CADTH Reimbursement Review

Tucatinib (Tukysa)

Sponsor: Seagen Canada Inc.

Therapeutic Area: Advanced or metastatic breast 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

Breast cancer is the most commonly diagnosed cancer among women.¹ Human epidermal growth factor receptor 2 (HER2) positivity may account for 15% to 20% of breast cancers. The development of metastases is common among patients with HER2-positive breast cancer. Brain metastases are frequently reported in patients with HER2-positive breast cancer, with some estimates suggesting that up to 30% of patients with HER2-positive breast cancer will develop metastases in the brain. Most HER2-positive breast cancers are detected early and can be treated with curable intent; however, for patients who develop metastases, treatment goals include prolonging life, controlling symptoms, limiting toxicities related to treatments, and improving or maintaining patients’ quality of life.

The treatment landscape for breast cancer has been evolving over the past 10 years and has resulted in paradigm shifts in the treatment of patients. Currently available treatments for patients with HER2-positive metastatic breast cancer (MBC) include taxane chemotherapy, trastuzumab, and pertuzumab in the first line, and trastuzumab emtansine (T-DM1) in the second line. No standard of care is currently available for the treatment of HER2-positive MBC patients in the third line. Patients who progress after second-line therapy may be treated with endocrine therapies or chemotherapies combined with HER2-targeted therapies. In particular, patients with brain metastases often do not have many effective therapy options. There is an unmet need for effective treatments for patients with HER2-positive MBCs in the third line and patients with brain metastases.

Tucatinib is a tyrosine kinase inhibitor (TKI) of the HER2 protein. Inhibition of the HER2 protein limits the growth of cancer cells.² Tucatinib, in combination with trastuzumab and capecitabine, is approved by Health Canada for the treatment of patients with locally advanced or metastatic HER2-positive breast cancer, including patients with brain metastases, who have received prior treatment with trastuzumab, pertuzumab, and T-DM1, separately or in combination. It is noted in the Health Canada product monograph that the clinical trial data supporting the effectiveness of tucatinib in combination with trastuzumab and capecitabine are limited to patients who have received at least 1 prior HER2-directed therapy in the metastatic setting.² The recommended dose of tucatinib is 300 mg orally twice daily, along with trastuzumab and capecitabine. Treatment with tucatinib should continue until disease progression or unacceptable toxicity.³ Previous CADTH reviews for advanced or metastatic HER2-positive breast cancer include T-DM1,⁴ apatinib in combination with letrozole,⁵ and eribulin mesylate.⁶ Of these previous reviews, the review for lapatinib in combination with letrozole did not receive a positive reimbursement recommendation. Of note, the CADTH review for eribulin mesylate was not specific to patients with HER2-positive cancer but was considered relevant to this review, as it can be used as a treatment option for patients with HER2-positive MBC.

The objective of this CADTH Reimbursement Review is to perform a systematic review of efficacy and safety of tucatinib in combination with trastuzumab and capecitabine in patients with locally advanced or metastatic HER2-positive breast cancer, including patients with brain metastases, who have received prior treatment with trastuzumab, pertuzumab, and T-DM1, separately or in combination.

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

Three patient groups provided input for the review of tucatinib: the Canadian Breast Cancer Network (CBCN), Rethink Breast Cancer, and the CanCertainty Coalition. Information from CBCN was obtained via online surveys. Information from Rethink Breast Cancer was obtained using an online patient survey and patient interviews. Input from CanCertainty was based on published reports on breast cancer statistics and patient drug coverage.

Quality of life was highlighted in the submitted inputs, as patients face difficulty in all aspects of life due to their condition. Patient groups stated that treatment options may vary for patients, depending on the line of therapy and patient characteristics. Trastuzumab and pertuzumab were reported by patients to be the most commonly received treatments, followed by T-DM1, capecitabine, paclitaxel, docetaxel, and trastuzumab/pertuzumab/T-DM1. Commonly reported side effects of treatment included fatigue, diarrhea, nausea, and insomnia. The patient groups identified a lack of effective treatment options for patients with brain metastases, who are typically offered local therapies including surgery and radiation.

Eight patients, 6 of whom had brain metastases, were identified as having experience with tucatinib. Commonly reported side effects due to treatment with tucatinib included diarrhea, decreased appetite, fatigue, nausea, hand-foot syndrome, and rash. In general, patients reported that side effects from tucatinib were manageable and did not negatively impact their quality of life. The patient groups highlighted that delayed progression, improved quality of life, and survival were important expectations for new treatments. Additional treatment options were also acknowledged as an important need for patients.

Clinician Input

Input From Clinical Experts Consulted by CADTH

The clinicians consulted by CADTH identified unmet treatment needs for patients with advanced or metastatic HER2-positive breast cancer, as patients lack an effective standard of care following progression on second-line therapy. In particular, patients with brain metastases lack effective systemic treatment options and are often excluded from clinical trials, resulting in significant unmet need in this patient subgroup. Tucatinib in combination with trastuzumab and capecitabine was suggested to be administered as per the HER2CLIMB trial eligibility criteria and dosing schedule, mainly in the third-line treatment setting. Tucatinib was suggested not to be used for patients with poor Eastern Cooperative Oncology Group performance status (ECOG PS) (i.e., an ECOG PS of 2 to 4); however, patients with an ECOG PS of 2 could be considered for treatment with tucatinib based on the judgment of the treating physician. As tucatinib is administered along with capecitabine and trastuzumab, the clinical experts stated that tucatinib would be administered in an outpatient clinical setting. Discontinuation of tucatinib should occur if there is evidence of disease progression or lack of benefit to patients with continued treatment, if a patient has poor performance status, or if a patient experiences severe treatment toxicity.

Clinician Group Input

Inputs from 2 clinician groups were received, 1 from the Ottawa Hospital Cancer Centre’s (OHCC) Breast Disease Site Group and the other from the Ontario Health (Cancer Care Ontario) Breast Disease Site Advisory Committee (OH-CCO B-DAC). Both groups stated that, after first-line treatment with taxane chemotherapy, trastuzumab, and pertuzumab, and second-line treatment with T-DM1, no standard third-line options are available for HER2-positive MBC patients. Third-line treatment options may differ across jurisdictions and across countries. Both groups also acknowledged there are limited treatment options for patients with brain metastases, aside from surgery and radiation. Both clinician group inputs suggested that tucatinib would be used in the third-line treatment setting. Both groups acknowledged that tucatinib-combination therapy addresses patient needs, as it demonstrated improved efficacy in patients with and without brain metastases.

Drug Program Input

Input from the CADTH pan-Canadian Oncology Drug Review (pCODR) Provincial Advisory Group identified factors pertaining to relevant comparators, generalizability, prescribing of therapy, companion diagnostics, and discontinuation criteria. The clinical experts consulted by CADTH weighed evidence from the HER2CLIMB trial and other clinical considerations to provide responses, which can be found in the Drug Program Input section.

Clinical Evidence

Pivotal Studies and Protocol Selected Studies

Description of Studies

One multi-centre, multinational, double-blind, randomized controlled phase II trial met the criteria for the CADTH systematic review protocol. The HER2CLIMB trial evaluated the efficacy and safety of tucatinib in combination with trastuzumab and capecitabine compared with placebo in combination with trastuzumab and capecitabine which, from here on, will be referred to as the tucatinib-combination group and the placebo-combination group, respectively. Eligible patients included adults with histologically confirmed HER2-positive advanced breast cancer confirmed using immunohistochemistry (IHC), in situ hybridization (ISH), or fluorescence in situ hybridization (FISH) testing. Patients must have had prior treatment with pertuzumab, trastuzumab, and T-DM1, measurable disease using Response Evaluation Criteria in Solid Tumors (RECIST) v1.1 criteria, and an ECOG PS of 0 or 1. Patients with brain metastases were also eligible for enrolment.⁷ The presence of brain metastases was identified based on medical history and screening contrast brain MRI, as assessed by an investigator.⁷ This international trial was conducted in 15 countries across 155 sites and included a total of 38 patients from Canada.^7,8 A total of 410 patients were randomized to the tucatinib-combination group and 202 patients were randomized to the placebo-combination group. Randomization was stratified according to the following: presence of brain metastases (yes versus no), ECOG PS (0 versus 1), and geographic region (US versus Canada versus the rest of the world).

The dosages for each treatment in the tucatinib-combination group were as follows:

• Tucatinib (300 mg) was administered orally twice daily.

• Capecitabine (1,000 mg/m²) was administered orally twice daily on days 1 to 14 of each 21-day cycle.

• Trastuzumab was administered with an initial loading dose of 8 mg/kg IV, after which it was administered at 6 mg/kg once every 21 days, except in specific circumstances where it was given weekly to compensate for modifications to the treatment schedule. Alternatively, trastuzumab could, after discussion with a medical monitor, have been administered at a dosage of 2 mg/kg IV every week (every 7 days), but only in circumstances where the trastuzumab infusion had been delayed and weekly infusions were therefore required to re-synchronize the cycle length to 21 days.

  • Subcutaneous use of trastuzumab was permitted; when subcutaneous trastuzumab was administered, a fixed dose of 600 mg was provided without a loading dose. Subcutaneous trastuzumab was administered once every 3 weeks, as there was no allowance for weekly administration. Crossover from IV to subcutaneous trastuzumab was permitted within the trial.

  • Where national regulatory authorities approved the use of a trastuzumab biosimilar, either IV or subcutaneous, a trastuzumab biosimilar could also be administered if considered by the investigator to be appropriate for the patient.

The treatment dosages in the placebo-combination group were the same as in the tucatinib-combination group, except that patients received placebo tablets in place of tucatinib; patients also received the placebo orally twice daily. Treatment continued until unacceptable toxicity, disease progression, withdrawal of consent, or study closure.

The primary end point of the trial was progression-free survival (PFS). Key secondary end points, which were part of a hierarchical testing scheme, included PFS among patients with brain metastases (PFS_BM) and overall survival (OS). Other secondary and exploratory end points included objective response rate (ORR), PFS assessed by investigator (PFS_INV), duration of response (DOR), and health-related quality of life (HRQoL) assessed using the EuroQol 5-Dimensions 5-Levels questionnaire (EQ-5D-5L).

Baseline characteristics of the HER2CLIMB trial were generally well balanced across both treatment groups in both the intention-to-treat (ITT) and ITT-PFS populations; baseline characteristics were also similar across both trial populations. In the ITT population, patients had a mean age of 54 years, with most patients (> 80%) being less than 65 years of age. Most patients were White (74%) and from the US (54%) or the rest of the world (40%). Relatively equal proportions of patients had an ECOG PS of 0 (48%) or 1 (51%).⁷ The majority of patients had metastatic disease (≥ 99%) and were positive for at least 1 hormone receptor (61%) or negative for both (38%). Non–central nervous system (CNS) metastases were reported among 98% of patients, with the most frequent metastasis sites being lung (49%), bone (55%), and liver (36%). Brain metastases were reported in 48% of patients.⁷ A mean of 4 lines of prior therapy were reported by all patients in both treatment groups, with a mean of 3 prior therapies specifically in the metastatic setting. As per eligibility criteria, all patients (100%) had received prior treatment with trastuzumab and T-DM1, and more than 99% of patients had also received prior therapy with pertuzumab.

Efficacy Results

Key efficacy results for the HER2CLIMB trial are reported in Table 2; they are based on a data cut-off date of September 4, 2019 and are considered to be the final analyses results. Results of the primary end point, PFS (stratified HR = 0.54; 95% confidence interval [CI], 0.42 to 0.71; stratified log-rank P value, < 0.00001), and key secondary end points, PFS_BM (stratified HR = 0.48; 95% CI, 0.34 to 0.69, stratified log-rank P value, < 0.00001) and OS (HR = 0.66; 95% CI, 0.50 to 0.88; stratified log-rank P value, 0.00480), indicated a statistically significant improvement in patients treated with the tucatinib-combination over the placebo-combination therapy. ORR was considered another secondary end point and also supported the results of the primary and key secondary analyses showing improved efficacy with the tucatinib-combination versus the placebo-combination treatment. A post-hoc analysis was conducted by the sponsor that provided an additional 15.6 months of follow-up. The post-hoc analysis provided updated data for OS and PFS, assessed among all randomized patients. Results of the post-hoc analyses continued to support trastuzumab-combination over placebo-combination therapy.⁹ The assessments conducted at the post-hoc analysis were not formally tested; therefore, they should be considered descriptive. HRQoL data did not indicate any differences in EQ-5D-5L scores between patients in the tucatinib- and placebo-combination groups.

Harms Results

Safety data were reported based on a data cut-off date of September 4, 2019. In general, adverse events (AEs) were reported more commonly among patients in the tucatinib-combination group. The most common AEs of any grade in both the tucatinib-combination group and the placebo-combination group were diarrhea (80.9% versus 53.3%), hand-foot syndrome (63.4% versus 52.8%), nausea (58.4% versus 43.7%), fatigue (45.0% versus 43.1%), and vomiting (35.9% versus 25.4%); however, the proportion of patients experiencing these AEs was greater in the tucatinib-combination group.⁷ A total of 223 patients (55.2%) in the tucatinib-combination group experienced a grade 3 or greater AE compared with 96 patients (48.7%) in the placebo-combination group. In both the tucatinib-combination group and placebo-combination group, the most commonly reported AEs that were grade 3 or greater were hand-foot syndrome (13.1% versus 9.1%) and diarrhea (12.9% versus 8.6%).⁷ A time-at-risk exposure-adjusted analysis of grade 3 or greater AEs for hand-foot syndrome, diarrhea, and increase in alanine aminotransferase (ALT) and aspartate aminotransferase (AST) were performed to adjust for the longer exposure to treatment that patients in the tucatinib-combination group experienced, as these patients had a longer duration of treatment than patients in the placebo-combination group. After adjustment, the crude incidence of grade 3 or greater AEs of hand-foot syndrome (13.1% versus 9.1%), diarrhea (12.9% versus 8.6%), and ALT (5.4% versus 0.5%) and AST increase (4.5% versus 0.5%) was higher, for all of these AEs, in the tucatinib-combination group than in the placebo-combination group, respectively; the time-at-risk exposure-adjusted incidence rate per 100 person-years was 21 versus 19, 21 versus 17, 8 versus 1, and 7 versus 1, respectively. Serious adverse events (SAEs) of any grade were reported in a similar proportion of patients in the tucatinib- and placebo-combination groups (25.7% and 26.9%, respectively). Grade 5 AEs were reported in 8 patients (2.0%) in the tucatinib-combination group and 6 patients (3.0%) in the placebo-combination group.

Critical Appraisal

The HER2CLIMB trial was an international, multi-centre, double-blind, placebo-controlled, phase II randomized controlled trial. The baseline demographic and clinical characteristics were balanced across the treatment groups overall and across important analysis populations (i.e., ITT and ITT-PFS populations). Patients were randomized based on the presence of brain metastases (yes versus no), ECOG PS (0 versus 1), and geographic region (US versus Canada versus the rest of the world). This helped to ensure the comparability of the subgroup analysis results between treatment arms, according to each pre-specified stratification factor. The sponsor also included specifications for using a biased-coin assignment in the randomization scheme to prevent imbalances between treatment groups and any given hierarchical level (i.e., ensure overall treatment group balance and balance within each stratification factor in each treatment group).

Results of the HER2CLIMB trial demonstrated statistically significantly improved OS and PFS among patients treated in the tucatinib-combination group compared with the placebo-combination group. In general, subgroup analyses favoured treatment with the tucatinib-combination group versus the placebo-combination group. However, it should be acknowledged that while the subgroups for the subgroup analyses were pre-specified, they were not adjusted for multiplicity, not powered to detect differences, and could be indicative of imprecision due to wide CIs. The lack of adjustment for subgroup analyses may increase the likelihood of type I error, resulting in an increased likelihood of detecting a treatment effect when 1 may not be present. The sponsor conducted a post-hoc analysis that provided 15.6 months of additional follow-up (resulting in a total of 29.7 months of follow-up for the tucatinib-combination group and 29.4 months of follow-up for the placebo-combination group), and provided additional efficacy (OS, PFS) and safety data. Of note, after the primary analysis, the trial was unblinded and assessments for PFS were conducted by the investigator. The results of the post-hoc analysis were consistent with the results of the primary analysis, which remained blinded and which used a blinded independent central review (BICR) for assessment of PFS.

It is possible that the choice of subsequent therapies could have affected the efficacy assessments of OS, as the analyses for OS included patients who received subsequent therapies. A total of 202 patients (69.2%) in the tucatinib-combination group and 139 patients (79.4%) in the placebo-combination group received subsequent anti-cancer therapies. There were disproportional differences noted between treatment groups in the types of subsequent anti-cancer therapies received, as more patients in the placebo-combination group received antibody (57.1% versus 50.0%, respectively) and TKI (24.0% versus 16.8%) anti-HER2 regimens and trastuzumab (12.2% versus 5.4%), while more patients in the tucatinib-combination group than in the placebo-combination group received trastuzumab plus chemotherapy (20.8% versus 15.8%, respectively). The differences in subsequent therapies are expected to have introduced bias in the efficacy analyses of OS and other patient outcomes. However, the direction and extent of the biases are difficult to predict.

Standard first-line therapies for patients with MBC may include treatment with pertuzumab in combination with trastuzumab and taxane followed by pertuzumab plus trastuzumab. Second-line therapies for these patients may then include T-DM1. Eligibility criteria in the HER2CLIMB trial specified that all patients must have had prior treatment with trastuzumab, pertuzumab, and T-DM1. Therefore, the population of patients in the HER2CLIMB trial is likely reflective of patients in the Canadian population and the treatment algorithms that are standard in Canadian clinical practice. Prior treatment with trastuzumab, T-DM1, and pertuzumab was not required to have taken place specifically in the metastatic setting. Although most patients did receive each drug in the metastatic setting, some patients received it in both neoadjuvant or adjuvant and metastatic settings; few patients received prior therapy in the neoadjuvant or adjuvant setting only. The sponsor noted that the treatment landscape for HER2-positive breast cancer patients has changed drastically since the completion of patient enrolment for the HER2CLIMB trial.¹⁰ During patient enrolment, T-DM1 was approved for and used only in the metastatic setting; however, since completion of patient enrolment, T-DM1 has been approved for use in the adjuvant setting. Almost all patients in the HER2CLIMB trial (> 98%) reported having received prior therapy with T-DM1 in the metastatic setting only.⁷ It is expected that a greater proportion of patients in clinical practice will have received prior therapy with T-DM1 in other treatment settings as well.

In the Health Canada–approved product monograph, tucatinib in combination with trastuzumab and capecitabine is indicated for patients who have received at least 1 prior HER2-directed therapy in the metastatic setting.² The treatment landscape for patients with MBC is complex and has changed to include new HER2-directed treatments, such as pertuzumab and T-DM1. Patients in the HER2CLIMB trial reported having received a mean of 3 prior therapies in the metastatic setting, and the sponsor confirmed that every patient in the HER2CLIMB trial received at least 1 prior therapy in the metastatic setting.^3,7 Therefore, given the changes to the treatment landscape for this setting and the characteristics of patients in the HER2CLIMB trial, it was considered appropriate for tucatinib, in combination with trastuzumab and capecitabine, to be used for patients who received at least 1 HER2-targeted therapy in the metastatic setting.

The HER2CLIMB trial eligibility criteria required patients to have prior treatment with trastuzumab, pertuzumab, and T-DM1, alone or in combination, and most patients (> 90%) reported having received each treatment. The median and mean number of therapies used among patients in the HER2CLIMB trial was 4, with most patients having received trastuzumab, pertuzumab, and T-DM1 in either the metastatic setting or in the metastatic and neoadjuvant or adjuvant setting. Therefore, patients would have received tucatinib-combination therapy in the second- or later-line setting. It may be unreasonable to suggest using tucatinib-combination therapy as a first-line treatment option for patients with MBC, as there is no evidence to support the use of this treatment in this context. The input received from the clinical expert consulted by CADTH and the Canadian clinician groups that provided input on this submission suggests that tucatinib-combination therapy would most likely be used as a third-line therapy.

Indirect Comparisons

Description of Studies

The sponsor-submitted indirect treatment comparison (ITC) compared the efficacy of tucatinib in combination with trastuzumab and capecitabine with relevant comparators, including lapatinib plus capecitabine, margetuximab plus capecitabine, neratinib, neratinib plus capecitabine, pertuzumab plus trastuzumab and capecitabine, trastuzumab plus capecitabine, capecitabine, T-DM1, and T-DM1 plus capecitabine, among patients with HER2-positive MBC who had received at least 1 prior therapy. The ITC was conducted using a network meta-analysis (NMA) that included 14 phase II and III trials identified by a systematic literature search.

Efficacy Results

Regarding PFS, the NMA results suggested that tucatinib-combination treatment was favoured compared with capecitabine monotherapy (HR = 0.33; 95% credible interval [CrI]: 0.23 to 0.47; P < 0.0001), neratinib (HR = 0.47; 95% CrI, 0.30 to 0.71; P = 0.0007), lapatinib plus capecitabine (HR = 0.55; 95% CrI, 0.40 to 0.76; P = 0.0003), trastuzumab plus capecitabine (HR = 0.53; 95% CrI, 0.42 to 0.68; P < 0.0001), and pertuzumab plus trastuzumab plus capecitabine (HR = 0.65; 95% CrI, 0.47 to 0.90; P = 0.0110). No differences were shown between the tucatinib combination and margetuximab plus capecitabine, neratinib plus capecitabine, T-DM1, and T-DM1 plus capecitabine.¹¹

Regarding OS, the NMA results suggested that the tucatinib-combination treatment was favoured compared with capecitabine monotherapy (HR = 0.45; 95% CrI, 0.27 to 0.77; P < 0.0017), neratinib (HR = 0.47; 95% CrI, 0.27 to 0.80; P = 0.0073), lapatinib plus capecitabine (HR = 0.59; 95% CrI, 0.41 to 0.83; P = 0.0030), and trastuzumab plus capecitabine (HR = 0.66; 95% CrI, 0.50 to 0.88; P = 0.0040). No differences were shown between the tucatinib combination and margetuximab plus capecitabine, neratinib plus capecitabine, pertuzumab plus trastuzumab plus capecitabine, and T-DM1.¹¹

Regarding ORR, the tucatinib-combination therapy was favoured over capecitabine (HR = 0.90; 95% CrI, 0.48 to 1.31; P < 0.0001), neratinib (HR = 0.82; 95% CrI, 0.29 to 1.33; P = 0.0010), and trastuzumab plus capecitabine (HR = 0.39; 95% CrI, 0.18 to 0.60; P = 0.003). There were no differences between tucatinib-combination therapy and lapatinib plus capecitabine, neratinib plus capecitabine, pertuzumab plus trastuzumab plus capecitabine, T-DM1, and T-DM1 plus capecitabine.¹¹

Harms Results

No comparisons for harms or safety were incorporated in the sponsor’s ITC.

