib should be discontinued if the patient experiences disease recurrence or unacceptable toxicity (e.g., SAEs or chronic AEs that do not respond to dose reduction). The clinical panel thought that patients whose disease progresses while on osimertinib would be discontinued and treated with platinum-doublet chemotherapy or be assessed for enrolment in a clinical trial.

Prescribing Conditions

The clinical panel indicated that osimertinib should be prescribed and monitored by medical oncologists, general practitioners of oncology, or nurse practitioners who have been trained in oncology. Patients would be treated in the outpatient community setting.

Additional Considerations

The clinical experts were uncertain about the time frame for re-treatment with osimertinib if a patient relapsed while off treatment (i.e., patient completed 3 years of osimertinib adjuvant therapy then experienced disease recurrence or relapsed with metastatic disease). They were also uncertain about defining an off-treatment interval (after completing 3 years of osimertinib adjuvant therapy) before restarting osimertinib due to the limited evidence available to inform on this issue. In patients who progress while off treatment with a treatment-free interval of 3 months or greater, the clinical experts indicated they would try rechallenging with osimertinib. In that situation, the clinical experts indicated they might re-biopsy the metastatic disease and look at genetic testing for a mechanism resistant to osimertinib. If there was no resistance, they would restart osimertinib treatment. The clinical panel noted that patients who restart treatment usually respond immediately (e.g., within 3 weeks), and this response can be observed clinically and/or with a chest X-ray rather than needing a CT scan. If the patient did not experience benefit after approximately 3 months of treatment, they would stop osimertinib. The clinical experts indicated they would be unlikely to use another TKI when rechallenging. The clinical experts also noted that, in their experience, patients with EGFR-mutated NSCLC do not respond well to immunotherapy (< 10% response rate); thus, they thought that immunotherapy is unlikely to be offered to these patients.

Clinician Group Input

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

Input was received from 2 clinician groups on the reimbursement review of osimertinib as adjuvant therapy after tumour resection for the treatment of patients with stage IB to stage IIA NSCLC whose tumours have EGFR exon 19 deletions or exon 21 (L858R) substitution mutations.

The OH-CCO’s L-DAC provides evidence-based clinical and health system guidance on drug-related issues in support of its mandate, including the Provincial Drug Reimbursement Programs and the Systemic Treatment Program. Information for this review was collected through joint discussions through emails. Four physicians (1 cardiothoracic surgeon and 3 medical oncologists) and 1 pharmacist provided input for this submission.

LCC is a national charitable organization that is a resource for lung cancer education, patient support, research, and advocacy. The organization is based in Toronto, Ontario and includes both regional and pan-Canadian initiatives. The organization is also a member of the Global Lung Cancer Coalition. Information for this submission was gathered through relevant published clinical data and expert evidence-based review among lung cancer medical oncologists across Canada. Sixteen physicians (14 medical oncologists and 2 pathologists) provided input for this submission.

Unmet Needs

The OH-CCO’s L-DAC noted that adjuvant chemotherapy is offered as standard therapy to patients with resected primary tumours greater than 4 cm or who have other high-risk features such as positive lymph nodes, invasion of the chest wall, or satellite nodules in the same lung or the ipsilateral lung. The clinician group added this includes most patients with stage IB to stage IIIA tumours, based on the AJCC 7th edition staging system, and equivalent stage IIA to stage IIIB tumours, per the 8th edition. LCC noted that in current Canadian practice, treatments for early-stage NSCLC do not differentiate between patients with and without EGFR mutations. Adjuvant platinum-doublet chemotherapy given after tumour resection typically consists of 4 cycles of treatment, with each cycle lasting 21 days for a total of 12 weeks of therapy.

LCC indicated that the primary goal of treatment for patients with stage IB to stage IIIA disease (AJCC 7th edition) is a cure, which is reflected in improvements in 5-year OS rates. Both clinician groups indicated that the goal of adjuvant therapy is to improve OS and HRQoL. The clinician groups thought that increasing DFS could translate into improved OS. The OH-CCO’s L-DAC added that in situations where recurrence is often symptomatic and unpredictable, improving DFS may improve HRQoL. Both groups also agreed that the side effects of adjuvant therapy should be minimal.

The OH-CCO’s L-DAC noted that despite adjuvant chemotherapy, the risk of disease recurrence and death remains high in patients with resected stage IIA to stage IIIB NSCLC. The clinician group reported that the 5-year OS rate ranges from 70% for stage IIA to 40% for stage IIIB. Similarly, LCC indicated that current therapies are inadequate to achieve high rates of cure and certain patient outcomes, particularly DFS and OS, remain poor despite adjuvant chemotherapy. In addition, the OH-CCO’s L-DAC noted that adjuvant chemotherapy is not used or accepted by a large portion of lung cancer patients and more than 30% of patients do not receive adjuvant chemotherapy. The clinician groups reported that while targeted therapies are routinely used in more advanced EGFR-mutated NSCLC, there are currently no targeted therapies available in the adjuvant setting.

The OH-CCO’s L-DAC reported that patients with EGFR-mutated NSCLC are often non-smokers and therefore would not be eligible for current screening programs for lung cancer; thus, these patients have a greater unmet need for an effective treatment. According to LCC, patients with stage II and stage IIIA NSCLC have the greatest unmet need, as their baseline DFS and OS rates are poorer than for patients with stage IB disease. Higher-risk individuals within the stage IB population (i.e., those with larger tumour sizes, perineural or lymphovascular invasion, or disease spread through air spaces) may also have an unmet need, even with adjuvant chemotherapy, and LCC believes that this subset of stage IB patients could benefit more from osimertinib compared with patients with stage IB disease without high-risk characteristics.

Place in Therapy

For patients with stage IB to stage II NSCLC, the standard treatment is complete surgical resection. A minority of patients with stage IB disease also have pathological findings indicating a high risk of relapse (e.g., larger tumours, lymphovascular or perineural invasion, or disease spread through air spaces) are offered adjuvant platinum-doublet chemotherapy. For patients with stage IIIA disease, if surgery is considered reasonable, neoadjuvant chemotherapy concurrent with radiation followed by complete surgical resection is typically offered. The OH-CCO’s L-DAC reported that adjuvant chemotherapy is routinely offered to patients with resected stage IIA to stage IIIB NSCLC, per the AJCC 8th edition. The clinician group added that patient uptake of adjuvant chemotherapy in stage IIA is lower than in stage IIB and stage III disease, and that patients with EGFR-mutated resected NSCLC are also offered adjuvant chemotherapy. The OH-CCO’s L-DAC commented that osimertinib would represent an additional treatment option for patients after adjuvant chemotherapy and those who do not receive adjuvant chemotherapy.

LCC thought that osimertinib should be added to the standard post-operative management of patients with resected stage IB to stage IIIA NSCLC disease carrying an eligible EGFR mutation. Both clinician groups thought that osimertinib should not be considered a replacement for adjuvant chemotherapy and that osimertinib would be used post adjuvant chemotherapy. LCC thought that adjuvant chemotherapy should be administered independent of the consideration of osimertinib, where appropriate.

LCC commented that 3 years of administration of an oral drug could change the current treatment paradigm because the only adjuvant treatment currently available is adjuvant chemotherapy, which is completed over approximately 3 months (4 treatment cycles). LCC indicated the impact on health care utilization might be modest because osimertinib is a home-based oral, low-toxicity drug.

With respect to sequencing, the OH-CCO’s L-DAC noted that osimertinib is currently used as first-line therapy for advanced or metastatic NSCLC. The clinician group noted that if patients received adjuvant osimertinib and relapse on treatment, they would not receive osimertinib for advanced or metastatic disease. For patients who complete 3 years of adjuvant osimertinib and relapse after a 6-month off-treatment interval, the clinician group thought these patients would be considered for re-treatment with osimertinib. Additionally, the clinician group noted that for patients who exhibit oligoprogression or, in some cases, “flare” (such as bone metastases after therapy) while on osimertinib, osimertinib would be expected to be continued.

Patient Population

The OH-CCO’s L-DAC noted that patients with resected stage IIA to stage IIIB NSCLC with exon 19 deletions or exon 21 (L858R) substitution EGFR mutations would be candidates for adjuvant osimertinib. LCC indicated that patients with a higher disease stage are more likely to benefit from adjuvant osimertinib, although patients with stage IB disease may also benefit.

Both clinician groups agreed that patients who do not carry EGFR mutations or who do not have surgically resected NSCLC would be least suitable for treatment with osimertinib. The OH-CCO’s L-DAC added that osimertinib is not indicated in patients with other types of EGFR mutations (i.e., who do not have exon 19 deletions or L858R-point mutation) or those with wild-type EGFR tumours.

Assessing Response to Treatment

The OH-CCO’s L-DAC indicated that a clinically meaningful response to adjuvant osimertinib therapy would be the absence of disease recurrence. Similarly, LCC indicated that DFS and OS are used to determine whether a patient is benefiting from adjuvant osimertinib in clinical practice.

Since adjuvant osimertinib therapy would be administered for 3 years, LCC thought that periodic follow-up for toxicity and disease recurrence is required. LCC commented that patients may have more frequent follow-up appointments (monthly to every 2 months) near the initiation of treatment, then less frequent follow-ups for up to every 6 months for the remainder of the treatment period. LCC noted that imaging scans at 3-month to 4-month intervals would be common near the beginning of the treatment period, followed by scans every 6 months or less frequently toward the end of adjuvant therapy. Similarly, the OH-CCO’s L-DAC thought that response to treatment should be assessed every 3 months near the initiation of therapy, and CT scans should be conducted at least every 6 months during the 3 years of osimertinib therapy to monitor for disease recurrence. LCC noted that patients with resected stage IB to stage IIIA NSCLC are generally followed up for at least 5 years postoperatively in current Canadian practice. LCC reported that with current treatments, recurrences of stage IB to stage IIIA NSCLC occur for most patients within 2 years to 3 years and OS typically requires a greater number of years of follow-up.

Discontinuing Treatment

Both clinician groups agreed the primary reason for drug discontinuation would be disease recurrence, intolerable toxicity, or completion of 3 years of adjuvant therapy.

Prescribing Conditions

The OH-CCO’s L-DAC indicated that adjuvant osimertinib therapy would be administered in the outpatient setting under the supervision of an oncologist. LCC indicated that adjuvant osimertinib is suitable in all oncology settings and is appropriate for treatment in the community setting, including in medical oncology outpatient clinics and the inpatient setting.

Additional Considerations

The OH-CCO’s L-DAC indicated that EGFR mutation testing should be reflexively performed on all patients newly diagnosed with non-squamous NSCLC so those with early-stage NSCLC who have eligible EGFR mutations might be readily identified. The clinician groups noted that an EGFR genetic testing infrastructure for patients with advanced or metastatic NSCLC has been set up across Canada. As a result, these groups think there will be no major infrastructure gaps when testing is expanded in the adjuvant setting, and there may be minimal-to-modest increases in costs. LCC noted that patients with other sensitizing EGFR mutations that are rarer than exon 19 deletions or exon 21 (L858R) substitution mutations may also benefit from osimertinib.

The LCC clinician group estimates that the total number of patients who will qualify for adjuvant osimertinib will range from 300 patients to 400 patients per year in the first few years, if osimertinib is reimbursed. The OH-CCO’s L-DAC reported that osimertinib is a generally well-tolerated therapy. LCC indicated that delaying or reducing disease recurrence has benefits to HRQoL, utilization of health care resources, economic loss of productivity, and overall costs to society.

The clinician groups indicated that re-treatment with osimertinib in the metastatic setting after completing adjuvant osimertinib therapy is a consideration, but data are not yet available to inform re-treatment. LCC suggested that with the lack of available data on re-treatment with osimertinib, it may be reasonable to consider allowing clinicians to re-treat with osimertinib at their discretion if they believe it will provide clinical benefit.

Drug Program Input

The drug programs provide input on each drug being reviewed through CADTH’s reimbursement review processes by identifying issues that may impact their ability to implement a recommendation. The implementation questions and corresponding responses from the clinical experts consulted by CADTH are summarized in Table 4.

The drug programs noted that osimertinib has the potential for drug–drug interactions, which could potentially increase pharmacy resource use. The drug programs also indicated that osimertinib adjuvant therapy may change the place in therapy of comparator drugs and drugs reimbursed in subsequent lines. The drug plans reported that EGFR genetic testing is not reflexively completed for early-stage NSCLC across most Canadian jurisdictions; thus, expansion of testing would be required to identify eligible patients. Lastly, the drug programs expressed concern that the budget impact may be substantial because the duration of therapy per patient is 3 years.

Clinical Evidence

The clinical evidence included in the review of osimertinib is presented in 3 sections. The first section, the systematic review, includes pivotal studies provided in the sponsor’s submission to CADTH and Health Canada, as well as those studies that were selected according to an a priori protocol. No indirect evidence, long-term extension studies, or other relevant studies were identified.

Systematic Review (Pivotal and Protocol-Selected Studies)

Objectives

To perform a systematic review of the beneficial and harmful effects of osimertinib (80 mg dose, oral) as adjuvant therapy after tumour resection in patients with stage IB to stage IIIA NSCLC whose tumours have EGFR exon 19 deletions or exon 21 (L858R) substitution mutations.

Methods

Studies selected for inclusion in the systematic review included pivotal studies provided in the sponsor’s submission to CADTH and Health Canada, as well as those meeting the selection criteria presented in Table 5. Outcomes included in the CADTH review protocol reflect outcomes considered to be important to patients, clinicians, and drug plans.

The literature search for clinical studies was performed by an information specialist using a peer-reviewed search strategy according to the PRESS Peer Review of Electronic Search Strategies checklist.¹⁹

Published literature was identified by searching the following bibliographic databases: MEDLINE All (1946‒) through Ovid and Embase (1974‒) through Ovid. The search strategy comprised both controlled vocabulary, such as the National Library of Medicine’s MeSH (Medical Subject Headings), and keywords. The main search concepts were Tagrisso (osimertinib) and NSCLC. Clinical trials registries were searched: the US National Institutes of Health’s clinicaltrials.gov, the WHO International Clinical Trials Registry Platform (ICTRP) search portal, Health Canada’s Clinical Trials Database, and the European Union Clinical Trials Register.

Search filters were applied to limit retrieval to RCTs or controlled clinical trials. Retrieval was not limited by publication date or by language. Conference abstracts were excluded from the search results. See Appendix 1 for the detailed search strategies.

The initial search was completed on April 19, 2021. Regular alerts updated the search until the meeting of the CADTH pan-Canadian Oncology Drug Expert Review Committee (pERC) on August 11, 2021.

Grey literature (literature that is not commercially published) was identified by searching relevant websites from the Grey Matters: A Practical Tool For Searching Health-Related Grey Literature checklist.²⁰ Included in this search were the websites of regulatory agencies (the FDA and European Medicines Agency). Google was used to search for additional internet-based materials. See Appendix 1 for more information on the grey literature search strategy.

In addition, the manufacturer of the drug was contacted for information regarding unpublished studies.