Critical Appraisal

The sponsor’s ITC included both phase II and III trials. Some phase II trials were not powered to detect differences between treatment groups, which may have affected the precision of the treatment estimates obtained from those studies. The inclusion of such studies in the sponsor’s ITC may have introduced uncertainty into the comparisons made within the network. The treatment crossover reported in trials is likely to have introduced bias into the comparisons of the ITC, as crossover is likely to have diluted estimates of treatments with investigational therapies. In addition, the differences in the patient characteristics across the studies introduce uncertainty regarding the comparability of patients across trials. For example, patients receiving treatment in later lines of therapy are likely to have worse clinical outcomes, as they have already progressed to more therapies than patients in earlier lines of treatment. Further, there were differences in patients’ ECOG PS, hormone receptor status, and presence of brain metastases. The sponsor’s ITC included trials published between 2008 and 2020. Due to changes in treatment paradigms for HER2-positive MBC, it is highly likely that patients across studies are not comparable due to the changing treatment landscapes, which would have affected overall patient outcomes over time. There were some methodological limitations, as some trials reported violation of the proportional hazard assumption and there was a lack of available data to incorporate relevant effect modifiers.

Other Relevant Evidence

The publication by Lin et al.¹² reported exploratory analyses of intracranial efficacy and survival in a subgroup of patients with brain metastases from the pivotal HER2CLIMB study. Patients with brain metastases were classified as follows:

• treated and stable (prior local treatment and no evidence of progression at baseline brain MRI, including patients treated during the screening period)

• treated and progressing (prior local treatment but evidence of progression of existing lesions, new lesions, or untreated lesions remaining after prior treatment at baseline brain MRI)

• untreated (no prior local treatment)

A total of 198 patients randomized to the tucatinib-combination group and 93 patients randomized to the placebo-combination group had brain metastases. The interventions have been previously described for the HER2CLIMB study. Treatment with dexamethasone (up to 2 mg per day) was permitted to control symptoms of brain metastases. The majority of patients were older than 65 years (83.5%); 60.8% resided in North America and 93.9% had non-CNS metastatic disease. Regarding ECOG PS, 44.7% of patients had a score of 0, 55.3% had a score of 1, and 57.0% of patients were hormone receptor–positive. The brain metastasis treatment status at baseline was treated and stable, treated and progressing, or untreated for 40.2%, 37.1%, and 22.7% of patients, respectively. Most patients (70.1%) had prior radiation therapy for brain metastases, 41.9% had whole-brain radiation therapy, 42.6% had targeted radiation therapy, and 15.8% had surgery.

The treatment groups were well balanced by baseline characteristics with the exception of the proportion of patients who were hormone receptor–positive (54.0% in the tucatinib-combination group versus 63.4% in the placebo-combination group), patients with an ECOG PS score of 1 (53.5% in the tucatinib-combination group versus 59.1% in the placebo-combination group), and patients with a history of prior targeted radiation therapy (46.5% in the tucatinib-combination group versus 34.4% in the placebo-combination group).

Efficacy Results

For patients treated in the tucatinib-combination group, 40.2% (95% CI, 29.5 to 50.6) of patients with brain metastases, 35.0% (95% CI, 23.2 to 47.0) of patients with active brain metastases, and 53.3% (95% CI, 31.4 to 71.0) of patients with stable brain metastases had CNS-PFS at 1 year. None of the patients receiving the placebo-combination had CNS-PFS at 1 year. A hazard ratio (HR) of 0.32 (95% CI, 0.22 to 0.48) was reported for the tucatinib-combination group compared with the placebo-combination group in all patients with brain metastases. Similar results were reported for patients with active brain metastases (HR = 0.36; 95% CI, 0.22 to 0.57) and patients with stable brain metastases (HR = 0.31; 95% CI, 0.14 to 0.67).

Among all patients with brain metastases, 1-year OS was reported for 70.1% (95% CI, 62.1 to 76.7) of patients in the tucatinib-combination treatment group and 46.7% (95% CI, 33.9 to 58.4) of patients in the placebo-combination treatment group. For patients with active brain metastases, 1-year OS was reported for 71.7% (95% CI, 61.4 to 79.7) and 41.1% (95% CI, 25.5 to 56.1) of patients randomized to the tucatinib- and placebo-combination groups, respectively. For patients with stable brain metastases, 1-year OS was reported for 67.6% (95% CI, 53.8 to 78.0) and 55.6% (95% CI, 34.1 to 72.6) of patients randomized to the tucatinib- and placebo-combination groups, respectively. These data for 1-year OS corresponded to an HR of 0.58 (95% CI, 0.40 to 0.85) for all patients with brain metastases, 0.49 (95% CI, 0.30 to 0.80) for patients with active brain metastases, and 0.88 (95% CI, 0.45 to 1.70) for patients with stable brain metastases.

Intracranial response was also reported for patients with active brain metastases and measurable intracranial lesions at baseline.

Harms Results

Safety outcomes were not assessed in this study.

Critical Appraisal

Information about reasons for or timing of discontinuation from treatment was not available. The proportion of patients who were hormone receptor–positive and who had a history of prior targeted radiation therapy was greater in the tucatinib treatment group, which may bias the results for PFS and OS against tucatinib. Additionally, a greater proportion of patients had received prior targeted radiation therapy in the tucatinib treatment group, which may also indicate bias against tucatinib. The analyses were exploratory and the statistical tests could not be interpreted as statistically significant. Lastly, CNS target lesions were assessed by the investigator and not externally validated. Issues of generalizability for the overall HER2CLIMB study also apply to the exploratory analyses described here. This study or exploratory analysis was specific to patients with brain lesions, which were identified using MRI and which is consistent with Canadian clinical practice. Trastuzumab was available for administration intravenously or subcutaneously; however, the Lin et al. study did not provide this level of detail for patients in the post-hoc analyses.

Conclusions

The results of a randomized phase II trial demonstrated significant improvement in PFS and OS. In particular, the improvement was observed in PFS among a subgroup of patients with brain metastases, a group of patients with limited effective treatment options available. HRQoL outcomes were reported to be important for patients; however, there were no differences observed. Despite the limitation of an early clinical development phase, the available evidence supports the comparative efficacy of tucatinib-combination therapy over trastuzumab plus capecitabine alone in the treatment of patients who are often difficult to treat if following the current treatment pathway in Canadian practice. There is a lack of evidence on the comparative effectiveness of other alternative therapies. One sponsor-submitted ITC suggested that tucatinib-combination therapy may be more efficacious than capecitabine alone, neratinib, lapatinib plus capecitabine, and trastuzumab plus capecitabine. However, the ITC had significant limitations that introduced uncertainty about the overall results. Specifically, the sponsor’s ITC included studies that reported differences in trial characteristics, differences in patient characteristics, lack of adjustment for relevant effect modifiers (e.g., prior exposure to treatments, line of therapy, and presence of brain metastases), and violation of the proportional hazard assumption, in particular, for PFS. While biases introduced in the sponsor’s ITC may introduce uncertainty in the magnitude of the estimates between treatments, the overall direction of estimates was considered reliable. Specific AEs were reported more frequently in the tucatinib-combination group, even though the overall AEs appeared to be similar based on the most updated results. Further study of AEs or SAEs may be required.

Introduction

Disease Background

Breast cancer is the most commonly diagnosed cancer among women in Canada. In 2020, it was expected to be the second-most commonly diagnosed cancer, with an estimated 27,400 new cases expected. Breast cancer was further expected to result in approximately 6% of all cancer-related deaths.¹ While not as common, breast cancers can also develop in men.¹³ Breast cancers are heterogeneous and can be categorized into subtypes, 1 of which relies on HER2 receptor status. Patients with HER2-positive breast cancer may further be distinguished by hormone receptor status, referring to the activation of estrogen and/or progesterone receptors. Based on input from the clinical experts consulting with CADTH for this review, HER2-positive breast cancers account for 15% to 20% of breast cancers; when HER2-positive breast cancers are diagnosed early, the majority of patients will be cured. A certain proportion of patients will develop metastatic disease, which is generally incurable.

Patients with HER2-positive breast cancer have a high likelihood of developing metastases. In particular, brain metastases among patients with HER2-positive breast cancer is a common occurrence; some estimates suggest that approximately 30% of patients with HER2-positive breast cancer will involve metastases in the brain.¹⁴ The prognosis of patients with brain metastases remains poor, despite advancements in treatments for breast cancer patients.¹⁵ In general, patients with recurrent breast cancer or MBC have a poor prognosis. Treatment options for such patients require considering clinical benefits and limiting harms associated with treatment. Therefore, treatment goals for patients with metastatic disease are to prolong life, control symptoms, limit toxicities related to treatments, and improve or maintain patients’ quality of life.¹⁴

Standards of Therapy

Treatment strategies may differ according to the subtype of breast cancer that patients present with and involve an interdisciplinary approach that can include both locoregional treatment (i.e., surgery and radiation) and systemic therapies.¹⁴ Input from the clinical experts consulted by CADTH indicated that, after a patient develops MBC, options for systemic therapy depend on the prior treatments the patient has received in adjuvant or neoadjuvant settings, as well as the amount of time that has elapsed between when they received their last treatment and when they developed metastatic disease. Other patient characteristics, such as functional status, comorbidities, and patient preference are incorporated into the decision-making for systemic treatment options. The clinical experts consulted by CADTH highlighted that patients who have not received prior adjuvant therapy with trastuzumab, and patients whose disease recurs after 1 year of completing adjuvant treatment, may typically be treated with a taxane chemotherapy combined with trastuzumab and another HER2 monoclonal antibody (e.g., pertuzumab).

The clinical experts suggested that patients are often treated with T-DM1, based on results observed in the EMILIA study.¹⁶ Input from 1 clinical expert acknowledged that treatment paradigms may be changing, as some patients may receive T-DM1 in the adjuvant setting or after receiving neoadjuvant trastuzumab and chemotherapy for early HER2-positive breast cancer. Optimal treatment for patients with advanced or metastatic HER2-positive breast cancer in the third line remains unclear. The clinical experts identified that current treatment options for patients include: capecitabine monotherapy, lapatinib plus capecitabine, trastuzumab plus capecitabine, trastuzumab plus a chemotherapeutic drug (e.g., vinorelbine), or chemotherapy. Patients with endocrine receptor–positive HER2-positive breast cancer may also be treated with trastuzumab combined with endocrine therapy or trastuzumab plus pertuzumab. The clinical experts identified that most patients who are endocrine-sensitive receive upfront treatment with HER2-targeted therapies and chemotherapy unless they have minimal disease or are unable to tolerate chemotherapy.

The clinical experts noted that drugs without Health Canada approval are not usually accessible in clinical practice and would be considered for patients only in the setting of a clinical trial (e.g., trastuzumab deruxtecan or margetuximab). Some HER2-targeted therapies were stated to be available to patients through special access programs, for example, neratinib, a pan-HER2 TKI, in combination with capecitabine.

The clinical expert consulted by CADTH acknowledged that expectations of a cure for patients with metastatic disease are sometimes not feasible. Therefore, treatment goals for patients with metastatic disease were stated to include prolonged survival, optimization of quality of life, and reduction of disease symptoms. Delayed progression of disease and minimal side effects were also stated to be important considerations for patients’ treatment.

Drug

Tucatinib is indicated by Health Canada to be used in combination with trastuzumab and capecitabine for the treatment of patients with locally advanced unresectable or metastatic HER2-positive breast cancer, including patients with brain metastases, who have received prior treatment with trastuzumab, pertuzumab, and T-DM1, separately or in combination.² As per the Health Canada indication, tucatinib in combination with trastuzumab and capecitabine should also be limited to patients who have received at least 1 prior HER2-directed therapy in the metastatic setting.²

Tucatinib is a TKI of the HER2 protein that promotes the growth of cancer cells. Tucatinib is selective for HER2 and results in inhibition of downstream cell signalling and proliferation; this inhibition ultimately results in the death of HER2-driven tumour cells.²

The Health Canada–recommended dosage for tucatinib is 300 mg (two 150 mg tablets) taken orally twice daily in combination with trastuzumab and capecitabine until disease progression or unacceptable toxicity.² Dosages for trastuzumab and capecitabine should be as follows:

• Capecitabine: 1,000 mg/m² taken orally twice daily on days 1 to 14 every 21 days.

• Trastuzumab: Administered via IV. It should be administered with an initial loading dose of 8 mg/kg followed by 6 mg/kg every 21 days. Subcutaneous trastuzumab should be administered at 600 mg every 21 days, without a loading dose.

The sponsor has requested reimbursement of tucatinib as per the indication under review, which is in combination with trastuzumab and capecitabine for the treatment of patients with locally advanced unresectable or metastatic HER2-positive breast cancer, including patients with brain metastases, who have received prior treatment with trastuzumab, pertuzumab, and T-DM1, separately or in combination. Tucatinib has not been previously reviewed by CADTH.

A table describing key characteristics of commonly used treatments for HER2-positive MBC is presented in Table 3.

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

Three patient groups provided input for the review of tucatinib: CBCN, Rethink Breast Cancer, and the CanCertainty Coalition. CBCN is a patient-directed, national health charity committed to ensuring the best quality of care for all Canadians affected by breast cancer through the promotion of information and education and advocacy activities. Rethink Breast Cancer’s mission is to empower young people who are concerned about and affected by breast cancer through education, support, and advocacy. The CanCertainty Coalition represents more than 30 Canadian patient groups, cancer health charities, and caregiver organizations from across the country that collaborate with oncologists and other cancer care professionals to significantly improve the affordability and accessibility of cancer treatment.

The CBCN’s submission included information from 2 online surveys of patients living with MBC and their caregivers conducted in 2012 (N = 87) and 2017 (N = 180). The 2012 survey was conducted in collaboration with Rethink Breast Cancer and was completed by 71 patients and 16 caregivers. In the 2017 survey, 36 respondents (20%) identified as being HER2-positive, according to CBCN. Patients were contacted through the membership databases of CBCN and other patient organizations. CBCN also conducted phone interviews in 2020 and 2021 with 2 Canadian patients living with HER2-positive MBC who had direct experience with tucatinib. They also reviewed published information to identify issues and experiences that are commonly shared among many women living with breast cancer.

Rethink Breast Cancer contracted a freelance writer to help develop the survey used for data collection for their patient input submission and to analyze the findings of the survey and interviews. Data were collected through an online patient survey conducted between March 2 and April 7, 2021. Respondents for the survey were actively identified through messages to Rethink Breast Cancer’s mailing list, a private Facebook group, and a partner organization, as well as passively through posts to social media and online discussion forums. The survey included 51 patients with HER2-positive locally advanced unresectable breast cancer or MBC from across Canada (73%), the US (24%), and Mexico (2%); 1 patient chose not to answer. Of the 51 respondents, 21 were receiving first-line treatment, 5 were receiving second-line treatment, 9 were receiving third-line treatment or higher, 8 were receiving treatment after recurrence, 2 were under surveillance following treatment, 3 had no evidence of disease, and 3 reported they were in a different phase of treatment at the time of completing the survey. A total of 14 respondents reported also having brain metastases along with their MBC. Six of the respondents had experience with tucatinib, and 5 of the patients agreed to further participate in telephone interviews; of these patients, 3 were from the US and 3 were from Canada.

CanCertainty developed its submission based on published reports relating to breast cancer statistics in patients under the age of 65 who have developed brain metastases and are without either private or public Canadian drug coverage. In addition, CanCertainty estimated current incidence rates for breast cancer for each Canadian province based on incidence rates dating from 2017 from the Canadian Cancer Registry. CanCertainty was also able to estimate the number of HER2-positive MBC patients under 65 years who would be left without private or public drug coverage. Information regarding patient-reported disease and treatment, or patient expectations for improved outcomes, was not provided in this patient input submission.

Disease Experience

The CBCN described MBC as the spread of cancerous cell growth to areas of the body other than where the cancer first formed and which is often more severe. They noted it most commonly spreads to the bones but can include the lungs, liver, brain, and skin. According to the patient group, the current treatment options for MBC are only effective at prolonging progression-free disease; most cases of advanced disease will progress and symptoms will worsen. Patients with a diagnosis of MBC understand the limitations of current treatment options and seek to live their remaining months and years with the best possible quality of life they can achieve.

The CBCN reported that the symptoms, progression, and experience with disease varies among patients living with MBC, but many of the effects have a significant or debilitating impact on quality of life. In a survey conducted by the CBCN in 2012, patients were asked to choose from the following options regarding their treatment experience: very acceptable, somewhat acceptable, or not acceptable. Patients were also asked to determine the impact of their treatments with the following options: significant impact, some impact, or no impact; or not important, important, somewhat important, or very important. From the survey, patients with MBC and caregivers reported that fatigue, insomnia, and pain had a notable impact on quality of life. More specifically, 54% of patients reported that fatigue resulted in a significant or debilitating impact, and 40% reported some or moderate impact; 39% of patients reported that insomnia resulted in a significant or debilitating impact, and 46% reported some or moderate impact; 37% of patients reported that pain resulted in a significant or debilitating impact, and 44% reported some or moderate impact. These results were reinforced when the survey was conducted again in 2017.

The CBCN also reported that the impact of living with MBC extends across all aspects of a patient’s life, restricting an individual’s employment and career, ability to care for children and dependents, and their ability to socialize and meaningfully participate in their community. In general, based on the 2012 survey, most patients were negatively impacted in their ability to work or to maintain a career, engage in familial responsibilities, and participate in social activities. Specifically, 71% of employed patients identified significant restrictions on their ability to work; among those with children or dependents, 21% and 53% identified significant and some or moderate restrictions, respectively, to their caregiving responsibilities; and 49% and 38% of patients identified significant and some or moderate restrictions, respectively, in their ability to exercise. Further, 42% of patients identified significant restrictions and 42% identified some or moderate restrictions on their ability to pursue hobbies and personal interests; 41% of patients identified significant restrictions and 41% identified some or moderate restrictions on their ability to participate in social events and activities; and 22% of patients identified significant restrictions and 52% identified some or moderate restrictions on their ability to spend time with loved ones.

Other experiences identified by patients, as noted by the CBCN, included: guilt, the feeling of being a burden on caregivers, fear of death, poor body image, not knowing what functionality will be lost, fear of the impact of cancer and the effect of the loss of a parent on children, not knowing what will happen to children, the loss of the support of loved ones, and marital stress or loss of fidelity and affection from their husband.

Experience With Treatment:

According to the CBCN, the goals of current treatment options for MBC include controlling the progression of the disease (extending life) and reducing cancer-related symptoms (extending or stabilizing quality of life). They also noted that treatment options and effectiveness vary depending on the type of cancer, location of the cancer, and how symptoms are experienced. The first line of treatment for patients with HER2-positive MBC is trastuzumab plus pertuzumab and a taxane, followed by second-line T-DM1. Similarly, patients who responded to Rethink Breast Cancer’s survey reported that trastuzumab and pertuzumab were the most common forms of treatment, with 96% and 88% of patients having reported experience with the 2 treatments, respectively, followed by T-DM1 (24%), capecitabine (20%), paclitaxel (18%), docetaxel (8%), and trastuzumab/pertuzumab/T-DM1 (8%). Additionally, respondents reported experience with 17 other treatments (each reported by no more than 3 patients). No therapies are considered standard of care for patients who progress after second line.

According to the Rethink Breast Cancer survey, fatigue was the most commonly reported side effect of treatments (86%), followed by diarrhea (71%), nausea (49%), and insomnia (45%). Diarrhea and fatigue were most commonly cited by respondents as the most difficult to tolerate side effects of these treatments. Loss of appetite, neuropathy, skin problems, and breathing difficulties were also cited by multiple respondents. Of note, respondents to the CBCN surveys indicated a willingness to tolerate treatment side effects such as fatigue, nausea, depression, problems with concentration, memory loss, diarrhea, and insomnia, as well as pain, to a certain degree, as survival and effectiveness and quality of life were noted as the most important factors when considering treatment of disease.

The CBCN reported that, specifically in the case of brain metastases, while systemic therapies have improved, incidence rates of brain metastases in breast cancer patients have increased, developing in about half of patients. The CBCN’s 2017 survey reported that 12% of metastatic patients who responded had metastases to their brain while 20% reported metastases to other body parts. Rethink Breast Cancer noted that 27% of respondents reported brain metastases from their breast cancer. Effective treatment options for HER2-positive breast cancer in patients with brain metastases are limited; treatment options include local therapies such as neurosurgical resection and stereotactic radiation therapy. Trial data showing the effectiveness of some of these systemic therapies are lacking.

A majority (69%) of respondents to Rethink Breast Cancer’s survey did not have difficulty accessing treatment; however, 22% reported they were unable to access treatment because it was unavailable in Canada. CanCertainty Coalition and the CBCN described a substantial financial barrier to accessing treatments, and Rethink Breast Cancer stated that 28% of respondents reported they needed financial assistance due to the costs associated with breast cancer. The submission from CanCertainty detailed the complexities of navigating care for HER2-positive MBC across Canada and the costs associated with it. Both CanCertainty and the CBCN highlighted that, in addition to the costs of treatment, patients also need to finance additional treatments for symptoms and side effects, the time required to travel to access treatments, and time off work due to appointments or symptoms and side effects. In particular, CanCertainty highlighted oral therapies that are not covered by all of the provincial plans and may require payment from patients. CanCertainty also noted that for patients without access to private health insurance, payment for treatment puts these patients at risk of severe financial burden. With the emergence of more oral oncological therapies, payment for treatment, or co-payments for those with insurance, may result in significant financial burden. As a consequence, the patient groups noted quality of life is significantly impacted and patients may end up taking less medication than prescribed because of the costs of treatment. One patient reported that this is “always a concern, as you never know if the next drug will be covered or how long it takes to get approval from private coverage. Many times it delays treatment and this weighs on one’s mind.”

Six respondents to Rethink Breast Cancer’s survey and 2 patients interviewed by CBCN had experience with tucatinib. Patients who were interviewed by CBCN were described as having different levels of experience with tucatinib. Notably, 1 patient was diagnosed with brain metastases and the other had lung, lymph node, and rib metastases. Patients from Rethink Breast Cancer’s submission reported having received tucatinib for less than 3 months (n = 3), 3 to 6 months (n = 2), and between 6 and 12 months (n = 1). Two patients reported receiving tucatinib as a second-line treatment, and 4 patients reported receiving it as a third-line treatment or higher. All but 1 of the patients reported brain metastases. Both patient groups reported that patients felt that the side effects from tucatinib were preferable to the side effects they had experienced on other treatments and therapies, and the side effects that did occur were minimal and manageable. As 1 patient described, “there was nothing not acceptable to me. They were mild based on all of the other lines experienced and it’s a small [price] to pay for the extension of life.” However, 1 patient responding to Rethink Breast Cancer’s survey reported discontinuation of treatment due to side effects, and another reported a dosage reduction due to side effects. The most commonly reported side effects were diarrhea, decreased appetite, fatigue, nausea, hand-foot syndrome, and an itchy rash. Despite this, patients generally felt that tucatinib did not negatively impact their quality of life and some patients reported they have been able to maintain their lifestyle, taking care of their home and family and travelling. Patients also positively reflected on the availability of tucatinib as an option for treatment of brain metastases, as no treatments are currently indicated for this.

Improved Outcomes

The CBCN reported that extension of PFS is a critical concern for metastatic patients. The patient group stated that, “patients living with MBC are aware that their advanced disease will progress with worsening symptoms until death, and embrace opportunities to try new treatments, even if benefits may be as little as a 6-month extension of progression-free disease.” They also emphasized the impact that delaying progression of disease and relieving cancer-related symptoms has on the improvement of quality of life, and the desire for patients to be able to provide care for families, continue with employment, and engage in social activities and personal interests. Maintaining quality of life was also rated as very important to patients from Rethink Breast Cancer’s submission, but long-term outcomes such as control of disease and prevention of recurrence and, ultimately, survival were reported as being of utmost importance to this group of patients as well. This was reiterated by 1 of the patient respondents, who reported that “survival rate chances are paramount, followed by quality of life.” Feedback from the CBCN and Rethink Breast Cancer submissions also indicated a desire among patients for targeted treatments and the importance of having treatment options.

Clinician Input

Input From Clinical Experts Consulted by CADTH

All CADTH review teams include at least 1 clinical specialist with expertise regarding the diagnosis and management of the condition for which the drug is indicated. Clinical experts are a critical part of the review team and are involved in all phases of the review process (e.g., providing guidance on the development of the review protocol, assisting in the critical appraisal of clinical evidence, interpreting the clinical relevance of the results, and providing guidance on the potential place in therapy). The following input was provided by 3 clinical specialists with expertise in the diagnosis and management of metastatic or advanced breast cancer.

Unmet Needs

More effective therapies were stated to be needed for patients, as patients with metastatic HER2-positive breast cancer will eventually experience progression of their disease. More convenience for therapies was also stated to be helpful for patients; subcutaneous formulations of trastuzumab and pertuzumab were stated to be in development, but currently not funded in Canada or commonly used. In addition, more effective treatments for patients with leptomeningeal disease are needed, as these patients are typically excluded from clinical trials, including the HER2CLIMB trial, and have extremely poor prognoses.

Development of brain metastases is common for patients with HER2-positive MBC, with approximately 50% of patients expected to develop intracranial disease, and is associated with poor prognoses. Patients with brain metastases were stated to be seen more frequently in clinical practice, as HER2-directed therapies are controlling extracranial disease and resulting in longer life for patients. It was stated that the brain may be protected from treatments with monoclonal antibodies (e.g., trastuzumab, pertuzumab and, potentially, T-DM1), as they are thought to be unable to cross the blood-brain barrier; however, as more data emerge, this belief is being challenged. Some HER2-directed TKIs, such as neratinib and lapatinib, have been shown to have CNS activity with intracranial responses, and have been shown to be particularly effective when combined with capecitabine chemotherapy. However, neither neratinib nor lapatinib combined with capecitabine was stated to have shown improved survival for patients in this setting. Therefore, for patients with brain metastases, unmet needs included improved survival, improved intracranial disease control, reduction of neurologic symptoms, need for supportive medications, improved functional status, and drugs that have both intracranial activity and tolerable toxicities.

Place in Therapy

Ideally, tucatinib was stated to be used for patients as a third-line treatment option for metastatic disease, after prior exposure to trastuzumab, pertuzumab, and T-DM1. Use of tucatinib was suggested to be used as per the HER2CLIMB trial, which enrolled patients who had prior exposure to trastuzumab, pertuzumab, and T-DM1; therefore, it was considered inappropriate by the clinical expert to use tucatinib for patients who had not received these treatments, unless patients are enrolled in a clinical trial. However, tucatinib may be considered for patients in an earlier line of therapy, depending on their disease-free interval and so long as the patient had received adjuvant trastuzumab, pertuzumab, and T-DM1. Tucatinib may also be considered for patients with other indications (e.g., intolerance or contraindications to other therapies); however, this is expected to represent a small subgroup of patients. Overall, introduction of tucatinib into current clinical practice may result in a shift in the treatment paradigm and push current third-line treatment options to later lines of therapy.

The clinical expert suggested administering tucatinib to patients who align with HER2CLIMB eligibility criteria. For patients with previous treatment with lapatinib, the clinical experts acknowledged that patients with prior treatment with lapatinib were eligible in the HER2CLIMB trial if it was administered 12 months before enrolment. However, they noted that lapatinib is usually administered to patients in combination with capecitabine, and prior treatment with capecitabine was an exclusion criteria of the HER2CLIMB trial. No specific subgroups of patients who might respond better to the tucatinib-combination treatment were identified.

The clinical experts acknowledged that, in clinical practice, some patients may not tolerate or may have a contraindication to certain treatments, such as T-DM1 or pertuzumab. For example, a patient may have peripheral neuropathy, which is a relative contraindication to T-DM1. The clinical experts agreed that for patients who cannot receive T-DM1 or pertuzumab due to contraindications or toxicity issues, treatment with tucatinib-combination therapy would be reasonable. However, there was agreement among all clinical experts that this population would represent a small group of patients.

In addition, the clinical experts considered that there may be rare scenarios in clinical practice where some patients may not have received standard first-line treatment with chemotherapy in combination with pertuzumab and trastuzumab if they experience recurrence within 6 months of completing treatment with trastuzumab in the adjuvant setting. Subsequent treatment for such patients would then be T-DM1. The clinical experts agreed that patients who did not receive prior treatment with pertuzumab may then be considered for tucatinib-combination treatment if subsequent disease progression occurs after treatment with T-DM1.

Patient Population

To identify patients eligible for treatment with tucatinib-combination therapy, patients with MBC are followed regularly by a medical oncologist who would determine the patient’s suitability with tucatinib. No issues related to diagnosis were identified, as these patients would already be diagnosed with MBC and have a history of prior treatments. Tucatinib-combination therapy was suggested not to be used for patients with poor ECOG PS (i.e., ECOG PS of 2 to 4); some patients with an ECOG PS of 2 might be considered, based on the judgment of the treating physician. Further, tucatinib may not be used for patients with unstable symptoms related to brain metastases, patients with leptomeningeal disease, or patients with contraindications to capecitabine or trastuzumab; for patients with unstable brain metastases, tucatinib could be considered as a treatment after patients receive and respond to local therapy. It was noted that patients in this setting typically survive less than 3 to 6 months, and that patients may also be receiving palliative care.

Assessing Response to Treatment

Patient’s response to treatments was stated to be based on tumour-response assessment on re-imaging of known sites of disease, usually with CT scans (with or without bone scans), clinical assessment (e.g., palpable disease, symptom control), and laboratory assessments (e.g., improving liver function, tumour marker levels). Frequency of assessments can vary, occurring every 3 to 6 months, depending on a patient’s disease status. For patients with brain metastases, a CT scan is considered reasonable, but a brain MRI can reveal other lesions that may be amenable to gamma knife stereotactic radiosurgery or surgery; therefore, brain MRIs are the preferred modality for brain imaging.

The outcomes used in the clinical trials were stated to have some alignment with the outcomes typically used in clinical practice. A meaningful response to treatment would typically result in improved survival, improved performance status, fewer disease symptoms, reduction in use of analgesic or other supportive medications, and less need for thoracentesis, paracenteses, and palliative radiotherapy. The magnitude of response to treatment was stated to be difficult to define and that physicians often need to use their own judgment. However, improvement in survival of greater than 2 to 3 months and improvements in symptoms and in patients’ ability to perform tasks of daily living are generally considered significant.

Discontinuing Treatment

The decision to discontinue a treatment should be made with the following considerations: imaging or clinical evidence shows a lack of benefit to patients or results in disease progression, a patient experiences intolerance to treatment or shows toxicity to treatment (e.g., diarrhea, fatigue, hand-foot syndrome), a patient’s performance status decreases, or the patient decides to stop treatment.

For some patients who progress on treatment with the tucatinib-combination therapy, a decision may be made to allow patients to continue receiving treatment. In such cases, patients may be considered eligible to continue on tucatinib-combination therapy if they have isolated progression in the brain that is amenable to local therapies with radiation or surgery.

Prescribing Conditions

Tucatinib-combination therapy was stated to be administered in an outpatient clinical setting.

Clinician Group Input

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

Inputs from 2 clinician groups were received for the Reimbursement Review of tucatinib (Tukysa) in combination with trastuzumab and capecitabine for the treatment of patients with locally advanced unresectable or metastatic HER2-positive breast cancer, including patients with brain metastases, who have received prior treatment with trastuzumab, pertuzumab, and T-DM1, separately or in combination.

The OHCC Breast Disease Site Group is a group of medical oncologists at the OHCC who treat breast cancer and are affiliated with the University of Ottawa in Ontario. They offer routine standard-of-care treatments and access to treatments through clinical trials. The group also serves as a large referral base for the Champlain Local Health Integration Network in Ontario. Input for this submission was collected through canvassing members electronically and in person for input and opinion. The recommendations from the clinicians were condensed and coalesced into summary statements, which reflect the breadth of opinions expressed. Clinician opinions were based on literature review and data from recent international congresses and publications.

The OH-CCO B-DAC provides evidence-based clinical and health system guidance on drug-related issues in support of OH-CCO’s mandate, including Ontario’s drug reimbursement programs and Systemic Treatment Program. Input on this review was discussed jointly via emails and at an OH-CCO B-DAC meeting.

Unmet Needs

Both clinician groups agreed that current first-line treatments include taxane chemotherapy combined with trastuzumab and pertuzumab followed by second-line T-DM1. Both groups acknowledged that there is no standard of care for patients in the third-line treatment setting following progression on T-DM1. Therefore, this particular population does not have access to effective treatments following progression on T-DM1, as treatments in this line of therapy do not have proven survival-prolonging benefits. The clinician groups highlighted the unmet need for HER2-targeted therapies in this treatment setting. They also stated that, other than radiation therapy, patients with brain metastases have limited effective treatment options. The OH-CCO B-DAC added that there is a considerable unmet need in patients with aggressive brain metastases who are not candidates for local therapies, as there are currently no other systemic drugs that offer significant CNS treatment options; as such, tucatinib may be preferred in this patient population. The OHCC’s Breast Disease Site Group noted that current standard third-line treatment options for patients may include endocrine or chemotherapy drugs used as single drugs sequentially.

The OHCC’s Breast Disease Site Group also noted that, internationally or for patients with private insurance, therapies in the third line can include: endocrine therapies with HER2-targeted options (e.g., fulvestrant and abemaciclib plus trastuzumab, aromatase inhibitors, or fulvestrant alone with trastuzumab), single-drug chemotherapy (i.e., capecitabine or vinorelbine) with trastuzumab, dual HER2-targeted combination therapy (i.e., trastuzumab with lapatinib), or chemotherapy and small molecule HER2-targeted approaches (i.e., lapatinib plus capecitabine or neratinib plus capecitabine). Treatments choices for patients are based on patient goals, performance status, symptoms, rapidity of disease progression, visceral organ involvement, and insurance coverage. OHCC’s Breast Disease Site Group added that clinical trials are recommended, when available for eligible patients.

Both clinician groups agreed that OS and quality of life are important treatment goals for patients. OHCC’s Breast Disease Site Group added that in this symptomatic patient population, improved response rates are also desirable to see more rapid symptom improvements. Both groups also agreed that dramatic PFS benefits are also valuable, deferring the need for earlier use of more toxic therapies. There was agreement that there is a need to prevent and delay brain metastases, which are common in this disease, and to effectively treat them, as other systemic therapies currently available to patients have not been effective in this line of treatment. Therefore, tucatinib, capecitabine, and trastuzumab fulfill a significant unmet need in treating HER2-positive breast cancer and have shown improved survival, including among those with brain metastases. The groups commented that, given the lack of other options in this setting, this regimen is critically needed to improve both mortality and morbidity in this patient population.

The OH-CCO B-DAC added that patients who received adjuvant T-DM1 and relapsed within 6 months are currently not eligible for pertuzumab funding in Ontario and are left without any anti-HER2-directed therapy. The group noted that although these patients were not eligible for tucatinib-combination therapy in the HER2CLIMB trial, funding should also be extended to this patient population.

With respect to which patients have the greatest unmet need for an intervention, both clinician groups agreed that patients with advanced HER2-positive breast cancer who have progressed on first- and second-line therapies are most in need, as they do not have any provincially funded options for third-line anti-HER2-directed therapy. OHCC’s Breast Disease Site Group added that patients receiving second-line treatment with contraindications to T-DM1 (e.g., persistent difficult residual peripheral neuropathy from preceding chemotherapy), recurrence or progression after preceding neoadjuvant or adjuvant trastuzumab (with or without pertuzumab or T-DM1), or those with active brain metastases (where the activity of T-DM1 alone is modest), would also be eligible for treatment with tucatinib-combination therapy.

Place in Therapy

Both clinician groups agreed that tucatinib in combination with capecitabine and trastuzumab would be added to existing treatment options and that, generally, tucatinib would fit in as a third line of treatment. OHCC’s Breast Disease Site Group noted that patients beyond the third-line treatment setting, such as those who were treated before tucatinib became available, would also be good candidates for this treatment, as long as their disease had not progressed while on capecitabine chemotherapy.

It was noted that this regimen under review would not, at present, alter the sequencing of therapies for the treatment of HER2-positive breast cancer because other available treatments could be offered beyond progression, as they have been in the past. The OH-CCO B-DAC commented that it is not expected that patients treated with tucatinib, capecitabine, and trastuzumab would be ineligible for subsequent treatment options. Additionally, the OHCC Breast Disease Site Group noted that, in light of expected attrition due to declining health and performance status, the number of subsequent lines of therapy would be fewer, commonly being between zero and 2 lines in practice.

Patient Population

Both clinician groups agreed that patients with advanced breast carcinoma (HER2-positive MBC as per the American Society of Clinical Oncology/College of American Pathologists HER2 criteria) on the basis of IHC analysis or ISH with good ECOG PS, and who had previously been treated with trastuzumab, pertuzumab, and T-DM1, would be best suited for treatment with the regimen under review. The OH-CCO B-DAC noted that testing using IHC, ISH, or FISH is considered standard in the treatment of breast cancer. The OHCC’s Breast Disease Site Group noted that patients who have been on active treatments, monitored closely by their medical oncologists and identified on the basis of objective disease progression, are best suited for treatment. Additionally, the OHCC’s Breast Disease Site Group noted that endocrine-based regimens should not disqualify from eligibility patients with more indolently behaving estrogen receptor–positive disease.

The OH-CCO B-DAC noted the patients least suitable for treatment with tucatinib include those patients previously treated with an anti-HER2 TKI, capecitabine (other than being considered on a time-limited basis), or patients with significant cardiac dysfunction. OHCC’s Breast Disease Site Group added that patients with an ECOG PS of 3 or 4, expected survival of less than 3 months, dysfunctional gastrointestinal tracts, or those ineligible for trastuzumab (based on past severe infusion reactions or cardiac dysfunction) or capecitabine, would be least suitable for treatment with tucatinib.

The OH-CCO B-DAC noted it is not possible to identify patients who would be most likely to exhibit a response to treatment with tucatinib. OHCC’s Breast Disease Site Group added that response rates in the HER2CLIMB trial were 40.6% per RECIST v1.1 criteria, and all eligible patients should have the opportunity to benefit based on the inclusion and exclusion criteria of the trial. The group added that no particular subgroups were found not to benefit from treatment.

Assessing Response to Treatment

The OH-CCO B-DAC noted the outcomes used in clinical practice are aligned with the outcomes that were used in the HER2CLIMB study. Treatment response should be assessed as per breast cancer guidelines and routine clinical and radiographic staging. The OHCC’s Breast Disease Site Group added that patients are evaluated before each cycle of treatment for efficacy and tolerability. The group noted that clinical examination, symptom improvement, biochemical parameters, and periodic radiographic restaging (usually every 3 to 4 cycles) are all used to determine whether a patient is responding to treatment in clinical practice.

Both clinician groups agreed that a clinically meaningful response to treatment would mean improving PFS, OS, and patients’ quality of life, which also includes improved control of CNS disease. The OHCC’s Breast Disease Site Group added that improvement in disease symptoms, which may vary depending on the extent of disease and location of metastatic involvement, or improvement of functional scores (e.g., ECOG PS, Edmonton Symptom Assessment Scale) is a clinically meaningful response. Additionally, radiographic responses by RECIST criteria are also important, signifying a disease-modifying effect that might imply longer PFS, delayed treatment with more toxic therapy, postponement of disability, and improved survival.

Discontinuing Treatment

Both clinician groups agreed that treatment should be discontinued due to disease progression (measured using radiographic RECIST criteria), intolerable or unmanageable toxicity, side effects, or patient or physician preference.

Prescribing Conditions

Both clinical groups agreed that tucatinib and capecitabine are oral therapies that can be given at the patient’s home (or in institutions), whereas IV trastuzumab is administered at specialized cancer clinics or infusion clinic environments. The OH-CCO B-DAC added that monitoring of oral anti-cancer medication would be helpful.

Additional Considerations

Both clinician groups noted that this regimen under review fills an urgent medical need and that the magnitude of benefit warrants approval and use based on the large phase II trial. The OHCC’s Breast Disease Site Group added that the activity seen in controlling or improving brain metastases (even if active) is unprecedented.

The OH-CCO B-DAC added that men were included in HER2CLIMB and should be eligible for tucatinib-combination therapy.

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.

Clinical Evidence

The clinical evidence included in the review of tucatinib 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. The second section includes indirect evidence from the sponsor and selected from the literature that met the selection criteria specified in the review. The third section includes sponsor-submitted long-term extension studies and additional relevant studies that were considered to address important gaps in the evidence included in the systematic review.

Systematic Review (Pivotal and Protocol Selected Studies)

Objectives

To perform a systematic review of the beneficial and harmful effects of tucatinib in combination with trastuzumab and capecitabine for the treatment of patients with locally advanced, unresectable, or metastatic HER2-positive breast cancer, including patients with brain metastases, who have received prior treatment with trastuzumab, pertuzumab, and T-DM1, separately or in combination.

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.²⁰

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 concept was Tukysa (tucatinib). Clinical trials registries were searched: the US National Institutes of Health’s clinicaltrials.gov, WHO’s International Clinical Trials Registry Platform (ICTRP) search portal, Health Canada’s Clinical Trials Database, and the European Union Clinical Trials Register.

No filters were applied to limit the retrieval by study type. 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 Review Expert Committee (pERC) on September 8, 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.²¹ Included in this search were the websites of regulatory agencies (US 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.

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.

Objectives and Methods for the Summary of Indirect Evidence

A focused literature search for NMAs dealing with Tukysa (tucatinib) and breast cancer was run in MEDLINE All (1946–) on April 20, 2021. The search was limited to documents published between January 1, 2010 and April 20, 2021.

Findings From the Literature

A total of 2 studies were identified from the literature for inclusion in the systematic review (Figure 1). The included studies are summarized in Table 6. A list of excluded studies is presented in Appendix 2.

Figure 1: Flow Diagram for Inclusion and Exclusion of Studies

Description of Studies

One multi-centre, multinational, double-blind, phase II randomized controlled trial met the criteria for the CADTH systematic review protocol. The HER2CLIMB trial (N = 612) evaluated the efficacy and safety of tucatinib in combination with trastuzumab and capecitabine compared with placebo in combination with trastuzumab and capecitabine, which will be referred to as the tucatinib-combination group and the placebo-combination group, respectively. Eligible patients included adults with histologically confirmed HER2-positive advanced breast cancer, confirmed using IHC, ISH, or FISH testing. Patients must have had prior treatment with pertuzumab, trastuzumab, and T-DM1, measurable disease using RECIST v1.1 criteria and an ECOG PS of 0 or 1. Patients with brain metastases were also eligible for enrolment.⁷ Patients were randomized in a 2:1 ratio using a hierarchical randomization scheme via an interactive response technology system to receive either tucatinib-combination or placebo-combination treatment. Patients were stratified according to the following: presence of brain metastases (yes versus no), ECOG PS (0 versus 1), and geographic region (US versus Canada versus the rest of the world). Presence of brain metastases was based on medical history and screening contrast brain MRI, as assessed by an investigator. Patients with a documented history of prior brain metastases and patients with brain lesions of equivocal significance on screening MRI were stratified into the “yes” category. The hierarchical randomization scheme included specifications for a biased-coin assignment to allow for randomization of approximately equal proportions of patients in each treatment group and stratification factor. The biased-coin assignment allowed for greater randomization of patients to specific groups when an imbalance at a given hierarchical level (i.e., overall treatment group balance, then treatment group balance within each stratification factor) exceeded a specified threshold.⁷

This international trial was conducted in 15 countries across 155 sites and included a total of 38 patients from Canada.^7,8 The study was double-blinded; patients, personnel in contact with study patients, data-collection personnel, and others associated with patient procedures or data handling were blinded to the treatment allocation. Access to treatment allocation information in the trial was permitted only in the event of a suspected unexpected serious adverse reaction where unblinding for regulatory purposes became necessary; all such events were reported in accordance with local regulatory requirements. The following data were unblinded and monitored regularly by an independent data monitoring committee: deaths, discontinuations, dose reductions, AEs, and SAEs). Efficacy assessments were performed once every 6 weeks for the first 6 months while on study, and then once every 9 weeks.⁷ An overview of the HER2CLIMB trial is illustrated in Figure 2.

Figure 2: HER2CLIMB Trial Schematic

BID = twice a day; CNS = central nervous system; PD = progressive disease; PO = orally; SC = subcutaneous.

^a Treatment continued until unacceptable toxicity, disease progression, withdrawal of consent, or study closure. Patients with CNS progression could undergo local therapy for CNS lesions and continue on study treatment with approval from the medical monitor for clinical benefit.

^b Contrast CT, PET/CT, and/or MRI, and brain contrast MRI scan at baseline, every 6 weeks for the first 24 weeks, and then every 9 weeks thereafter until PD, initiation of a new therapy, withdrawal of consent, or study closure. Patients without brain metastases at baseline did not require brain contrast MRIs while on treatment. A brain contrast MRI was required at the 30-day follow-up visit for all patients.

^c Assessment of overall survival and/or disease recurrence, as well as collection of information regarding any additional anti-cancer therapies administered after completion of study treatment.

^d If study treatment was discontinued for reasons other than disease progression or death, every reasonable effort was to be made to obtain contrast CT, PET/CT, and/or MRI, and contrast brain MRI (only in patients with known brain metastases) approximately every 9 weeks until disease progression, death, withdrawal of consent, or study closure.

Source: HER2CLIMB Clinical Study Report.⁷

Baseline Assessments

Patients were provided with the option of being pre-screened for HER2 status by central review before any study screening. The screening period lasted up to 28 days, during which all patients received an MRI of the brain and underwent high-quality spiral contrast CT, PET/CT, and/or contrast MRI scan imaging including, at a minimum, the chest, abdomen, pelvis, and appropriate imaging of other known sites of disease (e.g., skin lesion photography, bone imaging). Blood samples were taken for assessment of biomarkers, hematology, coagulation, chemistry, and liver function. Patients were also screened for cardiac function, hepatitis B and C, urinalysis, and pregnancy (except in men and women of non-child-bearing age).⁷

Post-Treatment Discontinuation Assessments

Approximately 30 days after patients discontinued study treatment, a final routine safety assessment was conducted that included a physical examination and laboratory assessments. All patients were also required to undergo a repeat contrast MRI of the brain. Patients who discontinued treatment for reasons other than disease progression, repeat imaging scans of all known areas of metastatic or locally advanced unresectable disease were requested until patients experienced either progression of disease or death. Bloodwork was conducted to assess for the presence of certain biomarkers of response (i.e., analysis of mutations in circulating tumour DNA), and performed locally at sites.