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

A focused literature search for network meta-analyses dealing with NSCLC was run in MEDLINE All (1946–) on April 16, 2021. No search limits were applied.

Findings From the Literature

A total of 5 reports presenting data from 1 unique study were identified from the literature for inclusion in the systematic review (Figure 1). The included studies are summarized in Table 6.

Figure 1: Flow Diagram for Inclusion and Exclusion of Studies

Description of Studies

The ADAURA trial is an ongoing, international, multi-centre, phase III, double-blind, randomized, placebo-controlled trial.¹³ The objective of the trial is to assess the efficacy and safety of osimertinib versus placebo in patients with stage IB to stage IIIA NSCLC with a centrally confirmed, common-sensitizing EGFR mutation (exon 19 deletions and/or exon 21 [L858R] substitution mutations, either alone or in combination with other EGFR mutations) who have undergone complete tumour resection, with or without post-operative adjuvant chemotherapy. Patients were staged according to the 7th edition of the AJCC cancer staging manual. A total of 682 patients were randomized in a 1:1 ratio to osimertinib (80 mg oral, once daily) or matching placebo: 339 patients to the osimertinib arm and 343 to the placebo arm. Randomization was stratified by disease stage (stage IB versus II versus IIIA), mutation type (exon 19 deletions or L858R), and race (Asian or non-Asian). Patients were randomized from 185 sites in 24 countries in Europe, Asia-Pacific, North America, and South America. One site in Canada enrolled patients into the ADAURA trial. The study design of the ADAURA trial is summarized in Figure 2.

Figure 2: ADAURA Study Design

AZD9291 = osimertinib; EGFR = epidermal growth factor receptor; Exdel19 = exon 19 deletions; L858R = exon 21 L858R substitution mutation; NSCLC = non–small cell lung cancer; w = week.

Source: Clinical Study Report.¹³

To be eligible for the ADAURA trial, patients must have sufficiently recovered from surgery and completed any standard-of-care adjuvant chemotherapy (if applicable) before randomization. Patients must have been randomized within 10 weeks of complete surgical resection if adjuvant chemotherapy was not administered, or within 26 weeks if adjuvant chemotherapy (platinum-based doublet chemotherapy, maximum of 4 cycles) was administered. Patients were required to have a baseline CT scan (chest and abdomen, including liver and adrenal glands) following surgical resection and within 28 days before treatment initiation to confirm that disease was not present.

The ADAURA trial employed a 2-part screening period. In part 1, patients provided informed consent in writing for their tumour sample to be sent to the central laboratory for EGFR genetic testing. The EGFR mutation status of patients was prospectively tested by the central cobas EGFR mutation test (Roche Molecular Systems). A separate informed consent form was then required before any other study procedures were done. After patients signed this second informed consent form, eligibility for the trial was confirmed in part 2 of the screening. Patients received study treatment for 3 years or until they experienced disease recurrence or unmanageable toxicity.

The primary end point of the ADAURA trial is DFS by investigator assessment. Radiological assessments for disease recurrence were done at 12 weeks, 24 weeks, every 24 weeks until 5 years, and then annually. In the event of disease recurrence, imaging procedures were undertaken to capture all sites of NSCLC relapse. Treatments received by the patient after relapse were determined by the treating physician, and post-recurrence cancer treatments and procedures were recorded.

The ADAURA trial is ongoing. An interim analysis with a data cut-off date of January 17, 2020, was conducted. As of the interim analysis, a total of 3 protocol amendments had been implemented during the ADAURA trial. The first amendment (December 18, 2018) consisted of updates to provide clarity on study procedures and additional information. The second amendment (April 28, 2020) included an OS extension period and a new multiple-testing procedure to account for 2 unplanned interim analyses of DFS. This amendment also added an analysis of time to new brain lesions or death. The third amendment (June 23, 2020) was to update the multiple-testing procedure again for the testing of OS.

Populations

Inclusion and Exclusion Criteria

The ADAURA trial enrolled adult patients aged 18 years and older (20 years and older in Japan and Taiwan) with a diagnosis of NSCLC of predominately non-squamous histology and a WHO performance status of 0 or 1.¹³ Patients must have had complete surgical resection of the primary NSCLC and their cancer must have been postoperatively classified as stage IB, II, or IIIA (AJCC 7th edition). Confirmation by the central laboratory that the tumour harboured 1 of the 2 common EGFR mutations associated with EGFR TKI sensitivity (exon 19 deletions or L858R), either alone or in combination with other EGFR mutations, was required. Patients must have sufficiently recovered from surgery and completed adjuvant chemotherapy (if applicable) before randomization. Standard post-operative adjuvant chemotherapy consisting of a platinum-based doublet for 4 or fewer cycles was allowed. No more than 26 weeks or 10 weeks could elapse between surgery and randomization for patients who had or had not received adjuvant chemotherapy, respectively. Patients were excluded from the ADAURA trial if they had previous treatment with pre- or post-operative radiation therapy, preoperative chemotherapy, EGFR TKIs, and/or potent cytochrome P450 3A4 (CYP3A4) inducers.

Baseline Characteristics

The baseline characteristics of patients were balanced between treatment arms. In the overall population, the mean age of the patients was 62.1 years.¹³ The majority of patients had stage II to stage IIIA disease (68.3%), a WHO performance status of 0 (63.6%), adenocarcinoma histology type (96.5%), undergone a lobectomy (95.3%), and were Asian (63.6%) and female (70.1%). Most patients had received post-operative adjuvant chemotherapy (60.0%). Overall, 54.7% of patients had exon 19 deletions and 45.2% had exon 21 (L858R) substitution mutations.

Interventions

Patients were randomized using an interactive voice response system to receive either osimertinib 80 mg orally once daily or matching placebo.¹³ Doses were to be taken approximately 24 hours apart and at the same time each day. Treatment continued until disease recurrence, a treatment discontinuation criterion was met (e.g., patient decision, unacceptable toxicity, AE, pregnancy), or until a maximum treatment duration of 3 years was reached.

Other anti-cancer therapies, investigational drugs, and radiotherapy were not permitted while the patient was receiving the study treatment. Pre-medication (including treatments for the management of diarrhea, nausea, and vomiting) was allowed after the first dose of the study drug. Other medication considered necessary for the safety and well-being of a patient could be given at the discretion of the investigator.

Outcomes

A list of the efficacy end points identified in the CADTH review protocol that were assessed in the ADAURA trial and included in this review is provided in Table 8.

A detailed discussion and critical appraisal of DFS as a surrogate outcome measure for OS is provided in Appendix 3.

In the ADAURA trial, HRQoL was measured using the SF-36 v2 questionnaire. Specifically, the trial reported the TTD in the MCS and PCS of the SF-36 v2. The SF-36 v2 is an instrument for measuring a person’s general health status over the past 28 days. The SF-36 v2 includes 8 domains: Physical Functioning, Role-Physical, Vitality, General Health, Bodily Pain, Social Function, Role-Emotional, and Mental Health. SF-36 v2 results can be grouped into 2 summary scores, the PCS and the MCS. Final scores for each scale range from 0 to 100, with higher scores indicating better health. The minimal important differences (MIDs) for this instrument have not been established in patients with NSCLC. The MIDs from the general population are 2 points for the PCS and 3 points for the MCS. A detailed discussion and critical appraisal of the SF-36 v2 is provided in Appendix 3.

Statistical Analysis

Sample Size and Power Calculation

It was planned that approximately 700 patients would be randomized in a 1:1 ratio (osimertinib to placebo) to the ADAURA study.¹³ It was estimated that 3,200 patients would be screened to randomize 700 patients. The primary end point of the study is DFS based on investigator assessment and was assessed first in patients with stage II to stage IIIA disease (primary analysis population) and then the overall population (i.e., patients with stage IB to stage IIIA).

The ADAURA trial is event-driven. The primary analysis of DFS was planned to occur when approximately 247 disease recurrence events had been observed in approximately 490 patients in the stage IIA to stage IIIA patient population. If the true DFS HR for the comparison of osimertinib versus placebo in this patient population is 0.70, 247 disease recurrence events would provide 80% power to demonstrate a statistically significant difference in DFS at a 5% 2-sided significance level (translating to an improvement in median DFS from 40 months to 57 months, assuming DFS is exponentially distributed). Based on these assumptions, the minimum DFS HR that would be considered statistically significant (2-sided P < 0.05) is 0.78.

If the true DFS HR for the comparison of osimertinib versus placebo in the overall population (i.e., patients with stage IB to stage IIIA disease) is 0.70, then 317 disease recurrence events would provide approximately 90% power to demonstrate a statistically significant difference in DFS at a 4% 2-sided significance level (translating to an improvement in median DFS from 46 months to 66 months, assuming DFS is exponentially distributed). Based on these assumptions, the minimum DFS HR that would be considered statistically significant (2-sided P < 0.04) is 0.79.

At the time of the planned primary analysis of DFS, it was anticipated that approximately 195 OS events (28% maturity) would have occurred (assuming a median OS of 96 months for the placebo arm). If the true OS HR for the comparison of osimertinib versus placebo is 0.66, then 195 death events would provide approximately 80% power to demonstrate a statistically significant difference in OS between the treatment arms. Assuming 195 death events and a true OS HR of 0.85, there would be an approximately 90% chance of observing an HR of less than 1.02.

Assuming 28 months of nonlinear recruitment, the data cut-off for the primary analysis was estimated to occur approximately 68 months to 70 months after the first patient was randomized.¹²

Analyses, Multiple-Testing Procedure, and Alpha Spending

The primary end point of DFS and the secondary end point of OS were tested in the stage II to stage IIIA population (primary analysis population) and in the overall population (i.e., patients with stage IB to stage IIIA disease).¹³ To control the type I error at the 5% 2-sided level, a hierarchical multiple-testing procedure was employed across these end points (Figure 3). This procedure was added as an amendment to the study protocol and statistical analysis plan to account for the unplanned interim analysis. The multiple-testing procedure was ordered such that DFS in patients with stage II to stage IIIA disease was tested first using the full alpha. If statistical significance was shown for DFS in the stage II to stage IIIA population, DFS in the overall population would be tested. If the DFS results in both the stage II to stage IIIA population and the overall population were statistically significant, OS would then be assessed first in the stage II to stage IIIA population, and then in the overall population if statistical significance was reached in the stage II to stage IIIA population.

Figure 3: Multiple-Testing Procedure in the ADAURA Trial

DFS = disease-free survival; OS = overall survival.

Source: ADAURA study protocol.¹²

In the original ADAURA trial protocol, a single, primary analysis of DFS was planned; a further analysis was to be performed only if there were fewer than 70 DFS events in the stage IB population. For the overall population analysis, it was estimated there would be approximately 317 DFS events in approximately 700 patients, with approximately 70 events in the stage IB subgroup. If, however, there were meaningfully fewer than 70 DFS events (defined as 63 DFS events or fewer) in the stage IB subgroup at the time of the primary analysis, additional follow-up of patients could be performed and a further analysis of DFS could be conducted when at least 70 DFS events were observed in the stage IB subgroup.

Regular independent data monitoring committee (IDMC) meetings took place during the study to monitor safety data, and a scheduled event-based futility analysis took place to support the sixth IDMC meeting in February 2019. After ruling out futility, the IDMC made an ad hoc request to evaluate key efficacy data (Kaplan-Meier curves for DFS and key recurrence data tables) at their seventh meeting, which was held on April 7, 2020, with a data cut-off date of January 17, 2020. Following the seventh IDMC meeting, the IDMC recommended that a full analysis of efficacy and safety be performed by the sponsor as soon as possible for public disclosure. The sponsor conducted the interim analysis, and the sponsor was unblinded at the time of this analysis. Patients and investigators remain blinded to individual treatment allocations, and the study is ongoing.

The unplanned interim analysis of DFS was conducted at the time of observing 156 DFS events in the stage II to stage IIIA population. The corresponding information fraction was 0.63, where the final number of events would have been 247. The interim analysis needed to be taken into account to control the type I error rate at 5% (2-sided) and the multiple-testing procedure was revised to the procedure depicted in Figure 3. The Lan-DeMets approach that approximates the O’Brien and Fleming spending function was used to maintain an overall 2-sided 5% type I error.

The ADAURA trial is ongoing. The sponsor plans to conduct the final analysis of OS when approximately 94 deaths have been observed in the stage II to stage IIIA population (approximately 20% maturity).

Analysis of Outcomes

All efficacy analyses were performed using the full analysis set (FAS), which included all randomized patients. The FAS included 2 populations: the overall population (i.e., patients with stage IB to stage IIIA disease) and the stage II to stage IIIA population. The primary analysis population is the stage II to stage IIIA population. Results of all statistical analyses were presented using a 95% CI and 2-sided P value.

The statistical analyses performed of the efficacy end points that were identified in the CADTH systematic review protocol are summarized in Table 9.

Primary Outcome

DFS was calculated as the time from the date of randomization until the date of disease recurrence or death by any cause in the absence of recurrence. Patients who were disease-free and alive at the time of the analysis were censored at the date of their last follow-up assessment. The DFS time was based on the scan (assessment) dates and not visit dates. The DFS rate at 2, 3, 4, and 5 years was defined as the proportion of patients alive and disease-free at 2, 3, 4, and 5 years, respectively, estimated from Kaplan-Meier plots of the primary end point of DFS at the time of the primary analysis.

DFS in the subset of patients with stage II to stage IIIA disease was analyzed using a log-rank test stratified by stage (II, IIIA), mutation type (exon 19 deletions, L858R), and race (Asian, non-Asian) for the generation of the P value and using the Breslow approach for handling ties. DFS in the overall population was analyzed using a log-rank test stratified by stage (IB, II, IIIA), mutation status (exon 19 deletions or L858R as confirmed by a central test) and race (Asian, non-Asian) for the generation of the P value and using the Breslow approach for handling ties. The HRs and CIs were obtained directly from the U- and V-statistics provided by the log-rank test and the LIFETEST procedure. Proportionality was assessed for the HR.

A Kaplan-Meier plot of DFS was presented by treatment arm.

Subgroup analyses were conducted by comparing DFS between treatment arms in the following groups of patients: stage (IB, II, IIIA), EGFR mutation status (positive or negative for exon 19 deletions or L858R), EGFR mutation type (exon 19 deletions, L858R) mutation detectable in plasma-derived circulating tumour DNA (ctDNA), pre-treatment T790M mutation status (positive, negative), race (Asian, non-Asian), adjuvant chemotherapy (yes, no), gender (male, female), age at randomization (< 65, ≥ 65), and smoking history (never, ever). No adjustment was made to the significance level for statistical testing since the subgroup analyses may only be supportive of the primary analysis of DFS. For each subgroup level, the HR and 95% CI were calculated from a single Cox proportional-hazards model that contained a term for treatment, the subgroup covariate of interest, and the treatment-by-subgroup interaction term. The HR was obtained from this model for each level of the subgroup.