Populations

Inclusion and Exclusion Criteria

Eligibility criteria for the HER2CLIMB trial are reported in Table 6. Briefly, eligible patients included adult women with histologically confirmed HER2-positive advanced breast cancer. HER2-positive status must have been confirmed via IHC, ISH, or FISH methodology. Patients must have had prior treatment with pertuzumab, trastuzumab, and T-DM1, measurable disease via RECIST v1.1 criteria, and an ECOG PS of 0 or 1.⁷ Patients with brain metastases were also eligible for enrolment so long as there was no need for immediate local intervention. Patients who required local intervention could subsequently enrol after they had received local therapy. Presence of brain metastases was based on medical history and screening contrast brain MRI, as assessed by an investigator. Patients were not eligible for enrolment if they had received prior treatment with capecitabine, lapatinib, neratinib, afatinib, or any other HER2-targeted EGFR or TKI; however, patients who were treated with capecitabine or lapatinib 12 or more months before initiating any trial treatment, or who received treatment within 21 days and who discontinued for reasons other than disease progression or severe toxicity, could enrol in the HER2CLIMB trial.⁷

Baseline Characteristics

Baseline characteristics of the HER2CLIMB trial are reported in Table 7.

In general, baseline characteristics were well balanced across both treatment groups in both the ITT and ITT-PFS populations; baseline characteristics were also similar across both trial populations. In the ITT population, patients had a mean age of 54 years with most patients (> 80%) being less than 65 years of age. In the tucatinib- and placebo-combination groups, most patients were White, with slightly fewer White patients in the tucatinib-combination group (70.0%) than in the placebo-combination group (77.7%). Most were from the US (53.7% versus 55.0%, respectively) or the rest of the world (40.0% versus 39.1%). Relatively equal proportions of patients had an ECOG PS of 0 (49.8% in the tucatinib-combination group and 46.5% in the placebo-combination group) or 1 (50.2% versus 53.5%, respectively).⁷

The majority of patients had metastatic disease (≥ 99%) at study entry. At initial diagnosis, patients were classified as having stage II (30.0% versus 26.2% in the tucatinib- and placebo-combination groups, respectively), stage III (23.7% versus 21.8%), or stage IV (34.9% versus 38.1%) disease, and were positive for at least 1 hormone receptor (59.3% versus 62.9%) or negative for both (39.3% versus 37.1%). Ninety-eight percent of patients in both treatment groups reported non-CNS metastases at baseline, reporting mostly lung (48.8% versus 49.5%), bone (54.4% versus 55.5%), or liver (33.4% versus 38.6%) metastases. Brain metastases were reported in 48.5% of patients in the tucatinib-combination group and 48.0% of patients in the placebo-combination group.⁷ A mean of 4 lines of prior therapy were reported by all patients in both treatment groups, with a mean of 3 prior therapies specifically in the metastatic setting. As per eligibility criteria, all patients (100%) received prior treatment with trastuzumab and T-DM1; most patients (> 99%) reported prior therapy with pertuzumab. Patients who received prior treatment with trastuzumab did so mostly in the metastatic setting (56.8% in the tucatinib-combination group versus 63.9% in the placebo-combination group), followed by the neoadjuvant or adjuvant and metastatic setting (37.1% versus 29.2%, respectively). Patients receiving prior T-DM1 did so primarily in the metastatic setting only (99.0% versus 98.0% in the tucatinib- and placebo-combination groups, respectively). Patients receiving prior pertuzumab did so mostly in the metastatic setting only (86.3% versus 86.1% in the tucatinib- and placebo-combination groups, respectively). Few patients received prior treatment with trastuzumab, T-DM1, and pertuzumab in the neoadjuvant or adjuvant setting only.⁷

Interventions

Treatments were administered in 21-day cycles in the HER2CLIMB trial. The experimental group of the trial received the tucatinib-combination, consisting of tucatinib in combination with capecitabine and trastuzumab. The control group received the placebo-combination, consisting of placebo in combination with capecitabine and trastuzumab.⁷ The treatment dosages in the tucatinib-combination group were as follows:

• Tucatinib (300 mg) was administered orally twice daily.

• Capecitabine (1,000 mg/m²) was administered orally twice daily on days 1 to 14 of each 21-day cycle.

• Trastuzumab was administered with an initial loading dose of 8 mg/kg IV, after which it was administered at 6 mg/kg once every 21 days, except in specific circumstances where it was given weekly to compensate for modifications to the treatment schedule. Alternatively, trastuzumab could have been administered at a dose of 2 mg/kg IV every week (7 days), but only in circumstances when the trastuzumab infusion has been delayed, and weekly infusions are required to re-synchronize the cycle length to 21 days, after discussion with a medical monitor.

  • Subcutaneous use of trastuzumab was permitted; when subcutaneous trastuzumab was administered, a fixed dose of 600 mg was provided without a loading dose. Subcutaneous trastuzumab was administered once every 3 weeks, as there was no allowance for weekly dosages. Crossover from IV to subcutaneous trastuzumab was permitted within the trial.

  • Where national regulatory authorities approved the use of a trastuzumab biosimilar, either IV or subcutaneous, the biosimilar could be administered if considered appropriate by the investigator.

The treatment dosages in the placebo-combination group were the same as the tucatinib-combination group, except that patients received placebo tablets in place of tucatinib; patients also received the placebo orally twice daily. Treatment for patients continued until unacceptable toxicity, disease progression, withdrawal of consent, or study closure.⁷

Patients with isolated progression in the brain and stable systemic disease were permitted to be treated with local therapy (i.e., radiation therapy, surgery). These patients were permitted to continue the assigned study treatment for clinical benefit after a PFS event in the brain, with approval from a medical monitor. For patients to continue the study treatment after CNS progression, the following criteria were required to have been met:

• No worsening of cancer-related symptoms. Patients who were clinically deteriorating and unlikely to continue to receive further benefit from the study treatment should be discontinued.

• Patient should be tolerating the study treatment.

• Continuation of study treatment reviewed and approved by the medical monitor.

• Patient had no evidence of unequivocal systemic progression.

• Patient had not had a previous isolated CNS progression while on study treatment.

Study treatments were permitted to be held for up to 6 weeks to allow for local CNS therapy. Oral therapies, including tucatinib, placebo, and capecitabine, were to be held 1 week before planned CNS-directed therapy. Capecitabine, which was stated to be a known radiation sensitizer, was required to be held before CNS-directed therapy. Trastuzumab was permitted to continue while patients received radiotherapy, as it has been shown not to increase the effects of radiation. The sponsor noted that this approach of continuing treatment after evidence of progression in the brain approximates common off-practice study in clinical practices; therefore, the duration of treatment after patients experience progression in the brain was assessed only as an exploratory objective.⁷

Patients who discontinued the study treatments received further care, as determined by their physician. For patients who did not have any clear evidence of clinical progression, development of CNS symptoms, or radiographic changes thought to pose immediate risk, efforts were made to maintain the study treatments until unequivocal evidence of radiologic or clinical progression as per RECIST v1.1 criteria. No crossover was permitted during the double-blind phase of the trial.⁷ In the absence of progressive disease, efforts were made to avoid radiation therapy or surgery to target lesions in the brain unless clinically necessary, in the opinion of the investigator. This was done because target lesions could not be adequately assessed for subsequent response to systemic therapy once treated with local CNS therapy; this might have interfered with assessment of PFS, which was the main end point of the HER2CLIMB trial.⁷

Dose Modifications or Reductions

Dose modifications were required for certain clinical AEs, significant changes in hepatotoxicities, measured cardiac ejection fraction, and QT interval. Dose modifications for other toxicities were also permitted at the investigator’s discretion.⁷ Dose modifications of tucatinib, placebo, and capecitabine were permitted. Tucatinib and placebo dose reductions were made in steps of 50 mg, 100 mg, or 150 mg per dose. Up to 3 dose reductions of tucatinib or placebo were permitted. However, dose reductions of tucatinib or placebo below 150 mg per dose were not permitted; patients requiring dose reductions lower than 150 mg per dose or who required a fourth dose reduction should discontinue from the study. Re-escalations of tucatinib or placebo were not permitted. Dose reductions of capecitabine occurred based on toxicities per Common Terminology Criteria for Adverse Events criteria, which specify modifications by grade of AEs; doses of capecitabine should not be re-escalated once a dose reduction was made. Dose reductions of trastuzumab were not permitted within the trial. Doses of any trial treatment were permitted to be held for up to 6 weeks for toxicity; however, doses held for periods longer than 6 weeks required approval from a medical monitor.⁷

Treatment Discontinuations

Patients who experienced toxicity from capecitabine or trastuzumab were permitted to discontinue 1 of these drugs and continue on tucatinib or placebo in combination with either capecitabine or trastuzumab. Patients were not permitted to continue the study treatments if they required discontinuation of both capecitabine and trastuzumab, or discontinuation of tucatinib or placebo; these patients were followed for efficacy assessments per the protocol schedule.⁷

Concomitant Medications

Concomitant medications, which included strong CYP3A4 or CYP2C8 inducers of inhibitors, warfarin therapy, or therapy with other coumarin derivatives that are known and accepted to prolong QTc, were not permitted while in the study. However, some treatments that are possibly associated with QTc prolongation were permitted to be administered with caution. Corticosteroids at a daily dose of greater than 2 mg dexamethasone or equivalent for control of symptoms of brain metastases were not permitted at the time of study entry. Any planned surgery that was not directly related to cancer required consultation with the sponsor’s medical monitor, and patients were required to suspend the study treatments 3 to 7 days before surgery and resume treatment 3 to 14 days post surgery. Standard supportive care measures (e.g., anti-emetics, antidiarrheal medications, permitted concomitant medications, and hematopoietic support) were permitted within the trial but were not required.⁷

Outcomes

A list of efficacy end points identified in the CADTH review protocol that were assessed in the clinical trials included in this review is provided in Table 8. These end points are further summarized subsequently. A detailed discussion and critical appraisal of the outcome measures is provided in Appendix 4.

Efficacy Assessments

Efficacy assessments included measurement of all known sites of metastatic or locally advanced unresectable disease (including, at a minimum, the chest, abdomen, and pelvis) by high-quality spiral contrast CT, PET/CT (if high-quality CT scan included), and/or MRI scan, as appropriate, as well as appropriate imaging of any other known sites of disease (e.g., skin lesion photography, bone imaging). Scans were taken at baseline, every 6 weeks for the first 24 weeks, and every 9 weeks thereafter, irrespective of dose holdings or interruptions.⁷ Efficacy assessments for each patient continued until PFS was confirmed and documented. Follow-up for survival will continue until study closure. A contrast MRI of the brain was required on this same schedule only for those patients with a prior history of brain metastases or if brain metastases were found at screening. Contrast brain MRI might also have been performed for patients without known brain metastases if there was clinical suspicion of new brain lesions. Additional imaging, such as nuclear medicine bone scan or other unscheduled scans, was performed at the discretion of the investigator. Treatment decisions were made based upon local assessment of radiologic scans. Patients in both arms of the study would continue to be followed for OS after completion of study treatment as well as after the occurrence of disease progression.⁷

Safety Assessments

Safety data were monitored and performed by the sponsor throughout the study on a blinded basis. An independent data monitoring committee regularly reviewed all safety and efficacy data. Patients were assessed throughout the study for safety. Safety assessments, including a physical exam and collection of AEs and laboratory abnormalities, were performed at least once every 3 weeks throughout study treatment and 30 days after the last dose of the study drugs. Laboratory assessments were performed locally.⁷ During cycle 1, an in-person safety assessment was performed on days 1 and 12. During cycle 2, an in-person safety assessment took place on day 1 and liver function tests (AST, ALT, and total bilirubin) were performed on day 12 of cycle 2. An in-person safety assessment was then performed on day 1 of each cycle throughout the remainder of the study or as clinically indicated. Cardiac ejection fraction was assessed by multiple-gated acquisition scan or echocardiogram performed at screening, once every 12 weeks thereafter until study discontinuation (irrespective of dose delays or interruption), and 30 days after the last dose of the study drugs (unless the last scan was done less than 12 weeks before the 30-day follow-up visit).⁷

Exploratory End Points

The objectives of the exploratory end points included assessment of the tucatinib combination compared with the placebo combination using Response Assessment in Neuro-Oncology Brain Metastases (RANO-BM) in the subgroup of patients with brain metastases, as determined by BICR, assessment of the effect of the tucatinib combination versus the placebo combination on progression in the brain in the subgroup of patients with brain metastases.⁷

Health-Related Quality of Life

HRQoL was assessed using the EQ-5D-5L questionnaire. The questionnaire was administered at cycle 1 on day 1 before the start of study treatment, after which patients were administered the questionnaire every 6 weeks for 24 weeks and then every 9 weeks until disease progression, death, toxicity, withdrawal of consent, or study closure. A post-treatment assessment also occurred approximately 30 days after the end of treatment.⁷

Statistical Analysis

Analysis Populations

A description of the main efficacy and safety analysis sets is provided in Table 9.

Sample Size

The sample of patients in the HER2CLIMB trial was pre-specified to ensure enough power for the assessment of PFS, PFS_BM, and OS. Details are provided in Table 10; this table and Figure 3 also describe the timing for the primary and interim analyses.

The accrual period for patients was expected to occur within 42 months, with 12 months of follow-up for PFS continuing after the last patient was randomized. Assuming that 10% of patients are lost to follow-up annually, it was expected that approximately 480 patients would have to be randomized to have 214 PFS events in the subgroup of patients with brain metastases at baseline.⁷ In addition, based on an accrual period of 48 months for OS and a 5% yearly dropout rate, 361 OS events were expected to occur after approximately 59 months after the first patient was randomized.³

Figure 3: Timing of Primary and Interim Analyses

ITT = intention to treat; OS = overall survival; PFS = progression-free survival; PFS_BM = progression-free survival among the subgroup of patients with brain metastases.

^a If these 2 conditions are not met, then this analysis will be skipped.

^b If PFS_BM is positive at the primary analysis, OS interim analysis 2 will be conducted at 75% OS events (approximately 271 events).

^c Only if conditions for second analysis timing are not met.

Source: HER2CLIMB statistical analysis plan.³

It should be noted that, as the trial included PFS among the overall trial population and within the subgroup of patients with brain metastases at baseline and OS as end points, the analysis of PFS was conducted hierarchically to control for the family-wise type I error rate of 0.05; therefore, the analysis of PFS among the overall population was tested first in the ITT-PFS population, and, if it was found to be statistically significant, testing was conducted for key secondary end points using the group sequential Holm variable procedure.^8,23

The alpha was split between the end points of OS and PFS_BM (with 0.03 and 0.02 allocated, respectively) and tested at both the interim and final analyses, if not rejected at the interim analysis. The statistical significance of the boundaries at the interim analysis was determined using a Lan-DeMets O’Brien-Fleming approximation spending function. The boundaries for PFS_BM and OS at the interim analysis are outlined in Table 11.³

If both OS and PFS_BM were reported to be statistically significant at the primary analysis, no formal testing for the 2 end points was specified to occur after that point. The protocol of the HER2CLIMB trial pre-specified the alphas to be used for the analyses of each efficacy end point, depending on whether 1 or both OS and PFS_BM were found to be statistically significant at the primary analysis. However, as both end points were found to be statistically significant at the primary analyses, the details of other possible scenarios are not reported here.³ The initial boundaries for the first interim analysis follow Table 11.³

If both the OS and PFS_BM end points were found to be statistically significant, then the secondary end point of ORR was to be formally tested between the treatment groups using a 2-sided alpha of 0.05.³

Data collected in the study were presented using summary tables, patient data listings, and figures. Continuous outcomes were summarized using descriptive statistics (e.g., mean, median, standard deviation, and minimum and maximum). Categorical outcomes were summarized by frequencies and percentages. Two-sided 95% CIs were presented where needed.

Disease evaluations were performed by both an independent central review and investigator using RECIST v1.1 criteria. Exploratory efficacy end points, which analyze patients with and without brain metastases, disease evaluations were performed using the bi-compartmental tumour assessment method, with non-CNS disease evaluated using RECIST v1.1 criteria and the CNS disease evaluated using the RANO-BM criteria.³

The study continued to collect relevant clinical and survival follow-up data until approximately 350 deaths were recorded, which was described as the end of the study.³

Efficacy Analyses

Primary End Point

Progression-free survival: The primary end point of PFS was assessed through BICR. Treatment groups were compared using a 2-sided log-rank test. A procedure based on re-randomization was implemented that took into account stratification factors used for randomization (known history of treated or untreated brain metastases, ECOG PS, and region of the world).³ The primary analysis for PFS included all randomized patients, and Kaplan–Meier methodology was used to estimate PFS time curves, including the median. A Cox proportional hazards model was used to estimate the HR and corresponding 95% CI, taking into account stratification factors. The P value for PFS was calculated using a re-randomization-based procedure²⁴ to reflect the dynamic, hierarchical allocation scheme used for randomization in this trial.³ Patients were censored under the following circumstances:

• Patients without post-baseline tumour assessments were censored on the date of randomization.

• Patients without documented disease progression or death were censored on the date of their last assessment for progression.³

• Patients who began new anti-cancer treatment before disease progression or death were censored on the date of their last assessment before beginning new anti-cancer treatment.

• Patients who died or progressed after consecutively missing 2 or more tumour assessments were censored on the date of their last assessment for progression.

Key Secondary End Points

The analyses of PFS in the subgroup of patients with CNS and OS were conducted in the same manner as the primary end point of PFS. Analyses of all other end points relied primarily on descriptive summary statistics and CIs.

Progression-free survival in patients with brain metastases at baseline: PFS_BM was tested if PFS among the overall population reached statistical significance at the 2-sided 0.05 level. PFS for this subgroup of patients was assessed using the methodology outlined for PFS in the overall population (see Primary End Point discussed previously).

Overall survival: Follow-up for OS was expected to continue until a sufficient number of events were recorded to have 90% power to test the effect of treatment on OS. It was assumed that the effect of OS would be smaller than the effect expected for PFS; therefore, an HR of 0.70 was expected for OS and required approximately 350 OS events. The primary analysis for OS was expected to occur approximately 1 to 2 years after the primary analysis of PFS.³

OS was analyzed in the same manner as the primary end point PFS. A log-rank test was used to compare OS between treatment groups. Patients who did not experience an OS event were censored on their last assessment for vital status. OS was only considered to be formally tested if PFS in the overall population and among the subgroup of patients with brain metastases at baseline were statistically significant at the 0.05 alpha level. OS was assumed to be immature at the time of the primary analysis for PFS; however, testing of OS was still planned to occur.³ OS was considered to be statistically significant at the primary analysis for PFS only if the following conditions were met:

• Both analyses for PFS in the overall population and among the subgroup of patients with brain metastases at baseline were statistically significant at the 0.05 alpha level.

• The P value for OS was 0.001 or smaller.

Other Secondary End Points

The following secondary end points were not adjusted for multiplicity. These end points were analyzed using conventional log-rank statistical methods.

Progression-free survival (investigator-assessed): A stratified Cox proportional hazards regression model controlling for study stratification factors was used to estimate the HR and 95% CI for PFS_INV. The comparison of 2 treatment groups was performed using a stratified log-rank test controlling for the study stratification factors with a nominal P value provided. Kaplan–Meier estimates of the median and corresponding 95% CIs were also computed for each treatment group.³

Censoring rules were the same as those specified for PFS as the primary end point.³

Objective response rate: Summaries for ORR included patients with measurable disease at baseline. The proportion of patients achieving an objective response (complete response or partial response) were calculated for each treatment group. The comparison of ORR between groups was performed using a 2-sided Cochran-Mantel-Haenszel test controlling for stratification factors. If both key secondary end points were statistically significant, then testing of ORR was conducted formally and a P value from the stratified Cochran-Mantel-Haenszel test was reported.³

Duration of response: Patients who achieved a confirmed response were included in the analysis for DOR. Estimates for medians and corresponding 95% CIs were computed using Kaplan–Meier methodology. A stratified log-rank test was used to calculate the nominal P value. Censoring for DOR occurred in the same manner as the primary end point, PFS.³

Subgroup Analysis

Subgroup analysis for the primary and key secondary end points included the following: history of brain metastases or brain metastases at baseline, geographic region, ECOG PS, age, race, and hormone receptor status. Analyses were conducted using stratified log-rank testing and stratified Cox proportional hazards regression models. For subgroups that are stratification factors, stratification occurred using the other stratification factors. Subgroup analyses for subgroups consisting of less than 10% of the number of patients in the total population were not to be performed.³

Safety

Safety data were assessed through summaries of AEs, changes in laboratory test results, changes in vital signs, changes in ECOG PS, changes in cardiac ejection fraction results, SAEs, treatment-emergent adverse events (TEAEs), AEs of special interest (i.e., any drug-induced liver injury, asymptomatic left ventricular systolic dysfunction, and/or cerebral edema), discontinuations, and deaths. AEs were classified by system organ class and preferred term using the Medical Dictionary for Regulatory Activities (MedDRA), with severity of AEs classified using the Common Terminology Criteria for Adverse Events v.4.03.⁷

Health-Related Quality of Life

The tucatinib-combination and placebo-combination groups index values using the EQ-5D-5L were compared using a t-test if the scores were distributed normally and using a Mann–Whitney test if the data were non-normal. HRQoL was analyzed using the ITT population.⁷

Results

Patient Disposition

Primary Analysis

The following data are based on the data cut-off date of September 4, 2019. Between February 23, 2016 and May 3, 2019, a total of 612 patients were randomized into the HER2CLIMB trial, including 410 patients in the tucatinib-combination group and 202 patients in the placebo-combination group (Table 12).⁷ Screening failures were most commonly due to patients not having HER2-positive breast cancer histologically confirmed by a central laboratory (inclusion criteria 1).

Of the patients randomized, a lower proportion in the tucatinib-combination group discontinued treatment (69.8%) compared with the placebo-combination group (84.2%). The main reason for treatment discontinuation in both treatment groups was disease progression, although fewer patients in the tucatinib-combination group experienced disease progression compared with the placebo-combination group (49.2% versus 66.3%, respectively). A greater proportion of patients in the tucatinib-combination group remained in treatment compared with the placebo-combination group (28.8% versus 13.4%, respectively), but fewer patients in the tucatinib-combination group than in the placebo-combination group remained in long-term follow-up (36.1% versus 42.1%). Discontinuations from the trial were lower in the tucatinib-combination group (35.1%) than in the placebo-combination group (44.6%); death was the main reason for trial discontinuation, accounting for 30.7% of discontinuations in the tucatinib-combination group versus 41.6% in the placebo-combination group.⁷

Of the randomized patients, 6 were not treated in the tucatinib-combination group and 5 were not treated in the placebo-combination group.⁷ The reasons for not being treated included the following:

• Two patients in the tucatinib-combination group and 4 patients in the placebo-combination group did not meet all eligibility criteria at the time of randomization.