DFS rates were derived using the same model used for the primary analysis of DFS. Kaplan-Meier plots were produced. Estimates of the DFS rate at 6 months, 12 months, 18 months, 24 months, 36 months, 48 months, and 60 months were obtained from the Kaplan-Meier plot and presented for each treatment arm.

Sensitivity analyses to assess for quantitative interactions, evaluation-time bias, and attrition bias were planned.¹⁴ The presence of quantitative interactions was assessed by means of an overall global interaction test. To assess possible evaluation-time bias that could occur if scans are not performed at the protocol-scheduled time points, the midpoint between the time of recurrence and the previous evaluable assessment was analyzed using a log-rank test stratified by stage, mutation status, and race. Possible attrition bias was assessed by repeating the primary DFS analysis such that the actual DFS times were included in the analysis, rather than the censored times of patients who had experienced recurrence or died in the absence of recurrence immediately following 2 or more non-evaluable assessments. In addition, patients who took subsequent therapy before recurrence or death were censored at their last evaluable assessment before taking the subsequent therapy.

DFS with disease recurrence in the CNS only was analyzed as a post hoc exploratory analysis.

The disease recurrence rates in the stage II to stage IIIA population and overall population were reported within the reporting of DFS events.

Secondary Outcomes

Overall Survival

OS was calculated as the time from the date of randomization until date of death due to any cause. Any patient not known to have died at the time of analysis was censored based on the last recorded date on which the patient was known to be alive. OS data were analyzed using the same methodology and model as for the analysis of DFS with the exception that the sensitivity, subgroup, and exploratory analyses were not performed.

A Kaplan-Meier plot of OS was presented by treatment group. The OS rate at 2, 3, 4, and 5 years was defined as the proportion of patients alive at 2, 3, 4, and 5 years, respectively, estimated from the Kaplan-Meier plot of OS.

HRQoL Assessed Using the SF-36

Patient-reported HRQoL was assessed using the SF-36 v2 questionnaire. The compliance with the SF-36 was summarized by visit. Compliance rates were calculated as the number of evaluable forms (i.e., a questionnaire with a completion date and at least 1 domain that is non-missing) divided by the number of expected forms (i.e., a questionnaire from a patient who has not died, is not lost to follow-up, or has not withdrawn from the study at the scheduled assessment time) multiplied by 100.

The absolute values and change from baseline were calculated for each of the 8 domain scores and for the PCS and MCS summary scores at each post-baseline assessment. The responses to each of the health domain scores and the 2 summary scores were also categorized as improved, worsened, or stable at each post-baseline assessment and reported as counts and proportions.

The primary HRQoL outcome measures were TTD of the 2 summary scores (PCS and MCS) in the stage II to stage IIIA population (pre-specified secondary end points). The TTD analyses in the stage IIA to stage IIIA population were analyzed using a log-rank test stratified by stage (II, IIIA), mutation type (exon 19 deletions, L858R) and race (Asian, non-Asian) using the Breslow approach for handling ties. The HRQoL end points were not included in the multiple testing procedures and thus these analyses were not controlled for multiple-comparison testing.

The TTD in PCS and MCS were analyzed in the overall population as a post hoc exploratory analysis.

Exploratory Outcomes

Progression-Free Survival

PFS was calculated as the time from the date of randomization to the date of disease progression or death. Patients alive and for whom disease progression had not been observed were censored at the last time known to be alive and without disease progression. Progression status at the time of analysis by treatment arm was categorized into radiological, symptomatic, other progression, or death, with reasons for censoring also summarized accordingly. PFS data were analyzed for the overall population using the same methodology and model as for the analysis of DFS with the exception that the sensitivity, subgroup, and exploratory analyses were not performed.

Time to Next Treatment

Time to next treatment was defined as the time from the date of randomization to the earlier of the date of first subsequent anti-cancer therapy or procedure start date following study drug discontinuation, or death. The time to next treatment was calculated for the overall population using the same methodology and model as for the analysis of DFS with the exception that the sensitivity, subgroup, and exploratory analyses were not performed.

Analysis Populations

The FAS included all randomized patients and was used for all efficacy analyses. The FAS included 2 populations: the overall population (i.e., patients with stage IB to stage IIIA disease) and stage II to stage IIIA population. The stage II to stage IIIA population is considered the primary analysis population in the ADAURA trial. Treatment groups were compared based on randomized study treatment, regardless of the treatment received.

The safety analysis set consisted of all patients who received at least 1 dose of the study treatment.

Results

The results from the ADAURA trial presented in this report are from the unplanned interim analysis with a data cut-off date of January 17, 2020.

Patient Disposition

A summary of the disposition of patients in the ADAURA trial as of the interim analysis is included in Table 10. A total of 2,447 patients were screened for EGFR mutations. A total of 1,250 patients (51.1%) were EGFR mutation–negative; for 110 patients (4.5%), the EGFR status was not evaluable. A total of 239 patients (9.8%) did not sign the main informed consent form for further screening of eligibility (screening part 2). The reasons why patients did not sign the informed consent form were not reported. A total of 791 patients were identified as having EGFR mutation–positive disease and entered screening part 2, of which 109 (13.8%) were not randomized: 106 (97.2%) due to not meeting the eligibility criteria, and 3 (2.8%) due to patient decision. A total of 682 were randomized to receive osimertinib (n = 339) or placebo (n = 343). A total of 680 patients (99.7%) were treated: 337 patients (99.4%) in the osimertinib arm, and all 343 patients (100%) in the placebo arm.

As of the data cut-off for the interim analysis, 91.2% of patients in the osimertinib arm and 89.5% of patients in the placebo arm were ongoing in the ADAURA trial. A total of 92 patients (27.3%) in the osimertinib arm and 174 patients (50.7%) in the placebo arm had discontinued study treatment. The most common reasons for treatment discontinuation in the osimertinib arm were AEs (n = 36; 10.7%), patient decision (n = 30; 8.9%), and disease recurrence (n = 24; 7.1%). The most common reason for treatment discontinuation in the placebo arm was disease recurrence (n = 148; 43.1%).

Exposure to Study Treatments

Exposure to study treatments as of the interim analysis is summarized in Table 11. Patients in the osimertinib arm had a longer median total exposure time to the study drug (22.5 months versus 18.7 months). As of the data cut-off date, 40 patients (12%) in the osimertinib arm and 33 patients (10%) in the placebo arm had completed 3 years of treatment.²²

The dose interruptions and reductions that occurred in the ADAURA trial by treatment arm are summarized in Table 12. Overall, 54.3% of patients in the osimertinib arm and 41.7% of patients in the placebo arm had a dose interruption for any reason during the study, with a median duration of 8.0 days and 5.0 days, respectively. The most common reasons for dose interruptions were patients forgetting to take a dose (29.7% in the osimertinib arm, 31.5% in the placebo arm) and patients experiencing an AE (31.5% in the osimertinib arm, 10.5% in the placebo arm). The most common AEs leading to treatment interruption in the osimertinib arm were diarrhea (n = 13; 3.9%) and stomatitis (n = 8; 2.4%). In the placebo arm, the most common AEs leading to study treatment interruption were diarrhea (n = 4; 1.2%) and vomiting (n = 3; 0.9%).

In the osimertinib arm, 14.5% of patients had a dose reduction compared with 0.9% of patients in the placebo arm. The most common reason for dose reduction was an AE (13.6% in the osimertinib arm, 0.9% in the placebo arm). AEs leading to dose reductions in 1 or more patients in the osimertinib arm were stomatitis (n = 5; 1.5%), paronychia (n = 4; 1.2%), hypertension (n = 2; 0.6%), diarrhea (n = 2; 0.6%), nausea (n = 2; 0.6%), and QT interval prolonged (n = 2; 0.6%). No AE leading to a dose reduction was reported by more than 1 patient in the placebo arm.

Efficacy

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

Overall Survival

The OS results from the ADAURA trial as of the interim analysis are summarized in Table 13. As of the data cut-off date, the sponsor considered the OS data to be immature. A total of 29 patients (4.3%) had died: 9 (2.7%) in the osimertinib arm and 20 (5.8%) in the placebo arm. Most patients were in survival follow-up (616 patients; 90.3%). This included 309 patients (91.2%) in the osimertinib arm and 307 patients (89.5%) in the placebo arm. The median follow-up for OS in the stage II to stage IIIA population was 26.1 months and 24.6 months in the osimertinib and placebo arms, respectively; the median follow-up for OS in the overall population was 26.1 months and 25.9 months in the osimertinib and placebo arms, respectively.

Per the multiple-testing procedure, OS was formally tested in patients with stage II to stage IIIA disease at the interim analysis. At the data cut-off, 25 deaths had occurred in the stage II to stage IIIA population (5.3% maturity of data), comprising 8 deaths (3.4%) in the osimertinib arm and 17 deaths (7.2%) in the placebo arm. The Kaplan-Meier plots of OS in the stage II to stage IIIA population are depicted in Figure 4. The HR was 0.40 (95% CI, 0.18 to 0.89; P = 0.0244), which did not reach statistical significance (P value < 0.0002 required).

Since OS did not reach statistical significance in the primary stage II to stage IIIA population, OS in the overall population was not formally tested for statistical significance per the multiple-testing procedure. The analysis of OS in the overall population should therefore be regarded as an exploratory analysis. The Kaplan-Meier plots of OS in the overall population are depicted in Figure 5. The HR was 0.48 (95% CI, 0.23 to 1.02). In the overall population, 9 patients (2.7%) in the osimertinib arm and 20 patients (5.8%) in the placebo arm had died.

Figure 4: Kaplan-Meier Plot of OS in the ADAURA Trial — FAS, Stage II to Stage IIIA Population

AZD9291 = osimertinib; FAS = full analysis set; OS = overall survival.

Source: Clinical Study Report.¹³

Figure 5: Kaplan-Meier Plot of OS in the ADAURA Trial — FAS, Overall Population

AZD9291 = osimertinib; FAS = full analysis set; OS = overall survival.

Source: Clinical Study Report.¹³

Disease-Free Survival

The DFS results from the ADAURA trial as of the interim analysis are summarized in Table 14. As of the data cut-off date, median follow-up time for DFS in the stage II to stage IIIA population was 22.1 months in the osimertinib arm and 14.9 months in the placebo arm. The median follow-up for DFS in the overall population was 22.1 months in the osimertinib arm compared with 16.6 months in the placebo arm.

In the stage II to stage IIIA population, 26 patients (11.2%) in the osimertinib arm and 130 patients (54.9%) in the placebo arm had experienced a DFS event. The Kaplan-Meier plots of DFS in the stage II to stage IIIA population are depicted in Figure 6. The HR was 0.17 (95% CI, 0.12 to 0.23), which was statistically significant (P < 0.0001). The Kaplan-Meier estimate of the median duration of DFS was not reached (NR) in the osimertinib arm (95% CI, 38.8 months to NR) compared with 19.6 months (95% CI, 16.6 to 24.5) in the placebo arm.

In the overall population, 37 patients (10.9%) in the osimertinib arm and 159 patients (46.6%) in the placebo arm had experienced a DFS event. The Kaplan-Meier plots of DFS in the overall population are depicted in Figure 7. The HR was 0.20 (95% CI, 0.15 to 0.27), which was statistically significant (P < 0.0001). The Kaplan-Meier estimate of the median duration of DFS was NR in the osimertinib arm (95% CI, NR to NR) compared with 27.5 months (95% CI, 22.0 to 35.0) in the placebo arm.

Figure 6: Kaplan-Meier Plot of DFS in the ADAURA Trial — FAS, Stage II to Stage IIIA Population

AZD9291 = osimertinib; DFS = disease-free survival; FAS = full analysis set.

Source: Clinical Study Report.¹³

Figure 7: Kaplan-Meier Plot of DFS in the ADAURA Trial — FAS, Overall Population

AZD9291 = osimertinib; DFS = disease-free survival; FAS = full analysis set.

Source: Clinical Study Report.¹³

Sensitivity analyses were performed in the stage II to stage IIIA population to assess evaluation-time bias, attrition bias, and the presence of quantitative interactions, and the results of these analyses are summarized in Appendix 2. There was no evidence of evaluation-time bias (Table 25) or attrition bias (Figure 15). The presence of quantitative interactions was assessed using an overall global interaction test, and the results of these tests are summarized in Table 26. These analyses indicated there was evidence of a quantitative interaction in EGFR mutation type (exon 19 deletions versus L858R) on DFS (P = 0.0132), suggesting that osimertinib showed a treatment benefit in both exon 19 deletions and L858R mutation subgroups, but with a difference in magnitude (significance level of 0.1).

The results of the pre-specified subgroup analyses of DFS are summarized in Table 15. Consistent with the FAS population, a benefit with osimertinib was consistently observed with an HR of less than 0.4 for all subgroups.

A post hoc exploratory analysis of DFS with disease recurrence in the CNS only (determined by investigator assessment) was performed in the FAS, and the results are summarized in Table 16. In the overall population, 45 patients experienced disease recurrence in the CNS or death, with most events occurring in patients with stage II to stage IIIA disease (36 patients). Among all patients randomized to the osimertinib arm (overall population), 4 patients (1.2%) had recurrence in the CNS, corresponding to 10.8% of all recurrence events in this arm. In the placebo arm, 33 patients (9.6%) had recurrence in the CNS, corresponding to 21% of all recurrence events in this arm. The results suggested a benefit with osimertinib compared with placebo in CNS DFS, with an HR of 0.14 (95% CI, 0.07 to 0.27) for stage II to stage IIIA patients and an HR of 0.18 (95% CI, 0.10 to 0.33) for the overall population. In the stage II to stage IIIA population, the median CNS DFS was NR (95% CI, 38.8 to NR) and NR (95% CI, NR to NR) in the osimertinib and placebo arms, respectively. In the overall population, the median CNS DFS was NR in the osimertinib arm versus 48.2 months (95% CI, NR to NR) in the placebo arm.

Figure 8: Kaplan-Meier Plot of CNS DFS in the ADAURA Trial — FAS, Stage II to Stage IIIA Population

AZD9291 = osimertinib; CNS = central nervous system; DFS = disease-free survival; FAS = full analysis set.

Source: Clinical Study Report.¹³

Figure 9: Kaplan-Meier Plot of CNS DFS in the ADAURA Trial — FAS, Overall Population

AZD9291 = osimertinib; CNS = central nervous system; DFS = disease-free survival; FAS = full analysis set.

Source: Clinical Study Report.¹³

Disease Recurrence Rate

A summary of disease recurrence in the overall population of the ADAURA trial is presented in Table 17. As of the data cut-off, 37 patients (10.9%) in the osimertinib arm and 157 patients (45.8%) in the placebo arm had experienced disease recurrence. In the osimertinib arm, disease recurrence was local or regional only in 23 patients (6.8%), distant in 10 patients (2.9%), and both distant and local or regional in 4 patients (1.2%). In the placebo arm, disease recurrence was local or regional only in 61 patients (17.8%), distant in 78 patients (22.7%), and both distant and local or regional in 18 patients (5.2%). The most common tumour-recurrence locations in the osimertinib and placebo arms were the lung (5.6% and 17.8%, respectively), lymph nodes (2.9% and 14.0%, respectively), CNS (1.5% and 9.9%, respectively), and bone (1.5% and 8.2%, respectively).