• One patient in each treatment group withdrew consent after randomization.

• One patient in each treatment group met eligibility criteria at the time of randomization, but were no longer considered eligible due to changes in laboratory values at cycle 1 day 1.

• One patient in the tucatinib-combination group consented to enrol and was subsequently randomized, but then decided to pursue treatment at a different institution.

Patients who did not receive treatment were not included in the safety population, which included 404 patients in the tucatinib-combination group and 197 patients in the placebo-combination group; a greater proportion of patients in the tucatinib-combination group were on treatment than in the placebo-combination group (38.8% versus 13.4%, respectively).⁷

A summary of patient disposition for the ITT-PFS population, which was the analysis set used for the primary end point of the HER2CLIMB trial, is also provided in Table 13. In general, trends for both treatment groups reported for the ITT population were consistent with the patient disposition for the ITT-PFS population.⁷

Post-Hoc Analysis

Data from the post-hoc analysis are based on a data cut-off of February 8, 2021. The disposition of patients at the post-hoc analysis showed an increase of patients in both the tucatinib- and placebo-combination groups who were off tucatinib or placebo (90.0% and 97.0%, respectively); mainly, more patients in the tucatinib-combination group were off treatment at this longer follow-up than at the primary analysis. At this time, patients in the placebo-combination group were eligible for crossover into the tucatinib-combination group; a total of 26 patients (12.9%) crossed over to the tucatinib-combination group (Table 14).⁹

A total of 119 patients (29.0%) in the tucatinib-combination group and 50 patients (24.8%) in the placebo-combination group remained in long-term follow-up. Within the tucatinib- and placebo-combination groups, 256 patients (62.4%) and 142 patients (70.3%), respectively, discontinued from the study. The primary reasons for discontinuation were death (55.9% and 67.3% in the tucatinib- and placebo-combination groups, respectively), followed by withdrawal of consent (5.4% versus 2.5%), lost to follow-up (1.2% versus 0%) and physician decision (0% versus 0.5%).⁹ Of the patients who were off treatment, the majority continued to receive subsequent therapies (76.8% in the tucatinib-combination group versus 81.2% in the placebo-combination group).⁹

Exposure to Study Treatments

Summaries of dose exposure and dose intensity are provided in Table 15 and Table 16 and are based on data from the primary analysis of the HER2CLIMB trial (data cut-off: September 4, 2019). In both the safety and ITT-PFS populations, patients in the tucatinib-combination group received treatment (with tucatinib, capecitabine, and trastuzumab) for longer mean periods of time compared with patients in the placebo-combination group (who received placebo, capecitabine, and trastuzumab). The cumulative dose of trastuzumab administered to patients was also greater in the tucatinib-combination group than in the placebo-combination group.⁷ Specifically regarding tucatinib versus placebo, the mean duration of treatment with tucatinib was 7.6 months (standard deviation [SD] = 6.3) compared with 5.6 months (SD = 4.3) for placebo.⁷ Additional data from a post-hoc analysis (data cut-off: February 8, 2021) showed consistent results with a greater mean length of treatment with tucatinib at 10.2 months (SD = 9.6) versus 6.1 months (SD = 5.0) with placebo.⁹

The absolute and relative dose intensities of tucatinib or placebo plus capecitabine were both greater in the placebo-combination group than in the tucatinib-combination group, although, the mean cumulative dose of capecitabine was greater in the tucatinib-combination group than in the placebo-combination group.⁷ The lower dose intensity of treatment for tucatinib or placebo plus capecitabine in the tucatinib-combination group, despite having a longer duration of treatment, may be due to a greater proportion of patients in the tucatinib-combination group reporting they had a dose held or had their dose reduced due to AEs (Table 17).

Dose Modifications

Regarding tucatinib, of 84 patients (20.8%) requiring dose reductions, most were reduced to 250 mg (14.9%). Most patients in both the tucatinib- and placebo-combination groups resumed treatment at the same dose after a dose of tucatinib or placebo was held due to AEs (57.6% versus 58.3%, respectively), and most patients had only 1 dose hold (30.9% versus 28.4%, respectively) versus 2 (12.6% versus 8.1%) or more (9.9% versus 4.1%) dose holds. More patients in the tucatinib-combination group were reduced to a lower dose of tucatinib or placebo (26.2%) after a dose hold compared with patients in the placebo-combination group (17.4%). Patient doses of placebo or tucatinib were held for a mean number of 10.4 days (SD = 8.0) in the tucatinib-combination group and 13.6 days (SD = 10.6) in the placebo-combination group.⁷

Regarding capecitabine, most patients required only 1 dose hold (31.2% and 29.9% in the tucatinib- and placebo-combination groups, respectively), versus 2 (19.6% and 15.2%) or more (17.6% and 12.2%) dose holds. Most patients were able to resume capecitabine at the same dose (40.4% versus 35.2% in the tucatinib- and placebo-combination groups, respectively) or at a lower dose (36.3% versus 33.3%). The mean number of days for which capecitabine doses were held was 8.7 days (SD = 4.8) in the tucatinib-combination group and 8.3 days (SD = 5.4) in the placebo-combination group.⁷

More patients in the tucatinib-combination group (25.5%) also had a dose of trastuzumab held due to AEs than did patients in the placebo-combination group (19.3%). The proportion of patients who were permanently discontinued from treatment was similar across both treatment groups.⁷ The majority of patients in the tucatinib- and placebo-combination groups required 1 dose hold due to AEs (20.3% and 15.2%, respectively), and most were able to return to the same dose (83.6% and 73.5%). More patients in the placebo-combination group (20.4%) did not have any further dosing of trastuzumab after a dose hold compared with the tucatinib-combination group (11.7%). The mean duration of dose holds for trastuzumab was reported to be 11.6 days (SD = 7.5) in the tucatinib-combination group and 13.0 days (SD = 7.6) in the placebo-combination group.⁷

Treatment discontinuations of tucatinib or placebo, capecitabine, and trastuzumab were similarly reported in both treatment groups.⁷

Concomitant Medications

Concomitant medications were received by most patients in both the tucatinib- and placebo-combination groups (99.3% versus 97.5%, respectively). The most commonly administered types of concomitant medications were antipropulsives (65.6% in the tucatinib-combination group and 36.5% in the placebo-combination group), proton pump inhibitors (38.9% and 37.1%), anilides (36.1% and 36.5%), serotonin (5-hydroxytryptamine [5HT3]) antagonists (36.1% and 34.0%), other anti-emetics (32.7% and 28.4%), benzodiazepine derivatives (31.9% and 32.0%), vitamin D and analogues (26.2% and 27.4%), natural opium alkaloids (23.8% and 33.5%), propionic acid derivatives (23.3% and 24.4%), and other analgesics and antipyretics (20.3% and 20.8%). In general, except for antipropulsives and natural opium alkaloids, the proportions of patients receiving each type of concomitant medication were similar.⁷

Types of concomitant procedures, including surgery (13.4% and 10.7%), radiation (6.2% and 6.6%), and other (6.2% and 5.1%) were similar across the tucatinib- and placebo-combination groups, respectively.⁷ Concomitant systemic corticosteroids were used for a median number of 16.0 days (range, 1 to 849) in the tucatinib-combination group and 23.0 days (range, 1 to 432) in the placebo-combination group.⁷

Subsequent Anti-Cancer Therapies

A summary of subsequent anti-cancer therapies is provided in Table 18. A total of 292 patients in the tucatinib-combination group and 175 patients in the placebo-combination group discontinued or never received tucatinib or placebo. In the ITT population, subsequent anti-cancer therapies were administered to 202 patients (69.2%) in the tucatinib-combination group and 139 patients (79.4%) in the placebo-combination group. In general, a greater proportion of patients in the placebo-combination group received subsequent anti-cancer therapies than in the tucatinib-combination group, including antibody (57.1% versus 50.0%, respectively) and TKI (24.0% versus 16.8%) anti-HER2 regimens, and trastuzumab (12.2% versus 5.4%). Of note, trastuzumab plus chemotherapy was more commonly reported among patients in the tucatinib-combination group (20.8%) than in the placebo-combination group (15.8%). Other types of subsequent therapies were reported in similar proportions across both treatment groups.⁷

Efficacy

Only those efficacy outcomes and analyses of subgroups identified in the review protocol are reported subsequently. See Appendix 3 for detailed efficacy data. A summary of efficacy end points is provided in Table 19 and is based on a data cut-off date of September 4, 2019. The HER2CLIMB trial met its primary and key secondary end points showing statistically significant improvement in patients treated with the tucatinib combination versus the placebo combination. As both OS and PFS_BM were statistically significant at the time of this primary analysis, the results were considered to be from a final analysis; therefore, no formal testing for these end points was conducted at later time points (see Statistical Analysis section). ORR was considered to be another secondary end point that also supported the results of the primary and key secondary analyses showing improved efficacy with the tucatinib-combination versus the placebo-combination treatment. A detailed summary of efficacy outcomes is provided subsequently.

A post-hoc analysis, which was conducted by the sponsor approximately 2 years after the last patient of the HER2CLIMB trial was randomized, provided an additional 15.6 months of follow-up. The post-hoc analysis provided updated data for OS and PFS, assessed among all randomized patients, which continued to support trastuzumab-combination over placebo-combination therapy (Table 20).

The assessments conducted at the post-hoc analysis were not formally tested; therefore, they should be considered descriptive.

Progression-Free Survival

At the time of the primary analysis (data cut-off: September 4, 2019), the median duration of follow-up for PFS was 10.4 months for the total ITT-PFS population, per BICR. There were 178 events (55.65%) in the tucatinib-combination group and 97 events (60.6%) in the placebo-combination group. The median PFS was 7.8 months (95% CI, 7.5 to 9.6) in the tucatinib-combination group compared with 5.6 in the placebo-combination group, resulting in a 46% reduction in the risk of disease progression or death (stratified HR = 0.54; 95% CI, 0.42 to 0.71; stratified log-rank P value < 0.00001) (Figure 4).⁷ The results of the sensitivity analyses of PFS_INV were consistent with the primary analysis (stratified HR = 0.56; 95% CI, 0.45 to 0.70; stratified log-rank P value < 0.00001).

At the updated analysis (data cut-off: February 8, 2021), the median follow-up for PFS was 28.1 months in the tucatinib-combination group and 20.1 months in the placebo-combination group. The post-hoc analysis of PFS_INV continued to show improved PFS in the tucatinib-combination group, with a longer median PFS of 7.6 months (95% CI, 6.9 to 8.3) versus the placebo-combination group, which showed a median PFS of 4.9 months (95% CI, 4.1 to 5.6) (HR = 0.57; 95% CI, 0.47 to 0.70; nominal P value of < 0.00001) (Figure 5).⁹

Figure 4: PFS per BICR by Treatment Group (ITT-PFS)

BICR = blinded independent central review; cap = capecitabine; CI = confidence interval; PFS = progression-free survival; ECOG PS = Eastern Cooperative Oncology Group performance status; HR = hazard ratio; ITT = intention to treat; pbo = placebo; tra = trastuzumab; tuc = tucatinib.

HR is computed from the Cox proportional hazards model using stratification factors (presence or history of brain metastases [yes/no], ECOG PS [0/1], and region of world [North America versus the rest of the world]) at randomization.

Two-sided P value based on stratified log-rank test and re-randomization procedure (Rosenberger and Lachin, 2002).

Source: HER2CLIMB Clinical Study Report.⁷

Figure 5: PFS per Investigator by Treatment Group (ITT Population)

Cape = capecitabine; CI = confidence interval; HR = hazard ratio; ITT = intention to treat; PFS = progression-free survival; Pbo = placebo; Tras = trastuzumab; TUC = tucatinib.

Subgroup Analysis for PFS

A summary of subgroup analyses of the primary analysis is depicted in Figure 6. Subgroup analyses supported the primary analyses for PFS and favoured treatment with the tucatinib combination in all subgroups except for patients aged 65 years and older.⁷ Subgroup analyses were not adjusted for multiplicity and should be interpreted with caution.

Figure 6: Subgroup Analyses for PFS per BICR (Primary End Point, ITT-PFS Population) (Data Cut-Off: September 4, 2019)

BICR = blinded independent central review; CI = confidence interval; ECOG = Eastern Cooperative Oncology Group; EDC = electronic data capture; ER = estrogen receptor; HR = hazard ratio; N = no; PFS = progression-free survival; PR = progesterone receptor; Y = yes.

HR was calculated from Cox regression model considering stratified factors from randomization.

“Race Non-White” included patients with race other than White.

“Hormone receptor status: ER and PR negative” included patients without positive estrogen or positive progesterone receptors.

“Baseline brain metastasis: Y” included patients with a history of brain metastases or presence of brain metastases or brain lesions of equivocal significance on screening MRI per EDC data.

Source: HER2CLIMB Clinical Study Report.⁷

Censoring occurred for 142 patients (44.4%) in the tucatinib-combination group and 63 patients (39.4%) in the placebo-combination group (Table 21). Fewer patients in the tucatinib-combination group were censored due to progression events than in the placebo-combination group, but more patients were censored due to starting a new anti-cancer therapy in the placebo-combination group versus the tucatinib-combination group.

Progression-Free Survival Among Patients With Brain Metastases (per BICR)

At the time of the data cut-off, there were 106 events (53.5%) in the tucatinib-combination group and 51 events (54.8%) in the placebo-combination group, accounting for 71% of planned PFS events in the ITT-PFS_BM population. As previously stated in the Statistical Analysis section of this report, a 2-sided alpha level of 0.0080 (under the total alpha of 0.03) was used for the analysis of PFS_BM.⁷

The median PFS_BM was 7.6 months (95% CI, 6.2 to 9.5) in the tucatinib-combination group compared with 5.4 months (95% CI, 4.1 to 5.7) in the placebo-combination group, resulting in a 52% reduction in the risk of disease progression or death (stratified HR = 0.48; 95% CI, 0.34 to 0.69, stratified log-rank P value < 0.00001) (Figure 7).⁷

Figure 7: PFS_BM per BICR in Patients with Brain Metastases by Treatment Arm (ITT-PFS_BM) (Data Cut-Off: September 4, 2019)

BICR = blinded independent central review; Cap = capecitabine; CI = confidence interval; ECOG PS = Eastern Cooperative Oncology Group performance status; HR = hazard ratio; ITT = intention to treat; PFS = progression-free survival; PFS_BM = progression-free survival among the subgroup of patients with brain metastases; Pbo = placebo; Tra = trastuzumab; Tuc = tucatinib.

Note: Statistically significant after adjustment for multiplicity. The threshold for statistical significance was 0.0080.

The brain metastases population was defined as a subset of patients with a history of brain metastases or presence of brain metastases or brain lesions of equivocal significance on screening MRI.

HR was computed from the Cox proportional hazards model using stratification factors (ECOG PS [0/1], and region of world [North America versus the rest of the world]) at randomization.

Two-sided P value based on stratified log-rank test and re-randomization procedure (Rosenberger and Lachin, 2002).

Source: HER2CLIMB Clinical Study Report.⁷

Subgroup Analysis for PFS_BM

Subgroup analyses for PFS_BM are illustrated in Figure 8. The subgroup analyses were consistent with the primary analyses for PFS_BM and favoured treatment with the tucatinib-combination in all subgroups except for patients aged 65 years and older, patients with an ECOG PS of 0, and patients who were from the rest of the world.⁷ The subgroup analyses were not adjusted for multiplicity.

Figure 8: Subgroup Analyses for PFS_BrainMets (Data Cut-Off: September 4, 2019)

CI = confidence interval; ECOG = Eastern Cooperative Oncology Group; ER = estrogen receptor; HR = hazard ratio; ITT = intention to treat; PFS_BrainMets = progression-free survival among the subgroup of patients with brain metastases; PR = progesterone receptor.

HR is calculated from Cox regression model considering stratification factors from randomization.

“Race Non-White” includes patients with race other than White. “Hormone receptor status: ER and PR negative” includes patients without positive estrogen or positive progesterone receptors.

Source: HER2CLIMB Clinical Study Report.⁷

Overall Survival

At the time of this pre-specified analysis, the median duration of follow-up for the ITT population for OS was 14 months. A total of 130 patients (31.7%) in the tucatinib-combination group experienced an OS event compared with 85 patients (42.1%) in the placebo-combination group; 60% of all planned OS events had occurred by the time of this analysis. The median OS was 21.9 months (95% CI, 18.3 to 31.0) in the tucatinib-combination group compared with 17.4 months (95% CI, 13.6 to 19.9) in the placebo-controlled group. A statistically significant improvement in OS was observed among patients in the tucatinib-combination group, with a 34% reduction in risk of death compared with the placebo-combination group (HR = 0.66; 95% CI, 0.50 to 0.88; stratified log-rank P value 0.00480) (Figure 9).

At the updated analysis (data cut-off: February 8, 2021), the median follow-up for OS was 29.7 months in the tucatinib-combination group and 29.4 months in the placebo-combination group. The results of the post-hoc analysis of OS were consistent with the primary analysis and continued to show improved OS in the tucatinib-combination group, with a longer median OS of 24.7 months (95% CI, 21.6 to 28.9) versus the placebo-combination group, which showed a median OS of 19.2 months (95% CI, 16.4 to 21.4) (HR = 0.73; 95% CI, 0.59 to 0.90; nominal P value: < 0.004) (Figure 10).⁹

Figure 9: OS by Treatment Group (ITT Population) (Data Cut-Off: September 4, 2019)

Cap = capecitabine; CI = confidence interval; ECOG PS = Eastern Cooperative Oncology Group performance status; HR = hazard ratio; ITT = intention to treat; OS = overall survival; Pbo = placebo; Tra = trastuzumab; Tuc = tucatinib.

Statistically significant after adjustment for multiplicity. The threshold for statistical significance was 0.0074.

HR was computed from the Cox proportional hazards model using stratification factors (presence or history of brain metastases [yes/no], ECOG PS [0/1], and region of world [North America versus the rest of the world]) at randomization.

Two-sided P value based on stratified log-rank test and re-randomization procedure (Rosenberger and Lachin, 2002).

Source: HER2CLIMB Clinical Study Report.⁷

Figure 10: OS by Treatment Group (ITT Population) (Data Cut-Off: February 8, 2021)

Cape = capecitabine; CI = confidence interval; HR = hazard ratio; ITT = intention to treat; OS = overall survival; Pbo = placebo; Tras = trastuzumab; TUC = tucatinib.

Source: Curigliano et al. (2021).⁹

Subgroup Analysis for OS

A summary of subgroup analyses conducted at the time of the primary analysis for OS is depicted in Figure 11. Results for subgroup analyses favoured treatment with the tucatinib-combination over the placebo-combination in all groups, consistent with the overall results for OS, except for patients: aged 65 years and older, with positive hormone receptor status, without brain metastases, with an ECOG PS of 1, or who were from the rest of the world.⁷ Subgroup analyses were not adjusted for multiplicity and should be interpreted with caution.

Subgroup analyses for OS were also conducted at the post-hoc analysis and continued to show benefit for patients treated with tucatinib-combination therapy over the placebo-combination therapy. The results of the post-hoc subgroup analyses for OS were consistent with results of the primary analysis; tucatinib-combination therapy was favoured for all subgroups except for patients: aged 65 years and older, with positive hormone receptor status, without brain metastases, with an ECOG PS of 1, or who were from North America.⁹ Subgroup analyses were not adjusted for multiplicity.

Figure 11: Subgroup Analyses for OS (Data Cut-Off: September 4, 2019)

BICR = blinded independent central review; CI = confidence interval; ECOG PS = Eastern Cooperative Oncology Group; EDC = electronic data capture; ER = estrogen receptor; ITT = intention to treat; N = no; OS = overall survival; PR = progesterone receptor; Y = yes.

Note: Hazard ratio was calculated from a Cox regression model considering stratified factors from randomization.

“Race Non-White” included patients with race other than White.

“Hormone receptor status: ER and PR negative” included patients without positive estrogen or positive progesterone receptors.

“Baseline brain metastasis: Y” included patients with a history of brain metastases or presence of brain metastases or brain lesions of equivocal significance on screening MRI per EDC data.

Source: HER2CLIMB Clinical Study Report.⁷

Figure 12: Subgroup Analyses for OS (Data Cut-Off: February 8, 2021)

CI = confidence interval; ECOG PS = Eastern Cooperative Oncology Group; HR = hazard ratio; N = no; OS = overall survival; Y = yes.

Source: Curigliano et al. (2021).⁹

Objective Response Rate (per BICR)

At the primary analysis, based on assessment by BICR, more patients in the tucatinib-combination group than in the placebo-combination group had a complete response (0.9% versus 1.2%, respectively) or partial response (39.7% versus 21.6%). A greater proportion of patients in the placebo-combination group (14.0%) had a reported response progressive disease than patients in the tucatinib-combination group (7.9%). A greater proportion of patients in the tucatinib-combination group had a confirmed response (40.6%; 95% CI, 35.3 to 46.0) versus the placebo-combination group (22.8%; 95% CI, 16.7 to 29.8). Patients treated with the tucatinib-combination therapy showed statistically significantly improved ORR compared with the control group (P < 0.00008, stratified Cochrane-Mantel-Haenszel test).⁷

Other Secondary End Points

The following end points were included as other secondary end points: PFS_INV, ORR per investigator, DOR. These end points were not adjusted for multiple testing and are at risk for type I error. The results of these end points were supportive of the primary and key secondary end points, which favoured treatment with tucatinib-combination therapy compared with the placebo-combination.⁷ These results are summarized in Appendix 3.

Health-Related Quality of Life (EQ-5D-5L)

Assessment of HRQoL was conducted in the ITT population using the EQ-5D-5L, which was incorporated into the HER2CLIMB trial after a protocol amendment on August 30, 2017. Therefore, only patients who consented to version 7 of the trial protocol could be included in the analysis for HRQoL, which included 217 patients in the tucatinib-combination group and 113 patients in the placebo-combination group.⁷ There were no differences between the 2 treatment groups in the 5 domains of the descriptive system of the EQ-5D (anxiety/depression, mobility, pain/discomfort, self-care, and usual activities) (Figure 29, Figure 30, Figure 31, Figure 32, and Figure 33 in Appendix 3).⁷

Further, there were no differences in the mean EuroQol Visual Analogue Scale (EQ VAS) scores between treatment groups; the results of the EQ-5D-5L assessment suggested maintenance of HRQoL in both the tucatinib- and placebo-control groups (Figure 13).^7,25 A meaningful time to deterioration, based on the EQ VAS, was defined as a decrease of 7 or more points. There was a 19% reduction in the risk of meaningful deterioration of HRQoL based on the EQ VAS in the tucatinib-combination group compared with the placebo-combination group, although the difference in risk of meaningful deterioration between the 2 treatment groups was not considered meaningful (HR = 0.81; 95% CI, 0.55 to 1.18). The median time to deterioration was not reached in the tucatinib-combination group (95% CI, 7.6 to not estimable), while the median time to deterioration in the placebo-combination group was 5.8 months (95% CI, 4.3 to not estimable).²⁶

Similar results were reported for patients with brain metastases, which indicated no difference between the tucatinib- and placebo-combination groups.²⁶

Figure 13: EQ-5D-5L of Health Score (Data Cut-Off: September 4, 2019)

Cap = capecitabine; EQ-5D-5L = EuroQol 5-Dimensions 5-Levels questionnaire; Pbo = placebo; Tra = trastuzumab; Tuc = tucatinib.