Time to Disease Recurrence

Data on the time to disease recurrence were not reported in the ADAURA trial.

Time to Intracranial CNS Metastasis

Data on the time to intracranial CNS metastasis were not reported in the ADAURA trial.

Progression-Free Survival

Data on PFS determined by investigator assessment as of the interim analysis are summarized in Table 18, and the Kaplan-Meier plot is depicted in Figure 10. As of the interim analysis, the sponsor considered the PFS data to be immature. As of the data cut-off date, 13 patients (3.8%) in the osimertinib arm and 46 patients (13.4%) in the placebo arm had experienced a PFS event. The comparisons for this end point were not controlled for multiple comparisons.

Figure 10: Kaplan-Meier Plot of PFS in the ADAURA Trial — FAS, Overall Population

AZD9291 = osimertinib; FAS = full analysis set; PFS = progression-free survival.

Source: Clinical Study Report.¹³

Time to Next Treatment

Data on time to next treatment in the ADAURA trial as of the interim analysis are summarized in Table 19, and the Kaplan-Meier plot is depicted in Figure 11. As of the data cut-off date, the sponsor considered the time to next treatment data to be immature. The comparisons for this end point were not controlled for multiple comparisons.

As of the data cut-off date, 31 patients (9.1%) in the osimertinib arm and 134 patients (39.1%) in the placebo arm had experienced an event (death or first subsequent anti-cancer therapy). Of these events, 30 patients (96.8%) in the osimertinib arm and 125 patients (93.3%) in the placebo arm received a subsequent anti-cancer treatment. The HR was 0.20 (95% CI, 0.14 to 0.27).

The types of subsequent anti-cancer therapy received by patients are summarized in Table 20. The most commonly received type was TKIs (5.0% and 27.4% in the osimertinib and placebo arms, respectively). The most common TKIs were gefitinib (2.7% and 10.5% in the osimertinib and placebo arms, respectively), osimertinib (0.9% and 5.5% in the osimertinib and placebo arms, respectively), afatinib (0.6% and 4.4% in the osimertinib and placebo arms, respectively), and erlotinib (0.6% and 4.4% in the osimertinib and placebo arms, respectively).

Figure 11: Kaplan-Meier Plot of Time to Next Treatment in the ADAURA Trial — FAS, Overall Population

AZD9291 = osimertinib; FAS = full analysis set.

Source: Clinical Study Report.¹³

Health-Related Quality of Life: SF-36 v2

The SF-36 v2 was used to measure HRQoL in the ADAURA trial. This generic instrument includes 8 domains that can be grouped into 2 summary scores, the PCS and MCS. The sponsor defined a clinically relevant change in PCS and MCS as 3.1 points or greater and 3.8 points or greater, respectively.¹² The rationale for these MIDs was not reported by the sponsor. CADTH identified the MID in the general population as 2 points for the PCS score and 3 points for the MCS score (Appendix 3). No MID in patients with NSCLC was identified.

The compliance rates (i.e., the number of evaluable forms divided by the number of expected forms) for the SF-36 over time are depicted in Figure 12. In the overall population, compliance rates for SF-36 completion were greater than 90% in both treatment arms from baseline until week 144, with a reduction to 87.2% and 84.8% at week 156 in the osimertinib and placebo arms, respectively. However, the number of forms expected decreased over time in both treatment arms and included a small proportion of patients at later time points.

Figure 12: Compliance Rates for Completion of the SF-36 v2 in the ADAURA Trial — Overall Population

SF-36 v2 = Short Form (36) Health Survey version 2.

* Compliance rates were calculated as the number of evaluable forms (i.e., a questionnaire with a completion date and at least 1 domain that is non-missing) divided by the number of expected forms (i.e., a questionnaire from a patient who has not died, is not lost to follow-up, or has not withdrawn from the study at the scheduled assessment time) multiplied by 100.

^† Number of expected forms.

Source: Sponsor’s submission to CADTH.¹⁴

In the overall population, the proportion of patients reporting improvements in the PCS of 3.1 points or greater over time increased in both the osimertinib and placebo arms from week 12 (29.9% versus 33.2%) to week 48 (41.0% versus 50.2%), declined transiently at week 72 (38.7% versus 50.0%), and then increased again at week 96 (43.0% versus 53.2%). In both the osimertinib and placebo arms, the proportion of patients reporting improvement in the MCS by 3.8 points or greater increased from week 12 (34.4% versus 41.5%) to week 48 (46.4% versus 49.3%), followed by a trend of decline to week 96 (37.0% versus 44.4%).

A conference abstract included in the sponsor’s submission to CADTH reported a post hoc exploratory analysis of HRQoL that was conducted in the overall population.¹⁴ For this analysis, the SF-36 was scored using norm-based scoring relative to the 2009 US general population (mean ± standard deviation, 50 ± 10), resulting in t scores. Higher t scores indicate better health. The adjusted mean change in the SF-36 PCS and MCS t scores from baseline to week 96 are depicted in Figure 13.

Figure 13: Adjusted Mean Change From Baseline in SF-36 PCS and MCS T Scores Until Week 96 — FAS, Overall Population (Post Hoc Exploratory Analysis)

FAS = full analysis set; MCS = Mental Component Summary; PCS = Physical Component Summary; SD = standard deviation; SF-36 = Short Form (36) Health Survey.

Note: t scores were calculated through a linear transformation of 0 to 100 scores to a metric with a mean of 50 and an SD of 10 in the 2009 US general population. Higher t scores indicate better health.

Source: Sponsor’s submission to CADTH.¹⁴

The pre-specified TTD analyses of PCS and MCS scores were conducted in the stage II to stage IIIA population only, though these analyses were not adjusted for multiple comparisons. For the PCS score, 58 patients (24.9%) in the osimertinib arm experienced a confirmed deterioration of 3.1 points or greater or death compared with 39 patients (16.5%) in the placebo arm (HR = 1.43; 95% CI, 0.96 to 2.13; P = 0.0817). For the MCS score, 52 patients (22.3%) in the osimertinib arm and 52 patients (21.9%) in the placebo arm experienced a confirmed deterioration of 3.8 points or greater or death (HR = 0.90; 95% CI, 0.61 to 1.33; P = 0.5949). The results suggested no difference between groups in the TTD in either summary score.

A conference abstract included in the sponsor’s submission to CADTH included the results of post hoc exploratory analyses of TTD in PCS and MCS in the overall patient population (i.e., stage IB to stage IIIA).¹⁴ The TTD curves are depicted in Figure 14. The results of these analyses were consistent with the pre-specified analyses of TTD in the stage II to stage IIIA population for both the PCS score (HR = 0.98; 95% CI, 0.82 to 1.67) and the MCS score (HR = 0.98; 95% CI, 0.70 to 1.39).

Figure 14: TTD in SF-36 v2 PCS and MCS Scores — FAS, Overall Population (Post Hoc Exploratory Analysis)

FAS = full analysis set; MCS = Mental Component Summary; PCS = Physical Component Summary; SD = standard deviation; SF-36 v2 = Short Form (36) Health Survey version 2; TTD = time to deterioration.

Note: Deterioration was defined as a change from baseline of ≤ −3.1 points for the PCS and ≤ −3.8 points for the MCS.

Source: Sponsor’s submission to CADTH.¹⁴

Harms

Only those harms identified in the CADTH review protocol are reported subsequently. See Table 23 for detailed harms data.

Adverse Events

As of the data cut-off date for the interim analysis, 97.6% of patients in the osimertinib arm and 89.2% of patients in the placebo arm experienced at least 1 treatment-emergent AE (any grade). The most frequently reported AEs in the osimertinib and placebo arms were diarrhea (46.3% and 19.8%, respectively), paronychia (25.2% and 1.5%, respectively), dry skin (23.4% and 6.4%, respectively), pruritis (19.3% and 8.7%, respectively), and cough (18.4% and 16.6%, respectively). Detailed data on AEs reported in 10% or more of patients in either treatment arm by grade is reported in Appendix 2 (Table 27).

Serious Adverse Events

In the ADAURA trial, 16.0% of patients in the osimertinib arm and 12.2% of patients in the placebo arm experienced an SAE as of the interim analysis. The most frequently reported SAEs in the osimertinib and placebo arms were pneumonia (1.5% and 1.2%, respectively), cataracts (0.9% and 0%, respectively), diarrhea (0.6% and 0%, respectively), acute kidney injury (0.6% and 0%, respectively), ureterolithiasis (0.6% and 0%, respectively), and femur fracture (0.6% and 0.3%, respectively).

Withdrawals Due to Adverse Events

Withdrawals from the ADAURA trial specifically due to AEs were not reported. As of the data cut-off for the interim analysis, a total of 33 patients (4.8%) had withdrawn from the trial: 19 (5.6%) in the osimertinib arm and 14 (4.1%) in the placebo arm.

As of the data cut-off date, 10.7% of patients in the osimertinib arm and 2.9% of patients in the placebo arm had discontinued study treatment due to AEs. The most common AEs leading to treatment discontinuation in the osimertinib arm were interstitial lung disease (n = 8; 2.4%), diarrhea (n = 3; 0.9%), and decreased appetite (n = 3; 0.9%). The most common AE leading to treatment discontinuation in the placebo arm was decreased ejection fraction (n = 3; 0.9%).

Mortality

As of the data cut-off date for the interim analysis, 9 patients (2.7%) in the osimertinib arm and 20 patients (5.8%) in the placebo arm had died. The majority of deaths were attributed to NSCLC (9 patients in the osimertinib arm, 18 patients in the placebo arm).

Notable Harms

In the osimertinib arm, 8 patients (2.4%) experienced interstitial lung disease and 2 patients (0.6%) experienced pneumonitis as of the interim analysis. No patients in the placebo arm experienced interstitial lung disease or pneumonitis.

Overall, the frequency of cardiac disorder AEs was greater in the osimertinib arm compared with the placebo arm (11.0% versus 5.2%). Twenty-two patients (6.5%) in the osimertinib arm and 4 patients (1.2%) in the placebo arm experienced QT interval prolongation. Four patients (1.2%) in the osimertinib arm experienced congestive heart failure, cardiac failure, or left ventricular dysfunction compared with none in the placebo arm. Four patients (1.2%) in the osimertinib arm experienced atrial fibrillation compared with 1 (0.3%) in the placebo arm; 6 patients (1.8%) in the osimertinib arm experienced an arrhythmia (other than atrial fibrillation) compared with none in the placebo arm.

Overall, 3 patients (0.4%) experienced keratitis: 2 (0.6%) in the osimertinib arm and 1 (0.3%) in the placebo arm.

Overall, 70.6% of patients in the osimertinib arm and 35.6% of patients in the placebo arm experienced a skin or subcutaneous tissue disorder AE. The most common skin disorders in the osimertinib and placebo arms were paronychia (25.2% and 1.5%, respectively), dry skin (23.4% and 6.4%, respectively), pruritis (19.3% and 8.7%, respectively), and dermatitis acneiform (11.0% and 4.7%, respectively). One patient in the osimertinib arm experienced erythema multiforme. No patients in the ADAURA trial experienced Stevens-Johnson syndrome.

Critical Appraisal

Internal Validity

The ADAURA trial was double blind to minimize bias. However, it is possible that the investigators and patients could have been unblinded due to the profile of AEs in the osimertinib arm. The primary outcome was DFS, which was assessed by the blinded investigators. Disease recurrence was not assessed by an independent central review committee. EGFR mutations were determined by a central laboratory. The enrolled sample size (N = 682 in overall population; N = 470 in stage II to stage IIIA population) was considered adequate.

Baseline characteristics were balanced between treatment arms and few randomized patients had been lost to follow-up as of the data cut-off date. However, a notable proportion of patients that tested positive for eligible EGFR mutations did not provide consent for further eligibility screening to participate in the trial. The reason(s) why these patients did not sign the main study informed consent form is unknown, and no information is available on how these patients compared with the study population in baseline characteristics. It is possible the exclusion of patients who did not sign the main informed consent form could have introduced selection bias.

Randomization was stratified by disease stage, EGFR mutation type, and race. Randomization was not stratified by prior adjuvant chemotherapy status, which was a subgroup that the clinical experts consulted by CADTH identified as important because adjuvant chemotherapy is known to confer an OS benefit in the patient population under review. However, the treatment arms were balanced with respect to the receipt of prior adjuvant chemotherapy and therefore adjuvant chemotherapy status would not bias the results in favour of either treatment arm. The ADAURA trial employed an intention-to-treat analysis, and efficacy outcomes were analyzed by the log-rank test stratified by disease stage, mutation, and race.

The interim analysis was not planned, and the trial is ongoing. The sponsor performed the interim analysis and was unblinded to treatment allocation at the time of this analysis. However, patients and investigators remain blinded to individual treatment allocations and the study is continuing. The primary end point was met at this interim analysis, since the DFS HR for patients with stage II to stage IIIA disease met statistical significance. The DFS HR for the overall population also met statistical significance at this time. A multiple-testing procedure was employed to control the overall type I error at the 5% 2-sided level for the end points of DFS and OS, which was modified to account for the unplanned interim analyses. At the reported interim analysis, formal statistical testing was done for DFS in the primary stage II to stage IIIA population and overall population, and for OS in the stage II to stage IIIA population, per the multiple-testing procedure. OS in the overall population was not formally tested for statistical significance per the multiple-testing procedure.

Due to the early reporting of the study, data maturity is lower than planned by the sponsor at the interim analysis. At the data cut-off date, the sponsor assessed the OS data to be immature, and assessed the data on time to next treatment and PFS to have limited clinical significance due to the immaturity of the data on patients who experienced a disease recurrence event. In addition, most patients have not had the opportunity to receive the planned study treatment duration of 3 years.

In the ADAURA trial, DFS was the primary outcome and OS was the key secondary outcome. DFS is a reasonable surrogate for OS in this treatment setting. Further details on DFS as a surrogate outcome measure for OS are available in Appendix 3. However, using DFS as an end point has disadvantages, such as including non-cancer deaths and having to balance the timing of assessments among treatment arms. The FDA guidelines recommend clearly defining the DFS end point and outlining the schedule for assessments, including an estimation of the treatment effect size and ensuring blinding of treatment assignments to help to reduce bias.²⁴ In the ADAURA trial, DFS and the schedule of assessments were clearly defined and the trial was double blinded, so investigators were unaware of treatment assignments. The European Medicines Agency recommends that when DFS is the primary end point, OS should be the secondary end point, which was done in the ADAURA trial.²⁵ The final analysis of OS data is planned to be conducted when approximately 94 deaths have been observed in the stage II to stage IIIA population (approximately 20% maturity); therefore, OS data are not yet available from the trial to assess whether the observed DFS benefit translates to a clinically meaningful OS benefit.