Baseline was defined as most recent non-missing assessment on or before first dose date.

n/N: n is the number of patients who completed the survey. N is the number of patients who completed baseline survey and are still on study. Cycles where the number of patients in each arm remained ≥ 20% of initial cohort size are presented.

The length of the box represents the interquartile range (the distance between the 25th and 75th percentiles). The horizontal line in the box interior represents the group median. The whiskers extend to the group minimum and maximum values.

Source: HER2CLIMB Clinical Study Report.⁷

Harms

Only those harms identified in the review protocol are reported subsequently. Harms data are reported based mainly on the primary analysis (data cut-off: September 4, 2019). Additional data were provided by the sponsor based on a post-hoc analysis (data cut-off: February 8, 2021); however, longer-term data were generally consistent with the harms data of the primary analysis and are not included in this report.

Adverse Events

A summary of AEs from any cause is reported in Table 22. Most patients in both treatment groups reported an AE of any grade. In general, AEs were more commonly reported among patients in the tucatinib-combination group. The most common AEs of any grade in both the tucatinib-combination group and the placebo-combination group were diarrhea (80.9% versus 53.3%), hand-foot syndrome (63.4% versus 52.8%), nausea (58.4% versus 43.7%), fatigue (45.0% versus 25.4%), and vomiting (35.9% versus 25.4%); however, the proportion of patients experiencing these AEs was greater in the tucatinib-combination group.⁷

Table 23 summarizes AEs that were grade 3 or greater. In general, AEs that were grade 3 or greater were more commonly reported in the tucatinib-combination group. A total of 223 patients (55.2%) in the tucatinib-combination group experienced a grade 3 or greater AE compared with 96 patients (48.7%) in the placebo-combination group. In both the tucatinib-combination group and placebo-combination group, the most commonly reported AEs that were grade 3 or greater were hand-foot syndrome (13.1% versus 9.1%) and diarrhea (12.9% versus 8.6%).⁷

A time-at-risk exposure-adjusted analysis of grade 3 or greater AEs of hand-foot syndrome, diarrhea, and increased ALT and AST was performed to adjust for the longer exposure to treatment that patients in the tucatinib-combination group experienced, as these patients had a longer duration of treatment than patients in the placebo-combination group.⁷ The incidence of each grade 3 or greater AE was lessened after adjusting for exposure, resulting in the following:

• Crude incidence of grade 3 or greater hand-foot syndrome was 13.1% in the tucatinib-combination group and 9.1% in the placebo-combination group, with a time-at-risk exposure-adjusted incidence of 21 versus 19 per 100 person-years.

• Crude incidence of grade 3 or greater diarrhea was 12.9% in the tucatinib-combination group and 8.6% in the placebo-combination group, with a time-at-risk exposure-adjusted incidence of 21 versus 17 per 100 person-years.

• Crude incidence of grade 3 or greater ALT increase was 5.4% in the tucatinib-combination group and 0.5% in the placebo-combination group, with a time-at-risk exposure-adjusted incidence of 8 versus 1 per 100 person-years.

• Crude incidence of grade 3 or greater AST increase was 4.5% in the tucatinib-combination group and 0.5% in the placebo-combination group, with a time-at-risk exposure-adjusted incidence of 7 versus 1 per 100 person-years.

Serious Adverse Events

SAEs of any grade were reported in similar proportions of patients in the tucatinib- and placebo-combination groups (25.7% and 26.9%, respectively). SAEs that were treatment-related were reported among 10.9% of patients in the tucatinib-combination group versus 6.6% of patients in the placebo-combination group. Serious TEAEs related specifically to tucatinib or placebo, capecitabine, or trastuzumab also occurred in similar proportions across both treatment groups (Table 24). Of the treatments in each regimen, trastuzumab-related serious TEAEs were reported the least compared with tucatinib, placebo, and capecitabine.⁷

Mortality

Grade 5 AEs were reported in 8 patients (2.0%) in the tucatinib-combination group and 6 patients (3.0%) in the placebo-combination group. In the tucatinib-combination group, grade 5 AEs included sudden death due to unknown cause (n = 2), cardiac failure, sepsis, multiple organ failure, dehydration, cardiac arrest, and septic shock (n = 1 each). In the placebo-combination group, grade 5 AEs included respiratory failure, cardiac arrest, systemic inflammatory response syndrome, myocardial infarction, sepsis, and multiple organ failure (n = 1 each). Grade 5 AEs were considered by the investigator to be related to treatment with tucatinib or placebo in 2 patients in the tucatinib-combination group due to sepsis and dehydration, and 1 patient in the placebo-combination group due to sepsis.⁷

Notable Harms

Diarrhea

In the tucatinib-combination group, 80.9% of patients experienced diarrhea compared with 53.3% of patients in the placebo-combination group, which required antidiarrheal medication for 77.1% and 58.1% of patients, respectively; most events of diarrhea were grade 1 (43.3% versus 32.0%, respectively) or grade 2 (24.8% versus 12.7%). Events of diarrhea were reported to be manageable with antidiarrheal medication and dose modification. Antidiarrheal medication was required for 49.7% of cycles where diarrhea was reported, and the mean length of treatment with antidiarrheal medication per cycle was 5.0 days (SD = 4.6) in the tucatinib-combination group.⁷ Of all events of diarrhea that took place, most (79.6%) resolved with a mean and median time of 37 days (SD = 67.6) and 8.0 days, respectively. Grade 3 or greater events of diarrhea occurred in 12.9% of patients in the tucatinib-combination group compared with 8.6% in the placebo-combination group. Doses of tucatinib or placebo were held due to diarrhea for 13.9% of patients in the tucatinib-combination group and 8.6% of patients in the placebo-combination group. Dose reductions due to diarrhea were reported for 5.7% of patients in the tucatinib-combination group and 4.6% of patients in the placebo-combination group. Few patients permanently discontinued treatment due to diarrhea (1.0% and 0.5% in the tucatinib- and placebo-combination groups, respectively).⁷

Potential Drug-Induced Liver Injury

A total of 9 patients (2.2%) in the tucatinib-combination group and 2 patients (1.0%) in the placebo-combination group met criteria for potential drug-induced liver injury. Of these patients, 3 in the tucatinib group were considered at high risk for fatal drug-induced liver injury under Hy’s Law (defined as AST or ALT elevations > 3 × the upper limit of normal [ULN] with concurrent elevation [within 21 days of AST and/or ALT elevations] of total bilirubin > 2 × ULN and alkaline phosphatase).⁷

Cerebral Edema Not Attributable to Progression of Disease

While no patients in the tucatinib-combination group reported AEs related to cerebral edema, 2 patients (1.0%) in the placebo-combination group did.

Left Ventricular Ejection Fraction

Few patients in the HER2CLIMB trial had a reported event of left ventricular ejection fraction. A total of 7 patients (1.7%) and 4 patients (2.0%) in the tucatinib- and placebo-combination groups, respectively, reported left ventricular ejection fraction. Few events required dose modification or discontinuation (1.7% in the tucatinib-combination group and 2.0% in the placebo-combination group). Dose modification was able to resolve events in 5 of 7 patients in the tucatinib-combination group. One patient in the tucatinib-combination group discontinued treatment due to left ventricular ejection fraction and 1 patient discontinued because they experienced a grade 5 event of cardiac failure; neither event was considered related to tucatinib treatment.⁷

Creatinine Increase

Creatinine increase was observed in 13.9% of patients in the tucatinib-combination group versus 1.5% of patients in the placebo-combination group, with all events being either grade 1 or grade 2. A mean creatinine increase of approximately 30% was observed within the first cycle of treatment with tucatinib. Mean levels of creatinine were reported to remain stable while on treatment in the tucatinib-combination group, with creatinine levels returning to baseline values upon discontinuation of treatment. No impacts on renal function for patients were reported, as acute kidney injury and renal failure AEs were infrequently reported and in similar proportions across both treatment groups.⁷

Critical Appraisal

Internal Validity

The HER2CLIMB trial was an international, multi-centre, double-blind, placebo-controlled, phase II RCT. Patients in the tucatinib-combination group received 300 mg tablets, and patients in the placebo-combination group received a tablet that did not contain the active ingredient but was identical in appearance to maintain the blinding. The baseline demographic and clinical characteristics were balanced across the treatment groups overall and across important analysis populations (i.e., ITT and ITT-PFS populations). Patients were randomized based on the presence of brain metastases (yes versus no), ECOG PS (0 versus 1), and geographic region (US versus Canada versus the rest of the world). This helped to ensure the comparability between treatment arms in the subgroup analysis results according to each pre-specified stratification factor. In general, subgroup analyses favoured treatment with the tucatinib combination versus the placebo combination across all subgroups.

The original protocol of the HER2CLIMB trial was dated August 11, 2015 and was amended 9 times. None of the amendments were considered to have negatively affected study outcomes, as none of the amendments were made before the unblinding of data. Also, the study amendments were not considered to have affected the interpretation of the study results. Important protocol deviations occurred in similar proportions across both the tucatinib- and placebo-combination groups (3.4% versus 3.5%). Important protocol deviations were due to dosing (n = 4), eligibility criteria (n = 4), missed assessments (n = 3), safety (n = 2), consent (n = 1), and other reasons (n = 8). As there were few important protocol deviations, and as both treatment groups reported similar proportions of important protocol deviations, it is unlikely that such deviations had a significant impact on the patient efficacy or safety analyses.

Patients were randomized in the HER2CLIMB trial using a hierarchical randomization scheme via interactive response technology. Patients were randomized based on the presence of brain metastases (yes versus no), ECOG PS (0 versus 1), and geographic region (US versus Canada versus the rest of the world). The sponsor also included specifications for a biased-coin assignment in the randomization scheme to prevent imbalances between treatment groups and any given hierarchical level (i.e., overall treatment group balance, then treatment group balance within each stratification factor). The methods for randomization were considered by the CADTH team to result in a low risk of bias, as patients were randomly assigned to a treatment group using technology that prevented unblinding of investigators and ensured that relatively equal proportions of patients would be randomized not only to each treatment group, but to each pre-specified stratification factor. The resulting baseline characteristics of the trial showed well-balanced demographic and clinical characteristics across both treatment groups in the HER2CLIMB trial and across important analysis populations (i.e., ITT and ITT-PFS populations). Therefore, there were no imbalances across treatment groups or analysis populations, which could have resulted in important differences in patient outcomes.

Geographic region was pre-specified as 1 of the stratification factors of the HER2CLIMB trial, stratifying patients into the following categories: US, Canada, and the rest of the world. The clinical experts consulted by CADTH indicated that stratification by geographic region is likely reflective of differences in treatment availability across the globe. Mainly, differences in post- and pre-protocol treatments may differ across countries and it is possible that the efficacy of tucatinib may be affected based on the treatment patterns of patients in these regions. It is unclear how differences in treatment patterns across the globe may affect patients who received treatment with tucatinib in the HER2CLIMB trial.

In general, the subgroup analyses favoured treatment with the tucatinib-combination group versus the placebo-combination group. However, it should be acknowledged that while the subgroups for the subgroup analyses were pre-specified, they were not adjusted for multiplicity, not powered to detect differences, and may be indicative of imprecision due to wide CIs. The lack of adjustment for subgroup analyses may increase the likelihood of type I error, resulting in an increased likelihood of detecting a treatment effect when 1 may not be present; as such, subgroup analyses should be considered descriptive and interpreted with caution.

Statistical analyses of OS and PFS were conducted using Cox proportional hazards models, which rely on the assumption of proportional hazards in both treatment groups. The sponsor conducted an assessment of proportional hazards by performing tests for statistical significance using log(-log(Survival)) plots. An analysis of proportional hazards did not reveal any significant violation of the assumption for either the OS or PFS analyses.

The analyses of the primary end point and key secondary end points were conducted hierarchically, whereby key secondary end points (PFS_BM and OS) were tested only if PFS was found to be statistically significant at the primary analysis. In addition, the sponsor pre-specified alpha to be split across key end points, and pre-specified the boundaries for statistical significance. Alpha values were allocated to key secondary end points, depending on whether the statistical significance of the primary end point and each key secondary end point of the trial was achieved. As all primary and key secondary end points were statistically significant at the time of the primary analysis, the results were considered final and no further formal testing were specified.

Other secondary end points of the trial included ORR, PFS_INV, and DOR. No adjustments for multiple testing were specified for these secondary end points. It is possible that analyses of other secondary end points are at greater risk for inflated type I error. The results of the other secondary end points are supportive of the results for the primary and key secondary end points, which favour treatment with tucatinib in combination with trastuzumab and capecitabine over placebo in combination with trastuzumab and capecitabine.

The sponsor conducted a post-hoc analysis that provided an additional 15.6 months of follow-up. This resulted in a total follow-up of 29.7 months for the tucatinib-combination group and 29.4 months for the placebo-combination group and provided additional efficacy (OS, PFS) and safety data. Of note, after the primary analysis, the trial was unblinded and assessments for PFS were conducted by the investigator. The results of the post-hoc analysis were consistent with the results of the primary analysis, which remained blinded and which used BICR for assessment of PFS.

It is possible the choice of subsequent therapies could have affected efficacy assessments of OS, as analyses for OS included patients who received subsequent therapies. A total of 202 patients (69.2%) in the tucatinib-combination group and 139 patients (79.4%) in the placebo-combination group received subsequent anti-cancer therapies. There were disproportional differences noted between treatment groups in the types of subsequent anti-cancer therapies received, as more patients in the placebo-combination group received antibody (57.1% versus 50.0%, respectively) and TKI (24.0% versus 16.8%) anti-HER2 regimens, and trastuzumab (12.2% versus 5.4%), while more patients in the tucatinib-combination group than in the placebo-combination group received trastuzumab plus chemotherapy (20.8% versus 15.8%, respectively). The differences in subsequent therapies are expected to introduce bias in the efficacy analyses of OS and other patient outcomes. However, the direction and extent of the biases are difficult to predict.

The analyses of PFS involved the censoring of patients who began new anti-cancer treatment before disease progression or death. Censoring patients for subsequent therapies is considered a form of informative censoring, especially considering the much greater proportion of patients in the placebo-combination group who were censored at the primary analysis for PFS compared with the tucatinib-combination group (74.6% versus 52.8%, respectively).^7,27 Sensitivity analyses for PFS that did not censor for patients receiving subsequent therapies were not conducted, and this definition of censoring was deemed conservative in the assessment of PFS. Post-hoc analysis revealed more equal proportions of patients in the tucatinib- and placebo-combination groups who eventually discontinued from study treatment and reported receiving subsequent therapies (76.8% versus 81.2%, respectively).⁹ The post-hoc analyses of PFS continued to support improved efficacy of tucatinib-combination over the placebo-combination therapy.

No treatment crossover was permitted within the HER2CLIMB trial.⁷ However, based on specific inclusion and exclusion criteria, patients could have crossed over from the placebo-combination group to the tucatinib-combination group after the double-blinded phase of the HER2CLIMB trial, which occurred after the primary analysis. There were 26 patients from the placebo-combination group who crossed over to the tucatinib-combination group, potentially biasing results against the tucatinib-combination group. As a minimal number of patients crossed over to the tucatinib-combination group, it is unlikely that the impact of crossover affected the efficacy results of the trial.

HRQoL was assessed using the EQ-5D-5L and was considered an exploratory end point for the HER2CLIMB trial; the results revealed no differences between the tucatinib- and placebo-combination groups. Analyses of patient-reported outcomes were conducted using the EQ-5D-5L, which previously demonstrated responsiveness among breast cancer patients following curative treatment, although small changes in health were not recognized as being meaningful.^28,29 Previous evidence also suggests that the EQ-5D-5L shows both reliability and validity for use among patients living with breast cancer.³⁰ However, none of the previously identified evidence included patients with brain metastases. The EQ-5D-5L is a generic tool used to measure HRQoL of patients across many diseases; the validity and reliability of this tool has been demonstrated across many patient populations, including breast cancer. It is possible that small differences in HRQoL that may be important to patients living with cancer were not detectable with the EQ-5D-5L.²⁹

A summary of dose exposure showed that patients in the tucatinib-combination group received treatment with tucatinib, capecitabine, and trastuzumab for longer mean periods of time than the patients in the placebo-combination group, who received placebo, capecitabine, and trastuzumab. Dose exposure was measured by the number of treatment cycles initiated among patients who received at least 1 dose. The total mean cumulative dose of trastuzumab was also greater in the tucatinib-combination group than in the placebo-combination group. The mean duration of exposure to tucatinib or placebo was longer in the tucatinib-combination group than in the placebo-combination group (7.6 months versus 5.6 months, respectively). The mean duration was also longer in the tucatinib-combination group than in the placebo-combination group for trastuzumab (7.3 months versus 5.4 months, respectively) and capecitabine (7.9 months versus 5.7 months). Overall, these differences likely resulted in a better treatment effect in the tucatinib-combination group than its counterpart. The longer treatment duration should be considered when interpreting HRQoL data, as longer exposure to treatment may influence toxicities and quality of life for patients.

Concomitant medications were administered in generally similar frequencies across both HER2CLIMB treatment groups, with most patients in both groups requiring a concomitant medication (99.3% in the tucatinib-combination group versus 97.5% in the placebo-combination group). However, more patients in the tucatinib-combination group required antipropulsive medications compared with the placebo-combination group (65.6% versus 36.5%, respectively), and more patients in the placebo-combination group than in the tucatinib-combination group received natural opium alkaloids (33.5% versus 23.8%, respectively).⁷ Consultation with clinical experts confirmed it is possible that differences in concomitant medications could result in differences in patient outcomes, particularly symptom or pain controls and HRQoL. However, the effect of concomitant medications on patients is unclear.

External Validity

The protocol of the HER2CLIMB trial stated that efforts were made to avoid radiation or surgery to target lesions in the brain in the absence of disease progression (per RECIST v1.1 criteria) unless clinically necessary in the opinion of the investigator. This was done to ensure accurate assessments for PFS, as target lesions cannot be adequately assessed for subsequent response to systemic therapy once treated with local CNS therapy and could, therefore, interfere with efficacy assessments. The clinical experts consulted by CADTH confirmed that avoidance of radiation or surgery to target lesions in the brain is not conducted in clinical practice. The clinical experts stated that radiation or surgery are typically preferentially chosen for the treatment of patients with progressive brain lesions, as previous systemic therapies in this setting were not associated with meaningful intracranial activity.

The protocol of the HER2CLIMB trial also stated that all patients underwent imaging for brain metastases at baseline. The clinical experts consulting with CADTH confirmed it is not typical clinical practice to scan all breast cancer patients for brain metastases upon initial assessment. Patients would undergo imaging only when brain metastases are suspected.

Efficacy assessments were performed once every 6 weeks for the first 6 months while on study and then once every 9 weeks. The sponsor stated that these assessment intervals were consistent with standard-of-care practice and the schedules used in prior studies in similar populations.⁷ The clinical experts consulted for this review stated that the assessments of patients in the HER2CLIMB trial were conducted more frequently than what might be typical in Canadian clinical practice; in general, imaging for patients occurs after approximately 4 cycles of treatment, or every 9 to 12 weeks. Assessments may occur more frequently for patients who require greater observation. The increased frequency of assessments in the HER2CLIMB trial may have allowed for greater detection of patients’ treatment response. It is not expected that the frequency of assessments in the HER2CLIMB trial negatively impacted patient outcomes; however, the frequency of assessments is likely too frequent and not reflective of clinical practice.

The interventions of the HER2CLIMB trial included tucatinib, trastuzumab, and capecitabine, which were administered at 300 mg orally twice daily, 6 mg/kg once every 21 days (with a loading dose of 8 mg/kg), and 1,000 mg/m² administered orally twice daily on days 1 to 14 of each 21-day cycle, respectively. These dosages align with the Health Canada–approved dosages for each treatment and align with what patients are administered in clinical practice. The baseline characteristics of patients in the HER2CLIMB trial included mostly women from North America or the rest of the world, good performance status (ECOG PS 0 or 1), and presence of metastases (including CNS and non-CNS metastases). While the trial included highly selective and mostly White patients, based on input from the clinical experts consulted by CADTH, the baseline characteristics of the HER2CLIMB trial were generally considered to be representative of Canadian patients. However, it should be noted the eligibility criteria of the HER2CLIMB trial were highly restrictive and may not capture all patients who may be treated with tucatinib-combination therapy in clinical practice. It is unclear whether the treatment effects of tucatinib-combination therapy will be generalizable to patients who were excluded from the HER2CLIMB trial. Key outcome measures of the trial included PFS and OS. These outcomes were identified by clinical experts consulted by CADTH as being important to patients. ORR was also captured as a secondary end point in the HER2CLIMB trial; response rates may also be useful for capturing the responsiveness of patients to treatment. In general, outcomes were clinically relevant and were considered appropriate measures for the assessment of patient outcomes. Overall, the interventions, patient characteristics, and outcomes measures were all considered generalizable to the Canadian context.

The HER2CLIMB trial enrolled mostly female patients but also enrolled 3 male patients (0.7%) in the tucatinib-combination group and 2 male patients (1.0%) in the placebo-combination group. Males are also at risk of developing breast cancer and are often excluded from clinical trials. The HER2CLIMB trial enrolled male patients in addition to women. While the number of male patients is low, generalizability to male patients should be considered and tucatinib in combination with trastuzumab and capecitabine should be provided to male patients with HER2-positive MBC as a treatment option.

Within the HER2CLIMB trial, both trial groups received trastuzumab, which could have been administered to patients in varying formulations. Patients may have been administered trastuzumab intravenously or subcutaneously. In addition, the IV and subcutaneous formulations of trastuzumab are administered at different dosages: 6 mg/kg once every 21 days (with a loading dose of 8 mg/kg) and 600 mg every 21 days (without a loading dose, respectively).⁷ While no direct evidence exists to compare the efficacy of the 2 formulations and doses of trastuzumab, the clinical experts agreed the efficacy of both versions of trastuzumab should be equivalent. Additionally, the clinical experts also confirmed that the IV formulation of trastuzumab is more commonly used among Canadian patients, and consultation with the Provincial Advisory Group confirmed that the biosimilar formulation of trastuzumab would be used in most Canadian jurisdictions.