Subgroup analyses for DFS were specified a priori and based on stratification variables to maintain randomization. However, they were not adjusted for multiplicity and, therefore, were not considered supportive evidence. The analysis of CNS DFS was performed post hoc and, therefore, should also be considered an exploratory analysis.

The ADAURA trial used the SF-36 v2 to measure HRQoL. The SF-36 is a generic, self-reported health-assessment questionnaire that has been used in clinical trials to study the impact of chronic disease on HRQoL. The SF-36 has been previously validated for a variety of health states and diseases. However, the SF-36 has not been validated in patients with NSCLC and no MID for this population was identified in the literature by CADTH. The SF-36 is less sensitive and responsive to changes when compared with the disease-specific European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30 (EORTC QLQ-C30) and its lung cancer–specific supplement (EORTC QLQ-L13). The clinical experts consulted by CADTH indicated it was appropriate to use a generic HRQoL instrument as opposed to a disease-specific instrument in this setting because patients with resected early-stage NSCLC are expected to have a lower symptom burden relative to patients with late-stage NSCLC. The threshold used by the sponsor to categorize change from baseline as improved, worsened, or stable for the PCS and MCS summary scores was greater than the MID identified in the literature by CADTH. The sponsor did not report its rationale for the MIDs used in its analysis of MCS and PCS. The TTD in MCS and PCS was analyzed as pre-specified secondary end points in the stage II to stage IIIA population only (i.e., patients —not in stage IB). The TTD in MCS and PCS analyses in the overall population were analyzed as an exploratory post hoc analysis and, thus, considered a supportive analysis only. Compliance rates for SF-36 v2 completion were high in both treatment arms. However, the number of expected forms decreased significantly at later time points. This is likely because the trial is ongoing and many patients had not yet reached the later time points, at which the SF-36 forms are collected (median duration of total treatment exposure was less than 2 years in both arms; 12% and 10% of patients in the osimertinib and placebo arms, respectively, had received 3 years of treatment as of the data cut-off date). The small number of patients contributing to the HRQoL analysis at later time points introduces uncertainty in estimates reported at these later time points.

External Validity

The dose and treatment regimen used in the ADAURA trial aligns with the Health Canada indication for osimertinib. However, due to the early reporting at the unplanned interim analysis, the majority of patients in the ADAURA trial had not yet received the recommended treatment duration of 3 years as of the data cut-off date. It is uncertain whether the results of the interim analysis are generalizable to a longer treatment duration of osimertinib (e.g., HRQoL, harms outcomes).

The comparator used in the ADAURA trial was placebo. Following tumour resection and adjuvant chemotherapy, Canadian patients with stage IB to stage IIIA (AJCC 7th edition) NSCLC receive active surveillance. Osimertinib will be the first drug to be used as adjuvant therapy following tumour resection (with or without prior adjuvant chemotherapy); therefore, it was appropriate that the ADAURA trial was placebo-controlled as opposed to active comparator–controlled.

The clinical experts consulted by CADTH thought that the inclusion and exclusion criteria used in the trial were appropriate and generally reflected the characteristics of the intended patient population in Canada. However, the ADAURA trial limited enrolment to patients with a WHO performance status of 0 to 1. The clinical experts noted that this criterion is common for clinical trials in NSCLC, but many patients in Canada with resected stage IA to stage IIIB NSCLC have a performance status of 2. The clinical experts indicated they did not believe the exclusion of patients with worse performance status from the ADAURA trial limits the generalizability of the results; however, the clinical experts did indicate that patients with a performance status of 3 or 4 would be unlikely to receive adjuvant treatment until their performance status improved.

The ADAURA trial included patients who had received standard-of-care adjuvant chemotherapy, which is commonly used in Canadian practice. This also aligns with the intended use for osimertinib in Canada, per the clinical experts consulted by CADTH and per the input from the clinician groups, both of which indicated that osimertinib is not intended to replace adjuvant chemotherapy.

The ADAURA trial enrolled patients with stage IB to stage IIIA disease according to the AJCC 7th edition staging system. The Health Canada indication aligns with the overall trial population (i.e., stage IB to stage IIIA). The primary analysis population in the ADAURA trial was stage II to stage IIIA (i.e., non-IB) patients, and results were reported for the overall population. The clinical experts consulted by CADTH noted that the proportion of patients enrolled in the ADUARA trial with stage IB disease was greater than they expected based on their clinical experience. However, the clinical experts noted that Canadian patients with stage IB disease may not be routinely referred to medical oncologists for adjuvant chemotherapy after surgical resection of their tumours.

In current Canadian practice, patients with early-stage NSCLC are not routinely tested for EGFR mutations. In the ADAURA trial, mutation status was determined by a central laboratory test only. It is unknown whether local laboratory testing would be sufficient, or if there could be discrepancies between results from central versus local laboratory tests, based on the trial data.

In the ADAURA trial, radiological assessments were performed more frequently in the trial than in clinical practice. The clinical experts reported that imaging is performed after surgery or after adjuvant chemotherapy every 4 months to 6 months for the first 2 years and annually thereafter until year 5 in regular clinical practice.

The majority of patients enrolled in the ADAURA trial were Asian. The proportion of Asian patients in the trial is higher than in the Canadian NSCLC population, per the clinical experts consulted by CADTH. However, the clinical experts noted that EGFR mutations are found more frequently in Asian patients. It is possible that the patient population with eligible EGFR mutations in Canada may also have a higher proportion of patients of Asian ethnicity. In addition, the majority of patients enrolled in the ADAURA trial were female. The clinical experts reported that this reflects their clinical experience, because EGFR mutations are more common in females.

The clinical experts noted that the ADAURA trial reported a higher EGFR mutation–positive rate than what is currently seen in Canada, where EGFR genetic testing is routinely offered only to patients with locally advanced or metastatic NSCLC. The clinical experts reported that, in their clinical experience, approximately 15% of Canadian patients test positive for EGFR mutations. The clinical experts noted that the higher rate of EGFR-positive genetic tests observed in the ADAURA trial was likely due to many of the study sites being in countries with larger Asian populations. The proportion of patients in Canada with early-stage NSCLC that are EGFR mutation–positive is unknown because testing is not routinely done in these patients. The EGFR mutation test–positivity rate seen in the ADAURA trial may or may not reflect the eligible patient population in Canada.

The patient advocacy groups that provided input for this review reported that patients value DFS and its association with improved quality of life. Common themes for improved outcomes reported by the patient groups also included desire for a cure, delaying disease recurrence, limiting side effects, and maintaining quality of life. This aligns with the outcomes of DFS and HRQoL assessed by the SF-36 v2 in the ADAURA trial.

The types of subsequent anti-cancer therapies received by patients in the ADAURA trial were generally consistent with current Canadian practice, per the clinical experts consulted by CADTH. The clinical experts noted that most patients in the ADAURA trial who received a subsequent anti-cancer therapy were treated with a TKI, which is consistent with current Canadian practice, although the type of TKIs available may vary across jurisdictions. Platinum-doublet chemotherapy also was received as a subsequent treatment by some ADAURA trial participants, which is used in Canadian practice per the clinical experts. However, the clinical experts indicated they would not use a monoclonal antibody in patients with an EGFR mutation.

Indirect Evidence

No indirect treatment comparisons were included in the sponsor’s submission to CADTH or identified in the literature search.

Other Relevant Evidence

No long-term extension studies or other relevant studies were included in the sponsor’s submission to CADTH.

Discussion

Summary of Available Evidence

The systematic review of osimertinib included 1 ongoing phase III RCT. The ADAURA study (N = 682) is an international, multi-centre, phase III, double-blind, placebo-controlled RCT that investigated the efficacy and safety of osimertinib in patients with stage IB to stage IIIA NSCLC (AJCC 7th edition) with a centrally confirmed common-sensitizing EGFR mutation (exon 19 deletion and/or exon 21 [L858R] substitution mutations, either alone or in combination with other EGFR mutations), who have undergone complete tumour resection, with or without post-operative adjuvant chemotherapy. Patients were randomized in a 1:1 ratio to either 80 mg osimertinib orally per day (N = 339) or matching placebo (N = 343). The primary outcome of the ADAURA trial was DFS by investigator assessment. The secondary outcomes were OS and TTD in HRQoL as assessed by the SF-36 v2 summary scores (MCS and PCS) in the stage II to stage IIIA population. Exploratory outcomes included CNS DFS, disease recurrence rate, PFS, time to next treatment, and TTD in HRQoL by the SF-36 v2 summary scores (MCS and PCS) in the overall population (i.e., stage IB to stage IIIA). The ADAURA trial is ongoing, and study results are from an unplanned interim analysis.

Most patients enrolled in the ADAURA trial had stage II to stage IIIA disease, had a WHO performance status of 0, adenocarcinoma histology type, had undergone a lobectomy, had received post-operative adjuvant chemotherapy, and were Asian and female. The mean age of all patients was 62.1 years. Overall, 54.7% of patients had exon 19 deletions and 45.2% had exon 21 (L858R) substitution mutations.

No indirect treatment comparisons or other evidence were included in the sponsor’s submission to CADTH or identified in the literature search.

Interpretation of Results

Efficacy

The results from the ADAURA trial are from an early, unplanned analysis, which limits the interpretation of efficacy results for some outcomes (OS, PFS, time to next treatment, and HRQoL). While the primary analysis was planned to be conducted at 50% DFS data maturity, this early interim analysis was conducted at 33.2% DFS maturity. Furthermore, most patients had not yet received the recommended 3-year treatment duration at the time of the data cut-off date, with an average treatment duration of 21.7 months and 18.6 months in the osimertinib and placebo arms, respectively. The average duration of follow-up was approximately 2 years, and the clinical experts consulted by CADTH indicated that patients with resected early-stage NSCLC are typically followed for 5 years in standard clinical practice. These limitations contribute to the uncertainty regarding the efficacy of osimertinib with respect to outcomes other than DFS.

DFS was considered an important outcome by the clinical panel consulted by CADTH, clinician groups, and the patient advocacy groups that provided input for this review. The ADAURA trial met its primary end point of DFS at the unplanned interim analysis. Osimertinib demonstrated a statistically significant improvement in DFS compared with placebo in both the stage II to stage IIIA (primary analysis) population and the overall population. These results were considered clinically significant by the clinical panel consulted by CADTH and the clinician groups that provided input for this submission. Results of the pre-specified subgroup analyses were consistent with the primary analysis in showing a DFS benefit with osimertinib compared with placebo in all patient subgroups. The clinical experts thought that the results of the supportive subgroup analyses suggest the magnitude of benefit may be smaller in the stage IB population compared with patients with stage II and stage IIIA disease. Furthermore, the clinical experts also thought that the subgroup analyses indicated the magnitude of benefit is larger in the later-stage groups (i.e., II and IIIA), which the clinical experts indicated are at a higher risk of relapse. DFS with recurrence in the CNS only was analyzed post hoc. This exploratory analysis suggested there may be benefit in CNS DFS with osimertinib, although this is associated with uncertainty due to the small number of patients who had events. The clinical panel consulted by CADTH indicated that these results have clinical significance, since CNS metastases are associated with high morbidity; however, firm conclusions cannot be drawn based on these results due to the post hoc nature of the analysis.

The overall disease recurrence rate was lower in the osimertinib arm compared with the placebo arm. In the osimertinib arm, most disease recurrences were local or regional only and the most common sites of recurrence were in the lung and lymph nodes. In the placebo arm, most disease recurrences were distant only or the patients had disease recurrence at both distant and local or regional sites. The most common sites of disease recurrence in the placebo arm were in the lung, lymph nodes, CNS, and bone. The clinical experts consulted by CADTH indicated that the site of disease recurrence is clinically significant because local or regional disease has the potential to be cured, whereas patients with distant metastases cannot be cured.

The clinical panel consulted by CADTH and clinician groups indicated that OS was an equally or more important outcome compared with DFS. It was noted that the relative importance of DFS versus OS is controversial among clinicians who treat NSCLC, and it is uncertain whether benefits in DFS translate into benefits in OS or just delay the time to recurrence. The patient advocacy groups indicated that patients want to be cured but they also value DFS and its association with improved HRQoL. At the time of the data cut-off date for the interim analysis, the OS data were immature, per the sponsor’s assessment (4% of randomized patients had died). The comparison of OS in the stage II to stage IIIA population was not statistically significantly different between the treatment groups. Thus, at the time of this review, it cannot be concluded that osimertinib confers an OS benefit compared with placebo. However, the patients providing input on this submission indicated they value quicker access to treatment and therefore are willing to take osimertinib based on its DFS benefit and do not want to wait for mature OS data to become available. Data on PFS and time to next treatment are also immature, per the sponsor’s assessment. These outcomes were not included in the multiple-testing procedure; thus, the results for these end points are associated with substantial uncertainty and are limited in their statistical and clinical significance.

The patient advocacy groups that provided input for this review emphasized the importance of treatments maintaining HRQoL. Similarly, the clinical panel indicated that the impact of adjuvant osimertinib on HRQoL is important for deciding if the treatment is worthwhile to patients in this setting. HRQoL was measured by the SF-36 v2. The HRQoL end points of interest specified in the study protocol were TTD in the MCS and PCS scores in the stage II to stage IIIA population, which were analyzed as secondary end points but not adjusted for multiple comparisons. Therefore, limited conclusions can be drawn on the effect of osimertinib on HRQoL end points and the results should be considered only as supportive evidence for the primary results of DFS and OS. As of the interim analysis, a numerically greater proportion of patients with stage II to stage IIIA disease in the osimertinib arm experienced a confirmed deterioration in the PCS score of the SF-36 of 3.1 points or greater or death compared with the placebo arm. A numerically similar proportion of patients in each treatment arm experienced confirmed deterioration in the MCS score of the SF-36 greater than the MID or death. The TTD in MCS and PCS scores in the overall population, which aligns with the reimbursement request, were analyzed as a post hoc exploratory analysis and the results were consistent with the results observed in the stage II to stage IIIA population.

Harms

AEs were reported more frequently in the osimertinib arm compared with the placebo arm and most AEs were grade 1 to 2. The overall frequency of AEs in the placebo arm was higher than expected by the clinical experts consulted by CADTH. The most frequently reported AEs in the osimertinib arm were diarrhea, paronychia, dry skin, pruritis, cough, and stomatitis; the most frequently reported AEs in the placebo arm were diarrhea and cough. The incidence of cough in both treatment arms was higher than expected by the clinical experts consulted by CADTH. SAEs were reported more frequently in the osimertinib arm compared with the placebo arm. The most frequently reported SAEs were pneumonia, cataracts, diarrhea, acute kidney injury, ureterolithiasis, and femur fracture. As of the interim analysis, 9 patients (2.7%) in the osimertinib arm and 20 patients (5.8%) in the placebo arm had died. As of the interim analysis, more patients in the osimertinib arm had discontinued study treatment due to AEs compared with the placebo arm.