Standard first-line therapies for patients with MBC may include treatment with pertuzumab in combination with trastuzumab and taxane followed by pertuzumab plus trastuzumab. Second-line therapies for these patients may then include T-DM1. Eligibility criteria in the HER2CLIMB trial specified that all patients must have had prior treatment with trastuzumab, pertuzumab, and T-DM1. Therefore, the patient population of patients in the HER2CLIMB trial is likely reflective of patients in the Canadian population and treatment algorithms standard in Canadian clinical practice. Prior treatment with trastuzumab, T-DM1, and pertuzumab was not required to have been specifically in the metastatic setting, although most patients did receive each drug in the metastatic setting, with some patients receiving it in both the neoadjuvant or adjuvant and metastatic setting, and few patients receiving prior therapy in the neoadjuvant or adjuvant setting only. The sponsor noted that the treatment landscape for HER2-positive breast cancer patients has changed drastically since the completion of the patient enrolment for the HER2CLIMB trial.¹⁰ During patient enrolment, T-DM1 was approved for and used only in the metastatic setting; however, since completion of patient enrolment, T-DM1 has been approved for use in the adjuvant setting. Almost all patients in the HER2CLIMB trial (> 98%) reported having received prior therapy with T-DM1 in the metastatic setting only.⁷ It is expected that a greater proportion of patients in clinical practice will have received prior therapy with T-DM1 in other treatment settings as well.

Treatments in the HER2CLIMB trial included tucatinib or placebo, trastuzumab, and capecitabine. The doses and administration schedules in the trial were in alignment with Health Canada and Canadian clinical practice. At the time the HER2CLIMB trial was conducted, the combination of trastuzumab and capecitabine was considered an appropriate comparator based on phase III data from the CEREBEL trial. This regimen supported improved PFS and OS for patients compared with lapatinib plus capecitabine therapy.¹⁹ Therefore, the choice of treatment comparator was considered appropriate, given the evidence at the time the HER2CLIMB trial was being conducted. However, the clinical experts consulted by CADTH highlighted that the results of the CEREBEL trial may not be conclusive and that lapatinib plus capecitabine might also be a relevant comparator for tucatinib-combination therapy. No direct evidence exists to compare the efficacy and safety of tucatinib in combination with trastuzumab and capecitabine versus lapatinib plus capecitabine. Regarding toxicity, the clinical experts commented that there may be fewer or equal rates of some AEs, such as diarrhea and nausea, with tucatinib-combination therapy, as lapatinib and capecitabine are associated with more of these toxicities than trastuzumab plus capecitabine.

In addition, the HER2CLIMB trial allowed for enrolment of patients with brain metastases who comprised 48.3% of the total study population and, traditionally, this subgroup of patients has been excluded from clinical trials. The inclusion of patients with brain metastases is highly relevant, as many patients with HER2-positive MBC develop brain metastases.³¹ The HER2CLIMB trial demonstrated statistically significantly improved efficacy in the tucatinib-combination group compared with the placebo-combination group, even with patients with brain metastases who may face worse prognoses. In addition, subgroup analyses for PFS and OS supported treatment with tucatinib-combination therapy over the placebo-combination treatment. The clinical experts confirmed that the patient population of the HER2CLIMB trial was reflective of Canadian patients and supported the results of the HER2CLIMB trial that suggested treatment with tucatinib in combination with trastuzumab and capecitabine for patients with brain metastases.

In the Health Canada–approved product monograph, tucatinib in combination with trastuzumab and capecitabine is indicated for patients who have received at least 1 prior HER2-directed therapy in the metastatic setting.² The treatment landscape for patients with MBC is complex and has changed to include new HER2-directed treatments, such as pertuzumab and T-DM1. Patients in the HER2CLIMB trial reported having received a mean of 3 prior therapies in the metastatic setting, and the sponsor confirmed that every patient in the HER2CLIMB trial received at least 1 prior therapy in the metastatic setting.^3,7 Therefore, it was considered appropriate, given the changes to the treatment landscape for this setting and the characteristics of patients in the HER2CLIMB trial, that treatment with tucatinib in combination with trastuzumab and capecitabine be used for patients who received at least 1 HER2-targeted therapy in the metastatic setting.

The HER2CLIMB trial eligibility criteria required patients to have prior treatment with trastuzumab, pertuzumab, and T-DM1, alone or in combination, and most patients (> 90%) reported having received each treatment. The median and mean number of therapies used among patients in the HER2CLIMB trial was 4, with most patients having received trastuzumab, pertuzumab, and T-DM1 in either the metastatic setting or in the metastatic and neoadjuvant or adjuvant setting. Therefore, patients would have received tucatinib-combination therapy in the second- or later-line setting. It may be unreasonable to suggest using tucatinib-combination therapy as a first-line treatment option for patients with MBC, as there is no evidence to support the use of this treatment in this context. The input received from the clinical expert consulted by CADTH and the Canadian clinician groups providing input on this submission suggests that tucatinib-combination therapy would most likely be used as a third-line therapy.

Regarding treatment discontinuations, the HER2CLIMB trial did not allow patients to continue the remaining study treatments if discontinuation was required for both trastuzumab and capecitabine (concurrently), or tucatinib or placebo. However, patients who experienced toxicity from capecitabine or trastuzumab were permitted to discontinue 1 of these drugs and continue on tucatinib or placebo in combination with either capecitabine or trastuzumab. The clinical experts consulted by CADTH agreed the discontinuation criteria of the trial were acceptable.

Indirect Evidence

Objectives and Methods for the Summary of Indirect Evidence

The HER2CLIMB trial compared tucatinib in combination with trastuzumab and capecitabine with placebo in combination with trastuzumab and capecitabine for patients with locally advanced, unresectable, or metastatic HER2-positive breast cancer, including patients with brain metastases, who have received prior treatment with trastuzumab, pertuzumab, and T-DM1, separately or in combination. In Canada, numerous treatment options are available for such patients in the second- or later-line setting. The objective of this section is to summarize and critically appraise available indirect evidence comparing tucatinib in combination with trastuzumab and capecitabine with relevant treatments for HER2-positive MBC patients.

The sponsor conducted a systemic literature review to identify relevant studies for its submitted ITC. The search for the systematic literature review was performed on January 22, 2020 for electronic databases (Table 25) and on January 27, 2020 for conference websites. Relevant systematic reviews and meta-analyses were handsearched for relevant publications that were not identified in the electronic searches.¹¹ It was noted by the sponsor that the protocol for the systematic literature review was amended in March 2020; the original protocol specified searching for literature pertaining to treatments administered to HER2-positive MBC patients who received at least 2 prior anti-HER2 regimens in any setting. However, due to limited studies fitting these criteria, the protocol was expanded to identify at least 1 prior anti-HER2 regimen in any setting.¹¹

Description of the Indirect Comparison

The sponsor-submitted ITC was summarized and appraised. A supplemental search of the medical literature for publicly available ITCs was conducted by CADTH staff, and no additional ITCs were identified that evaluated the efficacy or safety of tucatinib in combination with trastuzumab and capecitabine against relevant comparators for HER2-positive MBC patients.

Methods of the ITC

Objectives

The aim of the submitted ITC was to compare the efficacy of tucatinib and comparators of interest among adult patients with HER2-positive MBC who have had at least 1 prior anti-HER2 regimen in any setting.

Study Selection Methods

A literature search was conducted based on details in Table 25. Studies were screened by title and abstract followed by full-text screening, resulting in the inclusion of 39 studies that were considered for inclusion in the sponsor’s ITC; of those studies, 25 were not included in the ITC due to a lack of connected networks or lack of a connection between tucatinib and a comparator of interest (n = 24), or because the trial was a dose comparison study (n = 1). The sponsor’s systematic literature search identified 14 trials that included patients who had been treated previously with at least 1 anti-HER2 regimen (Figure 14).¹¹

Figure 14: PRISMA Diagram of Systematic Literature Review for Sponsor's ITC

HER2 = human epidermal growth factor receptor 2; ITC = indirect treatment comparison; PRISMA = Preferred Reporting Items for Systematic Reviews and Meta-Analyses.

Note: This figure demonstrates the results of the sponsor’s systematic literature search based on the initial criteria of patients receiving at least 2 prior anti-HER2 regimens. Once the criteria were updated to include patients who had received at least 1 prior anti-HER2 regimen, a total of 14 trials were identified by the systematic literature search and included into the sponsor’s ITC.

^a The category “Other” includes duplicate references and conference abstracts published before 2018.

Source: Sponsor’s ITC.¹¹

ITC Analysis Methods

Details of the methodology for the sponsor’s ITC are reported in Table 26.

Four trials reported treatment crossover (EGF100151,³² EMILIA,¹⁶ GBG-26,³³ Jiang et al.³⁴). The sponsor used the rank-preserving structural failure time model to adjust for treatment crossover.¹¹

Progression-free survival: The LANTERN³⁵ and TRAXHER2³⁶ trials did not report HRs for PFS, but reported median PFS for both the experimental and control groups. Methods reported by Hackshaw³⁷ were used by the sponsor to estimate HRs based on reported median PFS data. Therefore, all 14 trials were used in the network of evidence for PFS.¹¹

Overall survival: HRs were reported for 10 of the 14 trials. It was noted that the LANTERN³⁵ study did not report an HR for OS but did report data that were used to estimate the hazard for death in the trial. The following trials were not used for the network of evidence for OS in the NMA:

• The TRAXHER2³⁶ study did not report median survival in either treatment group, as the 50% survival rate threshold was not reached for the comparator group.

• The statistical analysis methods reported in the GBG-26³³ trial were not considered to robustly account for treatment switching. Almost half of all patients in the GBG-26 trial had crossover anti-HER2 treatment after receiving the treatment they were randomized to in the trial (capecitabine or trastuzumab plus capecitabine). The authors of the GBG-26 trial reported the analyses that were conducted excluded patients who received crossover treatment, but that the power of the analyses was limited due to the low number of patients (N = 31).³³ Therefore, to minimize the amount of bias being introduced into the ITC, the sponsor did not include this trial in the analysis for OS in the NMA.

• OS data were not reported in either the Jiang et al.³⁴ or Ma et al.³⁸ studies.

Best tumour response: The original probit model was used in the analysis of best tumour response, which analyzed patients’ best responses (complete response, partial response, stable disease, or progressive disease). Ten of the 14 trials were used for the analysis of best tumour response. It was noted by the sponsor that 7 of the 10 studies included in this analysis evaluated best tumour response for patients with measurable lesions at baseline; for 3 trials (Jiang et al.,³⁴ GBG-26,³³ TRAXHER2³⁶), there was no clear statement regarding measurable disease at baseline, but it was assumed by the sponsor that this was the case.¹¹

The sponsor considered prior exposure to HER2-targeted therapies, differences in line of therapy, and patients with brain metastases as potential effect modifiers. However, no analysis was deemed able to robustly account for the differences in these effect modifiers due to the lack of available data.

The proportional hazards assumption was reported not to be met for 3 studies included in the network of evidence for PFS and 1 study included in the network of evidence for OS. Fractional polynomials were fitted to reconstructed patient-level data to further examine whether proportional hazards assumptions were met. Patient-level data were reconstructed from publications of trials included in the ITC network using Kaplan–Meier survival curves and the number of patients at risk. Available data for reconstruction of patient-level data were available from 11 trials for the analysis of PFS and 8 trials for the analysis of OS. The sponsor conducted NMAs based on HRs and based on the fractional polynomial method and compared results for consistency. However, the primary analyses for both OS and PFS were based on HR data.¹¹

Random-effects fractional polynomial models were also conducted as sensitivity analyses. Random-effects models were compared with fixed-effects models but revealed difficulty with model convergence, inconsistency between direct trial data and results from random-effects Bayesian modelling, and generally had higher deviance information criterion (DIC) values compared with fixed-effects models. Overall, fixed-effects models were considered more robust.¹¹ Frequentist random-effects models were performed when there was at least 1 closed-loop or duplicate comparison. Bayesian random-effects models were performed with the use of informative priors presented by Turner et al. ^11,39,

Results of the ITC

Summary of Included Studies

The following trials were included for analyses in the sponsor’s ITC: CEREBEL,¹⁹ EGF100151,³² ELTOP,⁴⁰ EMILIA,¹⁶ GBG-26,³³ HER2CLIMB,²² LANTERN,³⁵ NALA,⁴¹ Martin et al.,⁴² Ma et al.,³⁸ Jiang et al.,³⁴ PHEREXA,⁴³ SOPHIA,⁴⁴ and TRAXHER2.³⁶

A summary of the overall network of trials is included in Table 27, which depicts the number of trials by treatment of interest. Most treatments of interest had only 1 or 2 trials that had information available for comparisons in the NMA, except for lapatinib plus capecitabine and trastuzumab plus capecitabine, which had 8 and 7 trials, respectively, with information available to inform on treatment efficacy.

Study Characteristics

A summary of the trial-level characteristics of the studies included in the sponsor’s ITC is reported in Table 28. Trials were phase II or III randomized controlled trials. Except for 2 studies, HER2CLIMB²² and Jiang et al.,³⁴ all trials were open label. Most trials (n = 12) reported prior trastuzumab treatment for nearly 100% of patients, with the remaining 2 trials reporting approximately half of patients receiving prior trastuzumab; the majority of trials involving prior treatment with trastuzumab was included as eligibility criteria for the trial. Most trials included patients who had received at least 1 prior therapy in the metastatic setting. Two trials^36,40 did not report information on the prior therapies received by patients in the metastatic setting. Most trials enrolled patients of many races, although most patients were White (n = 7); the remaining trials enrolled Chinese (n = 2) or Japanese patients (n = 1). The race of patients was not reported in 4 trials. Treatment crossover was reported for 4 of the studies; for 3 of these trials (EGF100151,³² EMILIA,¹⁶ and Jiang et al.³⁴), switching to the investigational treatment from the comparator group was permitted after patients had reported disease progression, while the fourth trial, GBG-26,³³ re-randomized patients from both treatment groups to receive either chemotherapy or chemotherapy combined with trastuzumab or lapatinib for treatment in the third line.

Baseline Characteristics

Baseline characteristics of patients in each treatment group of the trials included in the sponsor’s ITC are reported in Table 29. These trials enrolled between 43 and 271 patients in each treatment group; the LANTERN³⁵ trial enrolled only 30 patients across both treatment groups, as recruitment was slow and never reached the target sample size for the trial; it was thus underpowered for detection of treatment differences across treatment groups. To avoid selection bias, the LANTERN³⁵ trial was still included in the sponsor’s NMA. All trials, except for the HER2CLIMB trial,²² either enrolled few patients (less than 20%) with brain metastases or did not allow patients with brain metastases to enroll, as per eligibility criteria. Of the trials with available information, the median time to diagnosis was between 30 and 50 months. The median age was similar across most trials, with patients reporting a median age of between 48 and 59 years of age. In most trials, less than 50% of patients had an ECOG PS of 1; the CEREBEL,¹⁹ Ma et al.,³⁸ and Jiang et al.³⁴ trials included treatment groups in which more than 50% of patients had an ECOG PS of 1. Only the CEREBEL¹⁹ trial comprised mainly patients with an ECOG PS of 1. Across treatment groups and trials, between 1-third and 2-thirds of patients, approximately, reported estrogen receptor positivity.

In general, baseline characteristics were balanced across treatment groups in all of the trials included in the ITC. However, differences were observed between treatment groups regarding the proportion of patients with an ECOG PS of 1 in the ELTOP,⁴⁰ Ma et al.,³⁸ LANTERN,³⁵ and Jiang et al.³⁴ trials. In addition, the proportions of patients reporting estrogen receptor and/or progesterone receptor positivity varied in the Ma et al.³⁸ trial.

Risk-of-Bias Assessment

The sponsor also conducted a risk-of-bias assessment for each trial that categorized the following components of each study as low, medium, or high risk of bias: method of randomization, allocation concealment, balance across treatment groups, selective reporting, population representative of other studies, sample size, re-treatment switching, and other. In all categories, most studies reported either a low or medium risk of bias.¹¹ Overall, 6 of the trials were categorized as having a low risk of bias (EGF100151,³² EMILIA,¹⁶ HER2CLIMB,²² Martin et al.,⁴² NALA,⁴¹ and PHEREXA⁴³). The CEREBEL¹⁹ and SOPHIA⁴⁴ trials were categorized as having a medium risk of bias, and the ELTOP,⁴⁰ GBG-26,³³ Jiang et al.,³⁴ LANTERN,³⁵ Ma et al.,³⁸ and TRAXHER2³⁶ studies were classified as having a high risk of bias. Due to the sponsor’s risk-of-bias assessments, the following conclusions were made:

• The LANTERN³⁵ and TRAXHER2³⁶ trials were reported by the sponsor to be of particularly poor quality due to early termination with a small sample size; however, these studies were not excluded to avoid publication bias and were included in the sponsor’s ITC.

• Three studies recruited only Chinese or Japanese patients (ELTOP,⁴⁰ Jiang et al.,³⁴ Ma et al.³⁸); these 3 studies also had small sizes and issues with treatment crossover and selective reporting. The sponsor did not include these 3 studies in the main analysis of their ITC but did conduct sensitivity analyses that included these 3 trials.

• Subgroup data were used from the CEREBEL,¹⁹ Ma et al.,³⁸ and SOPHIA⁴⁴ studies. CEREBEL¹⁹ and SOPHIA⁴⁴ were reported by the sponsor as having a medium risk of bias; while these studies did not have a high risk of bias, the use of subgroup-level data, which could lead to imbalanced between treatment groups, resulted in a classification of a medium risk of bias.

Further details regarding the sponsor’s risk-of-bias assessment are reported subsequently.

Regarding the method of randomization, all studies were categorized by the sponsor as having a low or medium risk of bias; it should be noted that trials that did not report details for methods of randomization were categorized as having a medium risk of bias. Regarding allocation concealment, all but 2 trials, the HER2CLIMB²² trial and a study by Jiang et al.³⁴ (which were double-blind randomized controlled trials that were categorized as having a low risk of bias), had a medium risk of bias, as they were all open label. Some differences in baseline characteristics across trials are noted earlier; the ELTOP,⁴⁰ GBG-26,³³ and PHEREXIA⁴³ trials as well as the study by Ma et al.³⁸ were categorized as having a high risk of bias. The LANTERN,³⁵ TAXHER2,³⁶ and Jiang et al.³⁴ studies did not report baseline characteristics. The following studies reported a high risk of bias for selective reporting, while the remaining all reported a low risk of bias: CEREBEL,¹⁹ GBG-26,³³ Jiang et al.,³⁴ LANTERN,³⁵ Ma et al.,³⁸ and TAXHER2.³⁶

When assessing whether study populations were representative of other studies, the following studies were considered to have a low risk of bias, as they included patients who had progressed on prior therapy with trastuzumab: ELTOP,⁴⁰ EGF100151,³² EMILIA,¹⁶ GBG-26,³³ Jiang et al.,³⁴ LANTERN,³⁵ Martin et al.,⁴² PHEREXA,⁴³ and TAXHER2.³⁶ Studies that had a medium or high risk of bias included studies that enrolled mixed populations of patients with and without prior treatment with trastuzumab (CEREBEL,¹⁹ Ma et al.³⁸), enrolled patients from specific subgroups (i.e., only brain metastases) (LANTERN³⁵), enrolled patients who had received multiple (≥ 2) previous lines of therapy (NALA,⁴¹ SOPHIA, HER2CLIMB), and studies that included patients of races that differed from populations in the majority of studies (ELTOP,⁴⁰ Jiang et al.³⁴).

Studies with a medium or high risk of bias in terms of sample size included those with medium or low sample sizes (ELTOP,⁴⁰ GBG-26,³³ TRAXHER2,³⁶ LANTERN,³⁵ Ma et al.³⁸). As mentioned previously, most studies did not allow for crossover except for the EGF10015,³² EMILIA,¹⁶ GBG-26,³³ and Jiang et al.³⁴ studies. Specifically, 36% of patients in the EGF100151 trial³² switched from treatment with capecitabine monotherapy to lapatinib plus capecitabine, 27% of patients receiving lapatinib plus capecitabine crossed over to receive T-DM1 in the EMILIA trial,¹⁶ and 76% of patients in the capecitabine group switched over to pyrotinib after disease progression. In the GBG-26 trial,³³ patients who experienced disease progression were re-randomized to receive third-line therapy with either chemotherapy or chemotherapy combined with trastuzumab or lapatinib; in the trastuzumab plus capecitabine group, 51% of patients received chemotherapy, while 27% of patients in the capecitabine group received chemotherapy plus lapatinib or trastuzumab. Information about treatment crossover was not reported for the TRAXHER2³⁶ trial.¹¹

No “other” sources of bias were noted for any trials except for the SOPHIA⁴⁴ trial, where the sponsor noted that patients could have received the physician’s choice of capecitabine, eribulin, gemcitabine, or vinorelbine to be combined with either margetuximab or trastuzumab, and that approximately 27% of patients received capecitabine.¹¹ Upon review by the CADTH team, it was noted that the proportion of patients who received each type of chemotherapy was similar across both treatment groups in the SOPHIA⁴⁴ trial.¹¹

Assessment of Proportional Hazards

Progression-Free Survival: Visual inspection of log(-log(Survival)) plots for PFS indicated that the Ma et al.,³⁸ Martin et al.,⁴² NALA,⁴¹ ELTOP,⁴⁰ and EMILIA¹⁶ studies showed deviation of the proportional hazards assumption. Results for significance testing of the proportional hazard assumption for PFS are reported in Table 30. Three studies (Martin et al.,⁴² NALA,⁴¹ and EMILIA,¹⁶) showed violation of the proportional hazard assumption.¹¹

Overall Survival: Visual inspection of log(-log(Survival)) plots for OS by the sponsor suggested that deviation of the proportional hazards assumption was present in the following studies: Martin et al.,⁴² PHEREXA,⁴³ and Takano et al.⁴⁰ Results of significance testing for the proportional hazard assumption are reported in Table 31. Only Martin et al.⁴² showed deviation from the proportional hazard assumption.

Progression-Free Survival

The network of evidence for PFS in the NMA is depicted in Figure 15. The sponsor reported that the network for PFS contained minimal information on heterogeneity in the form of 1 closed loop (between the following treatments: capecitabine, lapatinib plus capecitabine, and trastuzumab plus capecitabine) and multiple duplicate comparisons (3 trials that provided information to compare lapatinib plus capecitabine versus trastuzumab plus capecitabine). The node-splitting analysis revealed no significant difference between the direct and indirect evidence for the comparisons within the closed loop containing capecitabine, lapatinib plus capecitabine, and trastuzumab plus capecitabine (P > 0.05). Regarding the 3 trials informing the comparison of lapatinib plus capecitabine and trastuzumab plus capecitabine, the node-splitting analysis did not reveal significant differences between treatments. However, it should be noted that the direct evidence from 1 trial (Takano et al.⁴⁰) suggested improved efficacy for lapatinib plus capecitabine versus trastuzumab plus capecitabine, whereas the other 2 trials (LANTERN³⁵ and Pivot et al.¹⁹) suggested the opposite. Also, the Higgins I² value was 20.8% and Cochran’s Q value was 1.879 (P = 0.391), suggesting low heterogeneity between the 3 trials.¹¹

Bayesian fixed-effects and random-effects models were conducted and compared. The fixed-effects models indicated convergence while the random-effects models suggested the models had difficulty converging. In addition, the random-effects models showed inconsistency with direct trial data in terms of significant treatment comparisons. The DIC value for the fixed-effects model was lower than the random-effects model (−0.8 versus −0.3, respectively); the difference of 5 points was considered meaningful. Therefore, the fixed-effects Bayesian models were used as the key results of the sponsor’s NMA.¹¹

Figure 15: Network of Evidence for the Primary PFS Hazard Ratio Analysis

PFS = progression-free survival; Rx = treatment; T-DM1 = trastuzumab emtansine.

Note: Blue ovals indicate treatment of interest; grey ovals indicate common comparator.

Source: Sponsor’s indirect treatment comparison.¹¹

Overall Survival

The network of evidence for OS is depicted in Figure 16. As there were no closed loops, limited information was available for assessment of heterogeneity. Multiple trials (Takano et al.,⁴⁰ LANTERN,³⁵ Pivot et al.¹⁹) were available to inform the HRs for comparisons of lapatinib plus capecitabine and trastuzumab plus capecitabine; there were no indications of significant heterogeneity between treatments. As previously mentioned, direct evidence from the trial by Takano et al.⁴⁰ suggested improved efficacy of lapatinib plus capecitabine versus trastuzumab plus capecitabine, whereas the other 2 trials (LANTERN³⁵ and Pivot et al.¹⁹) suggested the opposite. The I² value for this comparison was 38.8% with a Cochrane’s Q value of 2.759 (P = 0.252) suggesting a moderate level of heterogeneity between the Takeno et al.,⁴⁰ LANTERN,³⁵ and Pivot et al.¹⁹ trials.

The fixed-effects Bayesian models indicated convergence, and the results from the fixed-effects model were consistent with the direct trial treatment comparisons. The random-effects models suggested that models had difficulty converging and showed inconsistency with direct trial data in terms of significant treatment comparisons. The DIC value for the fixed-effects model was slightly lower than the random-effects model (4.2 versus 4.5, respectively). Therefore, the fixed-effects Bayesian models were used as the key results of the sponsor’s NMA.

Figure 16: Network of Evidence for the Primary OS Hazard Ratio Analysis

OS = overall survival; Rx = treatment; T-DM1 = trastuzumab emtansine.

Note: Blue ovals indicate treatment of interest; grey ovals indicate common comparator.

Source: Sponsor’s indirect treatment comparison.¹¹

Objective Response Rate

The network of evidence for the analysis of best tumour response is illustrated in Figure 17. There was limited information on heterogeneity in the single closed loop in the network, which consisted of the following treatments: capecitabine, lapatinib plus capecitabine, and trastuzumab plus capecitabine. The node-splitting analysis of inconsistency did not reveal any significant differences (P > 0.05) between the direct or indirect evidence used for the comparisons in the closed loop; therefore, all trials were included in the analysis for best tumour response. There was convergence of Bayesian fixed-effects models, and the results from the fixed-effects models were consistent with the direct evidence from the trials in terms of significant differences in ORR. The Bayesian random-effects models suggested difficulty with convergence, and there was inconsistency between the random-effects models and the direct trial data. The DIC value was slightly higher for the fixed-effects model (DIC = 241.3) compared with the random-effects model (DIC = 241.1), However, due to the inconsistencies of the random-effects model, the fixed-effects Bayesian model was chosen as the primary analysis for ORR.¹¹

Figure 17: Network of Evidence for the Best Tumour-Response Analysis

Rx = treatment; T-DM1 = trastuzumab emtansine.

Note: Blue ovals indicate treatment of interest; grey ovals indicate common comparator

Source: Sponsor’s indirect treatment comparison.¹¹

Results

Progression-Free Survival

Primary HR Analysis: The results for the pairwise comparisons of the HR analysis are shown in Figure 18. Results of the pairwise comparisons suggested that the tucatinib-combination treatment was favoured compared with capecitabine monotherapy (HR = 0.33; 95% CrI, 0.23 to 0.47; P < 0.0001), neratinib (HR = 0.47; 95% CrI, 0.30 to 0.71; P = 0.0007), lapatinib plus capecitabine (HR = 0.55; 95% CrI, 0.40 to 0.76; P = 0.0003), trastuzumab plus capecitabine (HR = 0.53; 95% CrI, 0.42 to 0.68; P < 0.0001), and pertuzumab plus trastuzumab plus capecitabine (HR = 0.65; 95% CrI, 0.47 to 0.90; P = 0.0110). No difference was shown between the tucatinib combination and margetuximab plus capecitabine, neratinib plus capecitabine, T-DM1, and T-DM1 plus capecitabine.¹¹

Both the Bayesian and frequentist fixed-effects models were consistent in terms of the results of the treatment comparisons (Figure 19); comparisons of treatments were made against lapatinib plus capecitabine. Both analyses suggested that lapatinib plus capecitabine was favoured over the tucatinib-combination treatment, T-DM1 plus capecitabine, T-DM1, neratinib plus capecitabine.¹¹

Sensitivity Analyses That Included Pyrotinib: When including information from trials that assessed pyrotinib plus capecitabine, the results of the fixed-effects Bayesian models were consistent with the results of the primary analysis of PFS, except that pyrotinib plus capecitabine was favoured compared with the tucatinib-combination treatment (HR = 1.75; 95% CrI, 1.09 to 3; P = 0.0177).¹¹

Figure 18: Pairwise Treatment Comparisons for the Primary PFS Hazard Ratio Analysis

NMA = network meta-analysis; PFS = progression-free survival; T-DM1 = trastuzumab emtansine.

Source: Sponsor’s indirect treatment comparison.¹¹

Figure 19: Bayesian Versus Frequentist Forest Plots for the Primary PFS Hazard Ratio Analysis

NMA = network meta-analysis; PFS = progression-free survival; T-DM1 = trastuzumab emtansine.

Source: Sponsor’s indirect treatment comparison.¹¹

Overall Survival

The results for the pairwise comparisons of the HR analysis for OS are reported in Figure 20. The results of the pairwise comparisons suggested that the tucatinib-combination treatment was favoured compared with capecitabine monotherapy (HR = 0.45; 95% CrI, 0.27 to 0.77; P < 0.0017), neratinib (HR = 0.47; 95% CrI, 0.27 to 0.80; P = 0.0073), lapatinib plus capecitabine (HR = 0.59; 95% CrI, 0.41 to 0.83; P = 0.0030), and trastuzumab plus capecitabine (HR = 0.66; 95% CrI, 0.50 to 0.88; P = 0.0040). No difference was shown between the tucatinib combination and margetuximab plus capecitabine, neratinib plus capecitabine, pertuzumab plus trastuzumab plus capecitabine, and T-DM1.¹¹

Both the Bayesian and frequentist fixed-effects models were consistent in terms of the results of the treatment comparisons (Figure 21); comparisons of treatments were made against lapatinib plus capecitabine. Both analyses suggested that lapatinib plus capecitabine was favoured over tucatinib-combination treatment, pertuzumab plus trastuzumab plus capecitabine, and T-DM1.¹¹

Sensitivity Analysis Excluding Takano et al.⁴⁰

After excluding the trial by Takano et al.,⁴⁰ the results of the fixed-effects Bayesian models were consistent with the results of the primary analysis of OS, which favoured tucatinib-combination therapy over lapatinib plus capecitabine, capecitabine, neratinib, neratinib plus capecitabine, and trastuzumab plus capecitabine.¹¹

Figure 20: Pairwise Treatment Comparisons for the Primary OS Hazard Ratio Analysis

NMA = network meta-analysis; OS = overall survival; T-DM1 = trastuzumab emtansine.

Source: Sponsor’s indirect treatment comparison.¹¹

Figure 21: Bayesian Versus Frequentist Forest Plots for the Primary OS Hazard Ratio Analysis

NMA = network meta-analysis; OS = overall survival; T-DM1 = trastuzumab emtansine.

Source: Sponsor’s indirect treatment comparison.¹¹

Objective Response Rate

The probit differences for the pairwise comparisons of ORR are depicted in Figure 22. Tucatinib-combination therapy was favoured over capecitabine (HR = 0.90; 95% CrI, 0.48 to 1.31; P < 0.0001), neratinib (HR = 0.82; 95% CrI, 0.29 to 1.33; P = 0.0010), and trastuzumab plus capecitabine (HR = 0.39; 95% CrI, 0.18 to 0.60; P = 0.003). There were no differences between tucatinib-combination therapy and lapatinib plus capecitabine, neratinib plus capecitabine, pertuzumab plus trastuzumab plus capecitabine, T-DM1, and T-DM1 plus capecitabine.¹¹

Both the Bayesian and frequentist fixed-effects models were consistent with the results of the treatment comparisons (Figure 24); comparisons of treatments were made against lapatinib plus capecitabine. Both analyses suggested that lapatinib plus capecitabine was favoured over capecitabine or neratinib, but that T-DM1 and T-DM1 plus capecitabine were favoured over lapatinib plus capecitabine.¹¹

Figure 22: Pairwise Treatment Comparisons for ORR

NMA = network meta-analysis; ORR = objective response rate; T-DM1 = trastuzumab emtansine.

Note: Probit differences for all pairwise comparisons: Bayesian NMA.

Source: Sponsor’s indirect treatment comparison.¹¹

Figure 23: Predicted Response Rates by Category for the Best Tumour Response

NMA = network meta-analysis; T-DM1 = trastuzumab emtansine.

Source: Sponsor’s indirect treatment comparison.¹¹

Figure 24: Bayesian Versus Frequentist Forest Plots for the Best Tumour-Response Analysis

NMA = network meta-analysis; T-DM1 = trastuzumab emtansine.

Source: Sponsor’s indirect treatment comparison.¹¹

Critical Appraisal of the ITC

The ITC included both randomized phase II and III trials. Phase II trials are typically not powered for hypothesis testing, for example, 2 trials included only 14 and 16 patients, respectively. Small sample sizes may affect the precision of treatment effect estimates. The inclusion of phase II trials may have added uncertainty in the comparison of treatment effect despite the fact these trials would have broadened the evidence base and perhaps strengthened the network of trials included in this ITC.

Treatment crossover was also reported for some trials, which may have biased the efficacy analyses of these trials. Treatment crossover, if due to lack of efficacy of the assigned treatment or disease progression, would most likely have created bias against investigational treatments, particularly in terms of OS; however, it may not bias the estimates of PFS. The patient population in the trials varied in terms of line of therapy. Patients receiving treatment in later lines of therapy were likely to have worse prognoses or to be more difficult to treat compared with patients in earlier lines of therapy (i.e., first or second line). The heterogeneity of the study patient populations resulted in significantly incomparable proportions of patients in terms of ECOG PS status (1 or greater), brain metastasis, and varied degrees of hormone receptor status (estrogen receptor–positive and/or progesterone receptor–positive). These all increased the difficulty of assessing the validity of ITC results.

The sponsor indicated there may be effect modifiers that could affect the ITC’s efficacy comparisons, including prior exposure to treatments, line of therapy, and presence of brain metastases. For all 3 potential effect modifiers, the sponsor considered it unnecessary to perform a meta-regression due to the lack of available data to provide reliable estimates. The lack of adjustment for effect modifiers was considered to introduce bias into the NMA analyses, as there was heterogeneity in the patient populations that was not accounted for and resulted in patient populations that were not comparable. However, due to the minimal amount of data available for exploration of bias, any additional analyses might also have raised other concerns about statistical integrity.

The trials included in the ITC were published between 2008 to 2020. The clinical experts consulting with CADTH for this review highlighted the complex and diverse treatment paradigms for patients with metastatic HER2-positive breast cancer who have received prior treatments. In addition, they acknowledged that new therapies have been introduced that have altered the treatment pathways for patients and, ultimately, the outcomes of patients who were advancing through the health care systems. It is unclear how these evolving differences over the past decades may have affected treatment comparisons in the network.

The studies included in the network involved mostly White patients. Three studies enrolled only Chinese^34,38 or Japanese⁴⁰ patients. These 3 studies were not included in the main analysis for the ITC; however, sensitivity analyses that included these 3 trials showed the results were consistent with the main analyses for PFS and OS.

The risk-of-bias assessment conducted by the sponsor classified 3 trials^19,38,44 as having a medium level of bias, while the trials themselves did not reveal any significant sources of bias. However, these trials relied on the use of subgroup-level data only, which could have resulted in an imbalance in patient characteristics between treatment groups. For the PFS and ORR end points, evaluations by independent review committees or by central assessment were preferred, particularly when all of the included trials except for 1 had an open-label design. However, 4 studies assessed ORR via individual investigator assessment and 7 assessed for PFS. This may have biased the assessment in favour of the study treatment in these trials.

The sponsor conducted an assessment of the proportional hazard assumption in each trial. For PFS, 3 studies^16,40,42 revealed a violation of the proportional hazard assumption, and 1 trial⁴² revealed a violation of the proportional hazard assumption in the analysis of OS. Regarding OS, only 1 trial demonstrated a violation of the proportional hazard assumption. The proportional hazard assumption is important for Cox proportional hazards models, which were used in the analyses of PFS and OS across the trials. The violation of the proportional hazard assumption is expected to impact the analyses of PFS more so than the analyses of OS, as there were greater examples of trials with a violation of the proportional hazard assumption among the trials included in the network of PFS. The sponsor conducted NMAs based on HRs and based on the fractional polynomial method and compared results for consistency.

For the NMAs, the sponsor explored the use of both fixed-effects and random-effects models. It was acknowledged that a fixed-effects model would likely underestimate the error, while a random-effects model might overestimate the error, and that the true values may likely be somewhere in between the results of these 2 models. In fact, more significant differences were found for the fixed-effects models than for the random-effects models, mainly due to insufficient data in the network. For PFS, OS, and ORR, the fixed-effects models exhibited convergence; meanwhile, the random-effects models showed examples of peaks in the iteration plots that suggested difficulties with convergence. In addition, the DIC values were lower (i.e., better) for the fixed-effects models for PFS and OS compared with the random-effects models for the same end points. For ORR, the DIC value was nearly the same for the fixed-effects model compared with the random-effects model (241.3 versus 241.1, respectively); however, due to the inconsistencies of the random-effects model, the fixed-effects model was considered appropriate for analyses of ORR. Based on the CADTH review, the choice to use fixed-effects over random-effects models was considered appropriate, given the improved model convergence and consistency versus the random-effects models.

Patients in most trials were required to have previous treatment with trastuzumab; thus, most trials reported nearly 100% of patients with exposure to trastuzumab. This aligned with the funding request for this review of tucatinib, as patients were also required to have previous exposure to trastuzumab. Overall, the sponsor’s ITC included most of the relevant treatments available in Canadian clinical practice. However, other treatments that could be administered to patients were noted to be missing from the ITC, including regimens that include chemotherapies other than capecitabine (i.e., vinorelbine, carboplatin, paclitaxel) as well as endocrine therapies (i.e., fulvestrant, tamoxifen, aromatase inhibitors). While the frequency of treatments involving only chemotherapy or endocrine therapy is expected to be less than other regimens, they remain options for patients.

Additional end points such as safety or quality of life were not considered in the sponsor’s ITC. It is expected the lack of comparisons for HRQoL end points was due mainly to the lack of available data, as some trials included in the analysis were published when collecting such information was not standard. However, without comparisons of safety data, it is not possible to confidently estimate how the harms of each treatment that would impact patients’ outcomes and choice of therapy.

Summary

One ITC submitted by the sponsor was summarized and appraised. The results of this ITC indicated that treatment with tucatinib plus trastuzumab and capecitabine was favoured when compared with capecitabine, neratinib, lapatinib plus capecitabine, trastuzumab plus capecitabine, and pertuzumab plus trastuzumab plus capecitabine when considering PFS. When considering OS, tucatinib-combination therapy was favoured over capecitabine monotherapy, neratinib, lapatinib plus capecitabine, and trastuzumab plus capecitabine. Regarding ORR, tucatinib-combination therapy was favoured over capecitabine, neratinib, and trastuzumab plus capecitabine. Due to the sparse network, considerable heterogeneity, and the limitations of the methods for analyses, the true magnitude of relative treatment effects is not certain. However, based on the opinions of the clinical experts consulting with CADTH, the overall directions of effects may be reliable. No analyses were conducted that compared the harms and HRQoL for patients treated with tucatinib-combination therapy versus the relevant comparator treatments.

Other Relevant Evidence

This section includes submitted long-term extension studies and additional relevant studies included in the sponsor’s submission to CADTH that were considered to address important gaps in the evidence included in the systematic review.

Other Studies: Efficacy in Patients With Brain Metastases, Study by Lin et al.

The publication by Lin et al.¹² reported exploratory analyses of intracranial efficacy and survival in patients with brain metastases derived from the pivotal HER2CLIMB study. This study has been summarized to provide additional evidence on the impact of tucatinib on intracranial efficacy and survival in patients with brain metastases who were randomized in the HER2CLIMB trial.

Methods

The design of the HER2CLIMB trial has been described earlier in this report. Randomization was stratified by the presence or absence of brain metastases, in addition to ECOG PS score and geographic region. Brain metastases were identified at baseline using MRI and were classified as follows:

• treated and stable (prior local treatment and no evidence of progression at baseline brain MRI, including patients treated during the screening period)

• treated and progressing (prior local treatment but evidence of progression of existing lesions, new lesions, or untreated lesions remaining after prior treatment at baseline brain MRI)

• untreated (no prior local treatment).

Treatment with dexamethasone (up to 2 mg per day) was permitted for patients with brain metastases to control symptoms of brain metastases.

Populations

Patients with untreated brain lesions greater than 2 cm were able to enrol in the HER2CLIMB trial if immediate local therapy was not required, although these patients could enrol after receiving radiation therapy or surgery and after a washout period. These patients were subsequently classified as treated and stable.

Patients were excluded if they had leptomeningeal disease.

A summary of baseline characteristics for patients with brain metastases included in the Lin et al. study is provided in Table 32. The majority of patients were older than 65 years (83.5%) and female (99.3%); 60.8% resided in North America and 93.9% had non-CNS metastatic disease. Regarding ECOG PS, 44.7% of patients had a score of 0 and 55.3% had a score of 1, while 57.0% of patients were hormone receptor–positive. The brain metastasis treatment status at baseline was treated and stable, treated and progressing, or untreated for 40.2%, 37.1%, and 22.7% of patients, respectively. Most patients (70.1%) had prior radiation therapy for brain metastases, 41.9% had whole-brain radiation therapy, 42.6% had targeted radiation therapy, and 15.8% had surgery. In addition, the median number of months from the diagnosis of metastatic disease to the development of brain metastases was 13.0 (range, < 0.1 to 100.7) for the tucatinib-combination treatment group and 9.8 (range, < 0.1 to 172.7) for the placebo-combination treatment group.

The baseline characteristics of the treatment groups were well balanced with the exception of the proportion of patients who were hormone receptor–positive (54.0% in the tucatinib-combination group versus 63.4% in the placebo-combination group), patients with an ECOG PS score of 1 (53.5% in the tucatinib-combination group versus 59.1% in the placebo-combination group), history of prior targeted radiation therapy (46.5% in the tucatinib-combination group versus 34.4% in the placebo-combination group).

Interventions

The interventions for the HER2CLIMB study have been described previously. Briefly, patients were randomized to receive tucatinib or placebo. Both were used in combination with trastuzumab and capecitabine.

Outcomes

Disease response and progression in the brain were evaluated using the RECIST 1.1 guidelines and assessed by the study investigator. Two exploratory end points, confirmed intracranial ORR (ORR-IC) and duration of intracranial response (DOR-IC), were reported in patients with measurable intracranial lesions at baseline and CNS-PFS. CNS-PFS was defined as time from randomization to disease progression in the brain or death resulting from any cause, whichever occurred first. DOR-IC was defined as the time from the first intracranial objective response (confirmed complete or partial) to documented intracranial disease progression or death resulting from any cause, whichever occurred first. CNS-PFS and OS were also reported.

Statistical Analysis

CNS-PFS, OS, and time to second progression were estimated using Kaplan–Meier curves and 95% CIs. A stratified Cox proportional hazards model was used to estimate HRs and 95% CIs for CNS-PFS and OS. An unstratified Cox proportional hazards model was used for the time to second progression.

All reported P values were nominal and obtained from a stratified log-rank test.

A between-group comparison of ORR-IC was analyzed using a 2-sided Cochran-Mantel-Haenszel test controlled for the stratification factors (presence or absence of brain metastases, in addition to ECOG PS score and geographic region). Kaplan–Meier estimates of median DOR-IC were also provided using methods similar to the analysis of PFS used for the DOR-IC analysis.

Patient Disposition

Patient disposition for the HER2CLIMB trial is summarized in Figure 25. A total of 612 patients were enrolled in the HER2CLIMB trial, 291 (48%) of whom had brain metastases at baseline or a history of brain metastases. The 281 patients with brain metastases were included in the exploratory analyses reported by Lin et al. A total of 198 patients randomized to the tucatinib combination and 93 patients randomized to the placebo combination had brain metastases. No additional information was provided for patient disposition.

Figure 25: Patient Disposition

BMs = brain metastases.

Source: Lin et al. (2020).¹²

Exposure to Study Treatments

Exposure to treatment was not reported.

Efficacy

CNS-PFS and OS

A summary of results for 1-year CNS-PFS and OS in patients with brain metastases is provided in Table 33. Kaplan–Meier curves for all patients with brain metastases (Figure 26), patients with active brain metastases (Figure 27), and patients with stable brain metastases (Figure 28) have also been provided. For patients treated with the tucatinib combination, 40.2% (95% CI, 29.5 to 50.6) of patients with brain metastases, 35.0% (95% CI, 23.2 to 47.0) of patients with active brain metastases, and 53.3% (95% CI, 31.4 to 71.0) of patients with stable brain metastases had CNS-PFS at 1 year. None of the patients receiving the placebo combination had CNS-PFS at 1 year. An HR of 0.32 (95% CI, 0.22 to 0.48) was reported for the tucatinib combination compared with the placebo combination in all patients with brain metastases. Similar results were reported for patients with active brain metastases (HR = 0.36; 95% CI, 0.22 to 0.57) and patients with stable brain metastases (HR = 0.31; 95% CI, 0.14 to 0.67).

Among all patients with brain metastases, 1-year OS was reported for 70.1% (95% CI, 62.1 to 76.7) of patients in the tucatinib-combination treatment group and 46.7% (95% CI, 33.9 to 58.4) of patients in the placebo-combination treatment group. For patients with active brain metastases, 1-year OS was reported for 71.7% (95% CI, 61.4 to 79.7) and 41.1% (95% CI, 25.5 to 56.1) of patients randomized to the tucatinib-combination and placebo-combination groups, respectively. For patients with stable brain metastases, 1-year OS was reported for 67.6% (95% CI, 53.8 to 78.0) and 55.6% (95% CI, 34.1 to 72.6) of patients randomized to tucatinib-combination and placebo-combination treatment, respectively. This data for 1-year OS corresponded to an HR of 0.58 (95% CI, 0.40 to 0.85) for all patients with brain metastases, 0.49 (95% CI, 0.30 to 0.80) for patients with active brain metastases, and 0.88 (95% CI, 0.45 to 1.70) for patients with stable brain metastases.