The product monograph for osimertinib contains warnings and precautions for interstitial lung disease and pneumonitis, QT interval prolongation, left ventricular dysfunction, and cardiomyopathy. Keratitis and skin disorders were also identified as harms of special interest by the clinical experts consulted by CADTH. Interstitial lung disease and pneumonitis occurred in a total of 3% of patients in the osimertinib arm, whereas no patients in the placebo arm experienced these AEs. Cardiac disorders occurred more frequently in the osimertinib arm, including QT interval prolongation. Two patients in the osimertinib arm experienced cardiomyopathy and no patients in either treatment arm experienced congestive heart failure. Keratitis occurred in a small number of patients in both treatment arms. Skin and subcutaneous tissue disorders occurred more frequently in the osimertinib arm compared with the placebo arm, which included dermatitis acneiform, dry skin, paronychia, pruritus, rash, and skin fissures. One patient in the osimertinib arm experienced erythema multiforme and no patients experienced Stevens-Johnson syndrome.

Although the safety data from the ADAURA trial are consistent with the known safety profile of osimertinib, the results are from an unplanned interim analysis and most patients in the ADAURA trial have not yet had the opportunity to receive the planned treatment duration of 3 years. As a result, evaluation of the harms associated with the planned 3-year treatment is limited by the data available at the time of this review.

Conclusions

One ongoing phase III, double-blind, placebo-controlled RCT (ADAURA) provided direct evidence regarding the safety and efficacy of osimertinib adjuvant therapy in adult patients with resected stage IB to stage IIIA (AJCC 7th edition) NSCLC. The trial included patients regardless of whether they had received standard adjuvant chemotherapy. Compared with placebo, patients treated with adjuvant osimertinib showed benefits in DFS. The ADAURA trial met its primary end point at an unplanned interim analysis that showed a statistically significant difference in DFS in both the stage II to stage IIIA population and overall population. This DFS benefit was consistently observed in all pre-specified subgroups. The results did not support conclusions for an effect of osimertinib on OS, PFS, and time to next treatment, and any potential clinical benefit for these outcomes is associated with uncertainty due to the immaturity of the data. Conclusions could not be drawn for the effect of osimertinib on HRQoL end points, as these end points were not adjusted for multiple comparisons. The majority of study participants reported treatment-emergent AEs. A greater proportion of patients in the osimertinib arm experienced an AE compared with the placebo arm. Interstitial lung disease, pneumonitis, cardiac disorders, and skin and subcutaneous tissue disorders were reported more frequently in the osimertinib arm. These AEs are consistent with the known safety profile of osimertinib. Keratitis was uncommon in both treatment arms. A greater number of patients in the osimertinib arm discontinued study treatment due to AEs or because they experienced an SAE. Overall, 4% of study participants had died as of the interim analysis.

References

1.Canadian Cancer Statistics Advisory Committee. Canadian Cancer Statistics: A 2020 special report on lung cancer. Toronto, ON: Canadian Cancer Society; 2020: https://cdn.cancer.ca/-/media/files/cancer-information/resources/publications/2020-canadian-cancer-statistics-special-report/2020-canadian-cancer-statistics-special-report-en.pdf?rev=15c66a0b05f5479e935b48035c70dca3&hash=3D51B0D0FB5C3F7E659F896D66495CE8. Accessed 2021 Apr 26.

2.BC Cancer. Lung. 2013; http://www.bccancer.bc.ca/health-professionals/clinical-resources/cancer-management-guidelines/lung/lung#Management. Accessed 2021 Jun 3.

3.Midha A, Dearden S, McCormack R. EGFR mutation incidence in non-small-cell lung cancer of adenocarcinoma histology: a systematic review and global map by ethnicity (mutMapII). Am J Cancer Res. 2015;5(9):2892-2911. PubMed

4.Graham RP, Treece AL, Lindeman NI, et al. Worldwide Frequency of Commonly Detected EGFR Mutations. Arch Pathol Lab Med. 2018;142(2):163-167. PubMed

5.Melosky B, Banerji S, Blais N, et al. Canadian consensus: a new systemic treatment algorithm for advanced EGFR-mutated non-small-cell lung cancer. Curr Oncol. 2020;27(2):e146-e155. PubMed

6.National Comprehensive Cancer Network. Non-small cell lung cancer. (NCCN Clinical practice guidelines in oncology). Plymouth Meeting (PA): National Comprehensive Cancer Network; 2021.

7.Alberta Health Services. Non-small cell lung cancer stage III. (Clinical practice guideline LU-003). 2012: https://www.albertahealthservices.ca/assets/info/hp/cancer/if-hp-cancer-guide-lu003-nlscs-stage3.pdf. Accessed June 3, 2021.

8.Pignon J-P, Tribodet H, Scagliotti GV, et al. Lung adjuvant cisplatin evaluation: a pooled analysis by the LACE Collaborative Group. J Clin Oncol. 2008;26(21):3552-3559. PubMed

9.Arriagada R, Auperin A, Burdett S, et al. Adjuvant chemotherapy, with or without postoperative radiotherapy, in operable non-small-cell lung cancer: two meta-analyses of individual patient data. Lancet. 2010;375(9722):1267-1277. PubMed

10.Artal Cortés Á, Calera Urquizu L, Hernando Cubero J. Adjuvant chemotherapy in non-small cell lung cancer: state-of-the-art. Translational lung cancer research. 2015;4(2):191-197. PubMed

11.Chouaid C, Danson S, Andreas S, et al. Adjuvant treatment patterns and outcomes in patients with stage IB-IIIA non-small cell lung cancer in France, Germany, and the United Kingdom based on the LuCaBIS burden of illness study. Lung Cancer. 2018;124:310-316. PubMed

12.Tagrisso (osimertinib tablets): 40 mg and 80 mg osimertinib as osimertinib mesylate Epidermal Growth Factor Receptor (EGFR) Tyrosine Kinase Inhibitor[product monograph]. Mississauga (ON): AstraZeneca Canada Inc.; 2021 Jan 15.

13.Clinical Study Report: D5164C00001. Phase III, double-blind, randomized, placebo-controlled multi-centre, study to assess the efficacy and safety of AZD9291 versus placebo, in patients with epidermal growth factor receptor mutation positive stage IB-IIIA non-small cell lung carcinoma, following complete tumour resection with or without adjuvant chemotherapy (ADAURA) [internal sponsor's report]. Cambridge (GB): AstraZeneca; 2020 Jul 31.

14.Drug Reimbursement Review sponsor submission: tagrisso (osimertinib), 40 mg and 80 mg osimertinib tablets[internal sponsor's package]. Mississauga (ON): AstraZeneca Canada Inc.; 2021 Mar 24.

15.Canadian Cancer Statistics Advisory Committee. Canadian Cancer Statistics 2019. Toronto, ON: Canadian Cancer Society; 2019: https://www.cancer.ca/~/media/cancer.ca/CW/cancer%20information/cancer%20101/Canadian%20cancer%20statistics/Canadian-Cancer-Statistics-2019-EN.pdf?la=en&la=en. Accessed 2021 Apr 26.

16.Detterbeck FC, Boffa DJ, Kim AW, Tanoue LT. The eighth edition lung cancer stage classification. Chest. 2017;151(1):193-203.

17.Birring SS, Peake MD. Symptoms and the early diagnosis of lung cancer. Thorax. 2005;60(4):268-269. PubMed

18.Ellis PM, Verma S, Sehdev S, Younus J, Leighl NB. Challenges to implementation of an epidermal growth factor receptor testing strategy for non-small-cell lung cancer in a publicly funded health care system. J Thorac Oncol. 2013;8(9):1136-1141. PubMed

19.McGowan J, Sampson M, Salzwedel DM, Cogo E, Foerster V, Lefebvre C. PRESS Peer Review of Electronic Search Strategies: 2015 guideline statement. J Clin Epidemiol. 2016;75:40-46. PubMed

20.Grey matters: a practical tool for searching health-related grey literature. Ottawa (ON): CADTH; 2019: https://www.cadth.ca/grey-matters. Accessed 2021 Jul 16.

21.Wu YL, Herbst RS, Mann H, Rukazenkov Y, Marotti M, Tsuboi M. ADAURA: phase III, double-blind, randomized study of osimertinib versus placebo in EGFR mutation-positive early-stage NSCLC after complete surgical resection. Clin Lung Cancer. 2018;19(4):e533-e536. PubMed

22.Wu YL, Tsuboi M, He J, et al. Osimertinib in resected EGFR-mutated non-small-cell lung cancer. N Engl J Med. 2020;383(18):1711-1723. PubMed

23.Health Canada reviewer's report: Tagrisso (osimertinib mesylate) [internal sponsor's report]. In: Drug Reimbursement Review sponsor submission: tagrisso (osimertinib), 40 mg and 80 mg osimertinib tablet. Mississauga (ON): AstraZeneca Canada Inc.; 2021 Mar 24.

24.Clinical trial endpoints for the approval of cancer drugs and biologics. (Guidance for industry). Silver Spring (MD): U.S. Food and Drug Administration; 2018: https://www.fda.gov/regulatory-information/search-fda-guidance-documents/clinical-trial-endpoints-approval-cancer-drugs-and-biologics. Accessed 2021 Apr 26.

25.Guideline on the clinical evaluation of anticancer medicinal products. London (GB): European Medicines Agency; 2019: https://www.ema.europa.eu/en/documents/scientific-guideline/draft-guideline-evaluation-anticancer-medicinal-products-man-revision-6_en.pdf. Accessed 2021 Apr 26.

26.Mauguen A, Pignon JP, Burdett S, et al. Surrogate endpoints for overall survival in chemotherapy and radiotherapy trials in operable and locally advanced lung cancer: a re-analysis of meta-analyses of individual patients' data. Lancet Oncol. 2013;14(7):619-626. PubMed

27.Savina M, Gourgou S, Italiano A, et al. Meta-analyses evaluating surrogate endpoints for overall survival in cancer randomized trials: A critical review. Crit Rev Oncol Hematol. 2018;123:21-41. PubMed

28.Fiteni F, Westeel V, Bonnetain F. Surrogate endpoints for overall survival in lung cancer trials: a review. Expert Rev Anticancer Ther. 2017;17(5):447-454. PubMed

29.Maruish ME. User's manual for the SF-36v2 health survey; 3rd edition. Lincoln (RD): Quality Metric Incorporated; 2011.

30.Ware JE, Sherbourne CD. The MOS 36-item short-form health survey (SF-36). I. Conceptual framework and item selection. Med Care. 1992;30(6):473-483. PubMed

31.Jenkinson C, Stewart-Brown S, Petersen S, Paice C. Assessment of the SF-36 version 2 in the United Kingdom. J Epidemiol Community Health. 1999;53:46-50. PubMed

32.The Short-Form-36 health survey (Rand Corporation and John E. Ware Jr., 1990, revised 1996). Excerpt from Ian McDowell Measuring health: a guide to rating scales and questionnaires. New York (NY): Oxford University Press; 2006: http://www.med.uottawa.ca/courses/CMED6203/Index_notes/SF36%20fn%20.pdf. Accessed 2021 Apr 26.

33.Möller A, Sartipy U. Long-term health-related quality of life following surgery for lung cancer. Eur J Cardiothorac Surg. 2012;41(2):362-367. PubMed

34.Zhang L, Su Y, Hua Y, et al. Validation of EORTC QLQ-LC43 for Chinese patients with lung cancer. Lung Cancer. 2014;85(1):94-98. PubMed

35.Bergman B, Aaronson NK, Ahmedzai S, Kaasa S, Sullivan M. The EORTC QLQ-LC13: a modular supplement to the EORTC Core Quality of Life Questionnaire (QLQ-C30) for use in lung cancer clinical trials. EORTC Study Group on Quality of Life. Eur J Cancer. 1994;30a(5):635-642. PubMed

36.Pompili C, Brunelli A, Xiume F, et al. Prospective external convergence evaluation of two different quality-of-life instruments in lung resection patients. Eur J Cardiothorac Surg. 2011;40(1):99-105. PubMed

37.Chiu CC, Lee KT, Wang JJ, et al. Preoperative health-related quality of life predicts minimal clinically important difference and survival after surgical resection of hepatocellular carcinoma. J Clin Med. 2019;8(5). PubMed

38.Jayadevappa R, Malkowicz SB, Wittink M, Wein AJ, Chhatre S. Comparison of distribution- and anchor-based approaches to infer changes in health-related quality of life of prostate cancer survivors. Health Serv Res. 2012;47(5):1902-1925. PubMed

Appendix 1: Literature Search Strategy

Note this appendix has not been copy-edited.

Clinical Literature Search

Overview

Interface: Ovid

Databases

• MEDLINE All (1946-present)

• Embase (1974-present)

• Note: Subject headings and search fields have been customized for each database. Duplicates between databases were removed in Ovid.

Date of search: April 19, 2021

Alerts: Bi-weekly search updates until project completion

Search filters applied: Randomized controlled trials or controlled clinical trials

Limits

• No publication date limits

• No language limits

• Conference abstracts: excluded

Table 24: Syntax Guide

Multi-Database Strategy

1. (tagrisso* or osimertinib* or AZD-9291 or AZD9291 or mereletinib* or AZ7550 or AZ-7550 or AZ5104 or AZ-5104 or 3C06JJ0Z2O or RDL94R2A16 or NI2ZUZ6F4O or 2DWZ6SE1E1 or Q27162944).ti,ab,ot,kf,hw,nm,rn.

2. 1 use medall

3. *osimertinib/

4. (tagrisso* or osimertinib* or AZD-9291 or AZD9291 or mereletinib* or AZ7550 or AZ-7550 or AZ5104 or AZ-5104).ti,ab,kw,dq.

5. 3 or 4

6. (conference abstract or conference review).pt.

7. 5 not 6

8. 7 use oemezd

9. 2 or 8

10. (Randomized Controlled Trial or Controlled Clinical Trial or Pragmatic Clinical Trial or Equivalence Trial or Clinical Trial, Phase III).pt.

11. Randomized Controlled Trial/

12. exp Randomized Controlled Trials as Topic/

13. "Randomized Controlled Trial (topic)"/

14. Controlled Clinical Trial/

15. exp Controlled Clinical Trials as Topic/

16. "Controlled Clinical Trial (topic)"/

17. Randomization/

18. Random Allocation/

19. Double-Blind Method/

20. Double Blind Procedure/

21. Double-Blind Studies/

22. Single-Blind Method/

23. Single Blind Procedure/

24. Single-Blind Studies/

25. Placebos/

26. Placebo/

27. Control Groups/

28. Control Group/

29. (random* or sham or placebo*).ti,ab,hw,kf,kw.

30. ((singl* or doubl*) adj (blind* or dumm* or mask*)).ti,ab,hw,kf,kw.

31. ((tripl* or trebl*) adj (blind* or dumm* or mask*)).ti,ab,hw,kf,kw.

32. (control* adj3 (study or studies or trial* or group*)).ti,ab,kf,kw.

33. (Nonrandom* or non random* or non-random* or quasi-random* or quasirandom*).ti,ab,hw,kf,kw.

34. allocated.ti,ab,hw.

35. ((open label or open-label) adj5 (study or studies or trial*)).ti,ab,hw,kf,kw.

36. ((equivalence or superiority or non-inferiority or noninferiority) adj3 (study or studies or trial*)).ti,ab,hw,kf,kw.

37. (pragmatic study or pragmatic studies).ti,ab,hw,kf,kw.

38. ((pragmatic or practical) adj3 trial*).ti,ab,hw,kf,kw.

39. ((quasiexperimental or quasi-experimental) adj3 (study or studies or trial*)).ti,ab,hw,kf,kw.

40. (phase adj3 (III or "3") adj3 (study or studies or trial*)).ti,hw,kf,kw.

41. or/10-40

42. 9 and 41

43. remove duplicates from 42

Clinical Trials Registries

ClinicalTrials.gov

Produced by the U.S. National Library of Medicine. Targeted search used to capture registered clinical trials.

Search: Tagrisso (osimertinib) AND Non-small cell lung cancer (NSCLC)

WHO ICTRP

International Clinical Trials Registry Platform, produced by the World Health Organization. Targeted search used to capture registered clinical trials.

Search terms: Tagrisso (osimertinib) AND Non-small cell lung cancer (NSCLC)

Health Canada’s Clinical Trials Database

Produced by Health Canada. Targeted search used to capture registered clinical trials.

Search terms: Tagrisso (osimertinib) AND Non-small cell lung cancer (NSCLC)

EU Clinical Trials Register

European Union Clinical Trials Register, produced by the European Union. Targeted search used to capture registered clinical trials.

Search terms: Tagrisso (osimertinib) AND Non-small cell lung cancer (NSCLC)

Grey Literature

Search dates: April 13 to April 15, 2021

Keywords: (Tagrisso OR osimertinib OR AZD-9291) AND (Non-small cell lung cancer OR NSCLC)

Limits: No publication date limits

Updated: Search updated prior to the completion of stakeholder feedback period

Relevant websites from the following sections of the CADTH grey literature checklist Grey Matters: A Practical Tool for Searching Health-Related Grey Literature were searched:

• Health Technology Assessment Agencies

• Health Economics

• Clinical Practice Guidelines

• Drug and Device Regulatory Approvals

• Advisories and Warnings

• Drug Class Reviews

• Clinical Trials Registries

• Databases (free)

• Health Statistics

• Internet Search

• Open Access Journals

Appendix 2: Detailed Outcome Data

Note this appendix has not been copy-edited.

Table 25: Sensitivity Analysis of DFS for Evaluation-Time Bias — FAS, Stage II to Stage IIIA Population

CI = confidence interval; DFS = disease-free survival; HR = hazard ratio; FAS = full analysis set.

^a The analysis was performed using a log-rank test stratified by stage (II versus IIIA), race (Asian versus non-Asian), and mutation type (Ex19del versus L858R).

^b An HR of less than 1 favours osimertinib. The HR and CI were obtained directly from the U and V statistics.

Source: Clinical Study Report.¹³

Table 26: Sensitivity Analysis of DFS for Quantitative Interactions — FAS, Stage II to Stage IIIA Population

DFS = disease-free survival; FAS = full analysis set.

^a Likelihood ratio P values.

^b Significance level for interaction test was 10%.

^c Quantitative interaction is defined as treatment effects in same direction but of different magnitude.

Source: Clinical Study Report.¹³

Figure 15: Sensitivity Analysis of DFS for Attrition Bias — FAS, Stage II to Stage IIIA

AZD9291 = osimertinib; DFS = disease-free survival; FAS = full analysis set.

Kaplan-Meier plot with censoring and event flags reversed from the primary DFS analysis.

Source: Clinical Study Report.¹³

Table 27: Adverse Events by Grade Reported in 10% or More of Patients in Either Arm of the ADAURA Trial — Safety Population

Source: Clinical Study Report,¹³ Wu et al. (2020).²²

Appendix 3: Description and Appraisal of Outcome Measures

Note this appendix has not been copy-edited.

Aim

To describe the following outcome measures and review their measurement properties (validity, reliability, responsiveness to change, and MID):

• DFS as a surrogate outcome for OS

• SF-36 v2

Findings

Table 28: Summary of Outcome Measures and Their Measurement Properties

DFS = disease-free survival; EORTC QLQ = European Organization for Research and Treatment of Cancer Quality of Life Questionnaire; HRQoL = health-related quality of life; MCS = Mental Component Summary; MID = minimal important difference; NSCLC = non–small cell lung cancer; OS = overall survival; PCS = Physical Component Summary; SF-36 v2 = Short Form (36) Health Survey version 2.

Disease-Free Survival as a Surrogate Outcome for Overall Survival

DFS by investigator assessment was the primary outcome used to evaluate the efficacy of osimertinib compared with placebo in the ADAURA trial.¹³ Radiological assessments for disease recurrence were conducted at 12 weeks, 24 weeks, every 24 weeks until 5 years, and yearly thereafter. DFS was defined as the number of days from the date of randomization to the date of recurrence or death (by any cause) during study treatment (up to 3 years) or beyond treatment discontinuation according to the study plan.

In a guidance document assessing end points used in clinical trials for cancer drugs and biologics, the FDA noted the benefits and drawbacks to using DFS.²⁴ Its advantages include being an objective measure based on quantitative assessment, being evaluable sooner, and needing a smaller sample size compared with studies using OS as an end point. Using DFS as an end point has its disadvantages such as including non-cancer deaths and having to balance the timing of assessments among treatment arms. For the traditional drug approval process (which is based on clinical benefit being demonstrated or the effect on a surrogate end point that is known to predict clinical benefit), DFS has been used in studies for breast cancer, colorectal cancer, gastrointestinal stromal tumours, melanoma, and renal cell carcinoma. The FDA guidelines also recommend clearly defining the end point, outlining the schedule for assessments, including an estimation of the treatment effect size and ensuring blinding of treatment assignments to help to reduce bias. The European Medicines Agency recommends that when DFS is the primary end point, OS should be the secondary end point.²⁵

Mauguen et al. assessed meta-analyses investigating DFS as a surrogate for OS in operable and locally advanced NSCLC.²⁶ In the adjuvant setting, 17 trials (N = 5,379 patients) evaluated chemotherapy versus no chemotherapy while another 7 studies (N = 2,247) assessed radiotherapy and chemotherapy versus just radiotherapy. Individual- and trial-level associations for DFS versus OS were ρ² (95% CI = 0.83 (0.83 to 0.83) and R² (95% CI) = 0.92 (0.88 to 0.95), respectively, for the chemotherapy/none studies. For the radio + chemotherapy/radiotherapy studies, ρ² (95% CI) = 0.87 (0.87 to 0.87) and R² (95% CI) = 0.99 (0.98 to 1.00). The 3-year DFS compared with the 5-year OS (R² [95% CI]) were estimated to be 0.88 (0.83 to 0.93) and 0.96 (0.93 to 0.99) for the different adjuvant treatment types.²⁶

Savina et al.²⁷ reviewed meta-analyses evaluating surrogate end points for OS, one of which was the Mauguen et al.²⁶ publication, and found that DFS was a good surrogate for studies investigating colon cancer, operable and locally advanced NSCLC, gastric cancer, and locally advanced head and neck cancer all of which were treated in an adjuvant setting with chemotherapy and/or radiotherapy. Based on the findings in the Mauguen et al.²⁶ publication, Savina et al. concluded the strength of the associations were high as per the German Institute of Quality and Efficiency in Health Care (IQWiG) guidelines and excellent according to the Biomarker-Surrogate Evaluation Schema (BSES) guidelines.²⁷ Fiteni et al. came to similar conclusions that the associations showed high reliability and correlation based on IQWiG guidelines and were valid surrogated end points for a specific disease and intervention class (i.e., level 2) according to the Fleming 4-level hierarchy.²⁸

Also in the review, Mauguen et al. commented that surrogate end points are validated for specific therapies, and in the case of the ADAURA trial of osimertinib, a targeted EGFR-specific treatment, validation would need to be reassessed.²⁶ The investigators also noted additional confounding factors, such as subsequent cancer therapies and treatment crossover in trials, that can impact OS.

Short Form (36) Health Survey Version 2

The SF-36 is a generic, self-reported health-assessment questionnaire that has been used in clinical trials to study the impact of chronic disease on HRQoL.²⁹ There are 2 versions of the instrument, the original SF-36³⁰ and SF-36 version 2 (SF-36 v2).^29,31 Compared with the original SF-36, the SF-36 v2 contains minor changes to the original survey, including changes to: instructions (reduced ambiguity), questions and answers (better layout), item-level response choices (increased), cultural/language comparability (increased), and elimination of a response option from the items in the mental health and vitality subscales.^29,31 The questionnaire consists of 36 items representing 8 subscales: Physical Functioning (PF; 10 items), Role-Physical (RP; 4 items), Bodily Pain (BP; 2 items), General Health (GH; 5 items), Vitality (VT; 4 items), Social Functioning (SF; 2 items), Role-Emotional (RE; 3 items), and Mental Health (MH; 5 items). The second question of the survey is a single item not used for scoring the 8 scales, but instead used to estimate the general health from a cross-sectional standpoint.³² The SF-36 has a recall period of 4 weeks and item response options are presented on a 3-point to 6-point, Likert-like scale.^29,31 Each item is converted to a score ranging from 0 to 100 where a higher value indicates a more favourable health state and item scores are averaged together to create the 8 subscale scores. The SF-36 also provides 2 component summaries, the PCS and the MCS, which are created by aggregating the 8 subscales according to a scoring algorithm. The first 4 subscales (PF, RP, BP, and GH) belong to the PCS while the next 4 subscales (VT, SF, RE, and MH) make up the MCS. Like the individual items, the PCS, MCS, and 8 subscale scores are each measured from 0 to 100. The subscale and summary scores (PCS and MCS) are t scores standardized to a reference population with a mean of 50 and standard deviation of 10.²⁹ Thus, a score of 50 on any scale would be at the average or norm of the reference and a score 10 points lower (i.e., 40) would be 1 standard deviation below the norm.

Möller and Sartipy (2011) conducted a prospective, population-based, cohort study of Swedish patients who had lung surgery and assessed HRQoL using the Swedish version of the SF-36.³³ Overall, 166 patients completed the questionnaire before as well as 6 months and 2 years after lung surgery. In total, 69.9% of patients were reported as having stage I to III tumours, 23.5% had received adjuvant chemotherapy, and 4.2% had adjuvant radiotherapy. After 2 years, the patients were compared with an age- and sex-matched reference population. Mean (SD) PCS scores decreased from baseline to 6 months post operation and were similar at 2 years (47.4 [10.3], 41.1 [10.5], and 41.4 [11.6], respectively). The patient PCS score was lower at 2 years compared with the reference population (44.0 [3.9]; P = 0.013). Mean (SD) MCS scores increased for the same 3 timepoints from 40.5 [13.8] to 44.6 [12.6] at 6 months to 46.9 [12.2] at 2 years but were also lower than the reference population (50.6 [1.7]; P = 0.01).

In a 2014 study by Zhang et al., 317 Chinese patients with lung cancer were given the SF-36 and the EORTC QLQ-C30 and supplemental lung cancer–specific module (QLQ-L13).³⁴ While the C30 measures HRQoL in relation to cancer, the LC13 focuses on symptoms and side effects specific to lung cancer and treatment, and both instruments have been validated.^34,35 The 2 EORTC QLQs administered together were denoted as the LC43.³⁴ Pearson’s correlation coefficients were calculated between the SF-36 and LC43 and were found to be highest among similar subscales such as 0.675 for PF on either questionnaire, 0.621 for PF of the SF-36 and role functioning of the LC43, 0.567 for both pain subscales, and 0.537 for SF on either questionnaire. Most of the other correlation coefficients were less than 0.5 which could be due to the SF-36 being a measure of general health and not being disease-specific like the LC43.

In a similar manner, Pompili et al. compared the SF-36 with the EORTC QLQ-C30 and LC13 in 33 patients with NSCLC who underwent pulmonary resection.³⁶ The questionnaires were completed before and 3 months after the operation. When assessing the pre- and post-operative changes using Cohen effect size (ES), the investigators found that none of the SF-36 subscales showed a large and clinically meaningful effect (less than −0.8 or greater than 0.8). An average moderate ES (0.5 < |ES| < 0.79) was noted for the PCS and RP scores while all others ESs were small demonstrating limited responsiveness of the SF-36 to changes in patients with NSCLC who have had surgery. To assess convergent validity, correlation coefficients were calculated between the 2 questionnaires. Two subscales had correlations greater than 0.5: the SF of either instrument (0.55) along with the MH of the SF-36 and emotional functioning of the EORTC QLQ-C30 and LC13 (0.57). Again, these results show that the SF-36 may be less sensitive to health changes in patients with lung cancer and that the questionnaires measure different aspects of HRQoL.

For the general population, the SF-36 User’s Manual proposed the following MIDs: a change of 2 points on the PCS, and 3 points on the MCS.²⁹ The manual also suggested the following minimal mean group differences, in terms of t score points, for SF-36 v2 individual subscale scores: PF, 3; RP, 3; BP, 3; GH, 2; VT, 2; SF, 3; RE, 4; and MH, 3. It should be noted that these MIDs were determined as appropriate for groups with mean t score ranges of 30 to 40; for higher t score ranges, values may be higher.²⁹ Furthermore, the MIDs do not represent patient-derived scores. The MIDs for the SF-36 v2 are based on clinical and other non–patient-reported anchors.

Chiu et al. assessed 369 patients in Taiwan who received surgical resection for hepatocellular carcinoma to estimate a MID for the SF-36.³⁷ Using a distribution-based method (one-half SD), they suggested 3.6 points for the PCS and 4.2 points for the MCS. In a separate study of 528 prostate cancer survivors, Jayadevappa estimated MIDs for individual items of the SF-36 using both distribution-based methods (one-half and one-third SD) and anchor-based methods (baseline global health and patient-reported symptoms).³⁸ Mean MIDs using distribution-based methods ranged from 7.0 to 16.1 points while mean estimates from anchor-based methods ranged from 5.0 to 12.4 points. The low and high ends of the ranges were for the Mental Health and Role-Physical items, respectively.

No MID was identified from the literature specific to patients with NSCLC.

Pharmacoeconomic Review

Executive Summary

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

Table 1: Submitted for Review

EGFR = epidermal growth factor receptor; L858R = sensitizing mutation in the EGFR gene with substitution of a leucine with an arginine at position 858 in exon 21; NOC = Notice of Compliance; NSCLC = non–small cell lung cancer.

Table 2: Summary of Economic Evaluation

1L DM = first-line treatment for distant metastatic disease; 2L DM = second-line treatment for distant metastatic disease; AE = adverse event; DF = disease-free; EGFR = epidermal growth factor receptor; ICER = incremental cost-effectiveness ratio; LR = local or regional recurrence; LY = life-year; NSCLC = non–small cell lung cancer; OS = overall survival; QALY = quality-adjusted life-year; RDI = relative dose intensity.

Conclusions

The CADTH clinical review found that compared with placebo, patients treated with osimertinib showed benefits in disease-free (DF) survival (DFS). However, the results did not support conclusions for an effect of osimertinib on OS, PFS, and time to next treatment, and any potential clinical benefit for these outcomes is associated with uncertainty due to the immaturity of the data.

CADTH undertook reanalyses to address limitations relating to uncertainty in long-term DF status and OS associated with adjuvant osimertinib, time-to-establish cure, subsequent therapies in local or regional recurrence (LR), local or regional disease management costs, radiotherapy and dialysis costs for progressed disease, health state utility values, relative dose intensities (RDIs), and transitions from second-line treatment for distant metastatic disease (2L DM) to death for patients not re-treated with osimertinib. Based on the CADTH reanalysis, the incremental cost-effectiveness ratio (ICER) for osimertinib relative to active surveillance was $328,026 per quality-adjusted life-year (QALY). A reduction of at least 82% in the price of osimertinib is required for osimertinib to achieve an ICER of $50,000 per QALY compared with active surveillance.

Crucially, the cost-effectiveness of osimertinib is contingent on long-term DFS and whether this translates into OS gains. The sponsor’s approach to model the link between DFS and OS is evidence-based and appropriately modelled; however, it lacks data to validate long-term DFS and OS for patients who receive osimertinib in the adjuvant setting. The CADTH clinical review concluded there is substantial uncertainty as to whether osimertinib will generate OS benefit in the adjuvant setting. Longer-term evidence is required to validate what the OS will be for patients taking osimertinib as adjuvant therapy. If relative to active surveillance, osimertinib only delays the rate at which disease progression occurs, then, based on optimistic OS benefits associated with DFS, a price reduction of 72% (log-logistic extrapolation) to 81% (Weibull extrapolation) may be sufficient to achieve cost-effectiveness at a threshold of $50,000 per QALY. If there is no OS benefit, the incremental QALYs generated from osimertinib are substantially less than the CADTH base-case estimate, requiring a price reduction of more than 95% to ensure cost-effectiveness. Conversely, if osimertinib leads to sustained long-term DFS and this translates into long-term overall survival (OS) benefits, then the cost-effectiveness of osimertinib is improved.

CADTH was unable to address limitations related to the model not explicitly incorporating a possibility of cure for patients with LR and incorporating only osimertinib-related adverse events (AEs) in the first month of treatment. Addressing these limitations would likely increase the ICER, making osimertinib less cost-effective. CADTH was unable to address the potential impacts of patients being on adjuvant chemotherapies before osimertinib. If efficacy is believed to be different among those who received prior chemotherapy, then this would have an impact on the cost-effectiveness conclusions. Finally, the impact of re-treatment with osimertinib is highly uncertain due to the lack of evidence as to when patients would be re-treated and how this might impact health outcomes.

Stakeholder Input Relevant to the Economic Review

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

Patient input was received from 4 Canadian-based patient groups: the Canadian Cancer Survivor Network (CCSN), Lung Cancer Canada (LCC), the Lung Health Foundation (LHF), and the CanCertainty Coalition (CanCertainty). The CCSN and LCC conducted patient interviews. LHF collected information through an online survey, with responses from 11 patients and 2 caregivers. The LHF also conducted 3 focus groups consisting of patients and caregivers. CanCertainty’s submission was based on published lung cancer statistical reports, Canadian drug coverage, and a past 2017 survey conducted by the group. Overall, patients’ disease experience was influenced by the physical symptoms associated with lung cancer (e.g., fatigue, shortness of breath, cough) and the impact of fear of low survival on their mental health and relationships with others. The patients who responded to the survey had experience with chemotherapy, with reports of side effects, including fatigue, nausea, and appetite loss, and reporting a fear of relapsing after adjuvant chemotherapy. Patients also reported having mostly successful experiences with surgery, apart from 1 respondent whose lifestyle was affected by their reduced lung function after resection. In terms of hopes for improved outcomes with a new therapy, patients expressed a desire for a cure, delayed disease recurrence, and for their quality of life to be maintained. Patients valued DFS with improved quality of life and medications that allow them to maintain their independence. Finally, patients expressed a desire for symptom reduction and management. Patients with experience using osimertinib reported mostly tolerating the medication well, along with some side effects; for 1 patient, side effects resulted in them discontinuing treatment after 4 months. It was also noted that EGFR genetic testing is not readily available until later stages, and an expansion to earlier stages would be helpful in this area.

Input from registered clinicians was received from 2 groups: LCC and the Ontario Health (Cancer Care Ontario) Lung Cancer Drug Advisory Committee. The clinician input noted that the current care pathway for patients is dependent on their stage after resection, with a minority of patients with stage IB disease offered adjuvant chemotherapy. Fit patients with stage II disease will be offered adjuvant chemotherapy. Patients with stage IIIA disease are offered adjuvant chemotherapy concurrent with radiation followed by complete resection. Some may be offered adjuvant chemotherapy after resection. Clinicians reported that improving OS and quality of life are the main expectations for treatment, but also noted that extending DF time, especially when recurrence is symptomatic, is also valuable. There was also an expressed desire to prevent recurrence. It was noted that osimertinib adjuvant treatment does not replace adjuvant chemotherapy but is an add-on to current care.

Drug plan input considered whether patients would be eligible for re-treatment with osimertinib in the metastatic setting if they had received it as adjuvant therapy. It was also noted that EGFR genetic testing would need to be expanded to identify those eligible for adjuvant osimertinib. Finally, the drug plans noted that, as the ADAURA trial data are immature, the time patients will spend on osimertinib is uncertain, leading to uncertainty in the 3-year budget impact.

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

• DFS and health state utilities capturing lung cancer symptoms and quality of life were included.

• AEs associated with osimertinib were included in the pharmacoeconomic analysis; chemotherapy-related AEs were not.

• EGFR genetic testing costs for those receiving osimertinib were included in the analysis.

• Patients who progress to distant metastatic disease more than 48 months after initiating osimertinib were eligible for re-treatment.

• Time-to-treatment discontinuation extrapolations from the ADAURA trial were used to estimate the duration spent on therapy in the budget impact analysis (BIA).

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

• incorporating adverse event disutility while remaining on treatment.

Economic Review

The current review is for osimertinib (Tagrisso) as adjuvant therapy for adult patients (aged ≥ 18 years) with stage IB to stage IIIA non–small cell lung cancer (NSCLC) whose tumours have EGFR exon 19 deletions or a sensitizing mutation in the EGFR gene with substitution of a leucine with an arginine at position 858 in exon 21 (L858R substitution mutations).¹

Economic Evaluation

Summary of Sponsor’s Economic Evaluation

Overview

The sponsor submitted a cost-utility analysis of osimertinib adjuvant therapy after tumour resection in patients with stage IB to stage IIIA NSCLC whose tumours have EGFR exon 19 deletions or exon 21 (L858R) substitution mutations, compared with active surveillance.¹ Both osimertinib-treated and active-surveillance patients could have received adjuvant chemotherapy after resection. The model population comprised adult patients (aged ≥ 18 years) with completely resected, early-stage EGFR mutation–positive, NSCLC, which was aligned with the Health Canada indication.²

Osimertinib is available as a 40 mg or 80 mg tablet. The recommended dose of osimertinib adjuvant therapy is 80 mg orally once daily for up to 3 years or until disease recurrence or unacceptable toxicity.² At the sponsor’s submitted price of $294.68 per 80 mg tablet, the annual cost of osimertinib adjuvant therapy would be $107,557 if patients remained on therapy for a full year. No drug-acquisition costs were modelled for active surveillance, which was assumed to consist of no active treatment.¹ It was noted that some patients would receive adjuvant chemotherapy; however, as osimertinib is an add-on therapy, the cost of any adjuvant chemotherapies would be the same regardless of the decision to use osimertinib. Therefore, no adjuvant chemotherapy costs were applied in the model.

The clinical outcomes of interest were QALYs and life-years. The economic analysis was undertaken over a lifetime (38-year) time horizon from the perspective of a Canadian public health care payer. Discounting (1.5% per annum) was applied to both costs and outcomes.

Model Structure

The sponsor submitted a semi-Markov, multi-state model with 5 mutually exclusive health states with 1-month cycle lengths (4.35 weeks) (Figure 1). All patients begin in the DF health state, where they could remain DF or transition to LR or first-line treatment for distant metastatic disease (1L DM). People with LR could remain in the LR health state or transition to 1L DM. From 1L DM, patients could transition to 2L DM. Patients receiving 2L DM could only transition to death. Patients in any health state could transition to death starting in the first cycle. To incorporate time-varying probabilities in each state, the sponsor used tunnel states so that the probability of transitioning to another state was dependent on how long the patient had spent in a given health state. For example, if a patient spends 4 years in the DF state, the likelihood of them moving to LR and 1L DM falls dramatically, as a proportion of this patient’s cancer is assumed to be cured.

Model Inputs

The model’s baseline population characteristics and clinical efficacy parameters for the adjuvant treatment space were characterized by the ADAURA trial, a randomized, double-blind, placebo-controlled, multi-centre phase III study designed to evaluate the efficacy of osimertinib 80 mg once daily as adjuvant therapy following complete tumour resection with curative intent compared with placebo.³ The sponsor assumed that the ADAURA population (baseline characteristics: mean age = 62.1 years; proportion male = 30%)³ reflected the Canadian population.

Transition probabilities were derived using a variety of data sources, with data from the ADAURA trial being used to model transitions from the DF health state to LR and 1L DM health states.³ Data sources for all transitions are summarized in Table 10. The FLAURA trial, a phase III, double-blind, randomized trial comparing osimertinib with standard of care (either gefitinib or erlotinib) for the first-line treatment of locally advanced or metastatic NSCLC, was used to model transitions from the distant metastatic health states (1L DM to 2L DM and death; 2L DM to death).⁴ To model transitions from LR to 1L DM, the sponsor selected an “ADAURA-like” cohort of patients with completely resected EGFR-mutated (EGFRm) stage IA to stage IIIB NSCLC from CancerLinQ Discovery, an American database of electronic medical records.^1,5

Parametric survival modelling was used to derive health state transition probabilities, with survival distributions being selected based on clinical plausibility of long-term projections, visual inspection of fit, and the Akaike information criterion and Bayesian information criterion.¹ A competing-risks approach was taken when modelling transitions from the DF state, as the DF survival curve from ADAURA contained progression and death events. Survival distributions were then transformed into transition probabilities.

The sponsor incorporated a cure assumption for some patients in the DF health states. It was assumed that, among those receiving both adjuvant osimertinib and active surveillance alone, a proportion of patients who remain in the DF state after 4 years will be cured. Patients receiving active surveillance gradually transition to cure over 2 years, whereas this occurs over 5 years for patients who received osimertinib adjuvant treatment. This longer transition-to-cure period is to account for the 36 months patients received osimertinib, during which time no patients would transition to cure. A maximum cure point of 95% was used, meaning that 95% of patients would be assumed to be cured if they remain DF after year 5 and year 8 (for patients receiving active surveillance or osimertinib, respectively). At the point of cure, patients are no longer at risk of death or relapse due to lung cancer, so will have the same outcomes as the general Canadian population.

The model incorporated the possibility of exploring a treatment-waning effect after discontinuation of osimertinib adjuvant therapy. This was not included in the sponsor’s base case because of limited data availability after osimertinib discontinuation in the ADAURA trial.

Mortality among those in the DF state was assumed to be equal to that of the general Canadian population.¹ For those in the LR state, mortality was modelled as that of the general Canadian population, with a standardized mortality ratio of 1.26 applied.⁶ For the distant metastatic states, the sponsor reported that mortality was derived from the time-to-treatment discontinuation data from the FLAURA trial.⁴ For 1L DM to death, the time-to-treatment discontinuation could not be lower than the risk of death in the LR state (i.e., it was superseded by general population mortality with a standardized mortality ratio of 1.26 applied).¹

The grade 3 or greater AEs observed in the ADAURA trial were incorporated into the model with an associated cost and disutility.³ These are applied for the first month only to patients on treatment in the DF state; after 1 month, no additional AEs are applied.

Health state utility values were sourced from the literature. A Canadian study on patients with metastatic NSCLC receiving EGFR tyrosine kinase inhibitors was used. That study found a utility of 0.85 for people using osimertinib in the metastatic setting.⁷ The sponsor assumed that quality of life would not be worse in the DF setting and, therefore, assumed that the utility of those using osimertinib in the metastatic setting was representative of DF patients. For LR, it was assumed that the utilities would be equal to that of the DF state.¹ Utilities for patients receiving 1L DM and 2L DM were sourced from a Canadian study of patients with EGFRm NSCLC.⁷ All utility values, along with the values examined in the sponsor’s scenario analyses, are in Table 11. Disutilities for AEs were sourced from the literature^8,9 and incorporated as a single disutility as a 1-off in the first cycle.

Costs in the model included treatment-acquisition costs for adjuvant osimertinib and subsequent therapies, disease management, AEs, and EGFRm testing. Dosing for osimertinib adjuvant therapy was incorporated by multiplying the price per 80 mg tablet by the number of days in a model cycle (30.44 days) and multiplying that by an RDI of ||||, leading to per-cycle and annual costs of $8,870.55 and $106,082, respectively.¹ Treatment costs for subsequent therapies were sourced from IQVIA. Patients who had not progressed 48 months after initiating osimertinib adjuvant therapy could be re-treated with osimertinib upon progression; those who progressed less than 48 months after initiation could not.¹ The proportion of patients receiving subsequent therapies and the costs used in the sponsor’s analysis are presented in Table 12.

Disease management costs included those for routine monitoring and were based on real-world studies for LR and distant metastatic health states.¹⁰ DF disease management costs were based on a study by the sponsor using the Institute for Clinical Evaluative Sciences database.⁵ AE costs were sourced from the Ontario Case Costing Initiative and applied as a 1-time cost in cycle 1.¹¹ EGFRm testing costs (sourced from a Cancer Care Ontario report⁵) were applied in cycle 1 for patients receiving osimertinib adjuvant therapy and when patients receiving active surveillance progressed to distant metastatic disease. The sponsor did not incorporate testing costs for those patients who are tested to assess eligibility for osimertinib but test negative.

Summary of Sponsor’s Economic Evaluation Results

All analyses were run probabilistically (1,500 iterations for the base-case and scenario analyses). The deterministic and probabilistic results were similar. The probabilistic findings are presented subsequently.

Base-Case Results

Osimertinib adjuvant therapy was associated with a gain of 3.20 QALYs at an additional cost of $114,513, resulting in an ICER of $35,811 compared with active surveillance. The vast majority of the QALY benefit was derived through the life extension occurring after the trial period. The sponsor reported that, compared with active surveillance, osimertinib alone was cost-effective at a willingness-to-pay threshold of $50,000 per QALY in 66.3% of iterations.¹