CADTH Reimbursement Review

Alpelisib (Piqray)

Sponsor: Novartis Pharmaceuticals 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.

Table 1: Submitted for Review

CDK4/6 = cyclin-dependent kinase 4 and 6; HER2 = human epidermal growth factor receptor 2; NOC = Notice of Compliance.

Introduction

Breast cancer is the most commonly diagnosed cancer in Canadian females and the most common subtype of breast cancer is hormone receptor–positive, human epidermal growth factor receptor 2 (HER2)–negative breast cancer. It was estimated that in 2020, there would be 27,200 new cases of breast cancer in Canada, with age-standardized incidence rates of 1.1 per 100,000 males and 128.2 per 100,000 females. It was also estimated that there would be 5,100 deaths from breast cancer, with age-standardized mortality rates of 0.3 per 100,000 males and 22.0 per 100,000 females. In an analysis of females diagnosed with breast cancer from 2012 to 2016 in the Ontario Cancer Registry, the percentages of patients with 5-year survival ranged from 24.0% to 94.7%, depending on disease stage at diagnosis.

There are no curative treatments for advanced or metastatic breast cancer. According to input from clinicians consulted by CADTH for the purpose of this review, standard first-line therapy for advanced or metastatic hormone receptor–positive HER2-negative breast cancer is a cyclin-dependent kinase 4 and 6 (CDK 4/6) inhibitor in combination with an aromatase inhibitor (AI). A luteinizing hormone–releasing hormone (LHRH) receptor agonist is also given for ovarian suppression, depending on menopausal status. Patients whose disease recurs while on or shortly after stopping adjuvant AI therapy are considered resistant and are frequently offered CDK4/6 inhibitor with fulvestrant instead of an AI. Upon disease progression, second-line therapy options include a different single-agent AI, a single-agent fulvestrant, an investigational therapy in a clinical trial, chemotherapy (commonly single-agent capecitabine or a taxane), combined everolimus and exemestane, and tamoxifen. Despite the number of options beyond first-line therapy, there is no high-quality evidence for prolongation of survival with these therapies. Coverage of fulvestrant and everolimus plus exemestane is inconsistent across jurisdictions in Canada.

Alpelisib is an oral phosphatidylinositol 3-kinase (PI3K) inhibitor that, in combination with fulvestrant, is indicated for the treatment of postmenopausal women, and men, with hormone receptor–positive, HER2-negative, PIK3CA-mutated advanced or metastatic breast cancer after disease progression following an endocrine-based regimen. Alpelisib 300 mg is taken daily on a continuous basis along with intramuscular (IM) fulvestrant 500 mg on day 1, day 15, and day 29, and every 28 days thereafter. Dose reductions for alpelisib are permitted to 250 mg or 200 mg daily for the management of adverse drug reactions.

Alpelisib has not been previously reviewed by CADTH. The sponsor’s reimbursement request is for alpelisib in combination with fulvestrant for the treatment of postmenopausal women, and men, with hormone receptor–positive, HER2-negative, PIK3CA-mutated advanced or metastatic breast cancer after disease progression following an endocrine-based regimen with a CDK4/6 inhibitor. The reimbursement request differs from the Health Canada indication in that it specifies that patients must have received CDK4/6 inhibitor with a previous endocrine-based regimen.

The objective of this report is to perform a systematic review of the beneficial and harmful effects of 50 mg, 150 mg, and 200 mg alpelisib tablets in combination with fulvestrant for the treatment of postmenopausal women, and men, with hormone receptor–positive, HER2-negative, PIK3CA-mutated advanced or metastatic breast cancer after disease progression following an endocrine-based regimen with a CDK4/6 inhibitor.

Stakeholder Perspectives

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

Patient Input

Three patient groups provided input for this review: the Canadian Breast Cancer Network (CBCN), Rethink Breast Cancer, and CanCertainty. The CBCN provided information gathered from patient and caregiver responses from 2 online surveys conducted in 2017 (90 patients) and 2012 (87 patients and caregivers) and a telephone interview with 1 patient. Rethink Breast Cancer’s submission was informed by an online survey in 2021 of 24 patients and telephone interviews with 6 of the survey respondents. CanCertainty developed its submission based on published reports relating to breast cancer and oral cancer drugs.

The CBCN reported that the physical impact of metastatic breast cancer is variable across individuals, with the vast majority of patients in the 2012 survey reporting a significant/debilitating or some/moderate impact on their quality of life due to the symptoms of fatigue, insomnia, and pain. Many negative impacts on patients and their families’ daily lives were identified, including restrictions in patients’ ability to remain employed, care for children and dependents, be social, exercise, pursue hobbies and interests, and spend time with loved ones. The patient groups identified the following measures of effectiveness as the most important: progression-free survival (PFS), overall survival (OS), quality of life, and adverse effects. Survey results indicated that patients are willing to tolerate side effects for drugs that can improve long-term health outcomes.

Clinician Input

Input From Clinical Experts Consulted by CADTH

The following input was provided by 2 clinical specialists with expertise in the diagnosis and management of breast cancer.

Current therapies for advanced or metastatic breast cancer beyond the first-line setting have low response rates and have not been shown to improve OS. Chemotherapy options are more poorly tolerated than endocrine therapy and many available chemotherapy options are administered intravenously, requiring more hospital visits and reliance on institutions. Alpelisib would be the first treatment available specifically for patients with PIK3CA-mutated cancer.

For patients harbouring a PIK3CA mutation, alpelisib would be added to an already established standard of care (SOC) option for the second-line treatment of advanced or metastatic hormone receptor–positive HER2-negative breast cancer (i.e., fulvestrant). Alpelisib would not be used as a first-line treatment given the strong evidence for the use of CDK4/6 inhibitors with endocrine therapy in that setting. Patients with advanced or metastatic hormone receptor–positive HER2-negative breast cancer, activating mutations in the PIK3CA gene (identified using liquid biopsy or tissue testing on archival or newly obtained tumour tissue), good performance status, expected survival of longer than 3 months, and no type 1 diabetes mellitus or uncontrolled type 2 diabetes mellitus would be best suited for treatment with alpelisib plus fulvestrant. Alpelisib with fulvestrant would not be reserved for patients who are intolerant of other treatments or for whom other treatments are contraindicated. In patients with life-threatening visceral organ metastases, chemotherapy would be recommended before considering treatment with alpelisib and fulvestrant. Patients would not be suited for treatment with alpelisib plus fulvestrant if they have poor performance status, have type 1 or uncontrolled type 2 diabetes mellitus, are unable to understand and manage potential toxicities and dosing and monitoring requirements, or are non-compliant with follow-up.

Treatment response is monitored using a combination of clinical examination, laboratory evaluation (markers of organ function with or without tumour markers), and radiographic evaluation. Treatment continues as long as the disease is stable or responding on radiographic scans according to the Response Evaluation Criteria in Solid Tumours (RECIST) criteria. Treatment with alpelisib and fulvestrant should be discontinued if there is disease progression, intolerable or dangerous toxicity (especially uncontrollable hyperglycemia), or an event or development of a comorbidity that adversely impacts performance status or survival (e.g., stroke).

Treatment with alpelisib and fulvestrant would be prescribed by medical oncologists or associated team physicians with expertise in cancer therapies and toxicity management. Patients would be treated on an outpatient basis under medical oncology supervision and fulvestrant injections would be administered in a hospital outpatient clinic or family doctor’s office.

Clinician Group Input

One clinician group submission was received from 6 clinicians with the Breast Medical Oncology group at the Ottawa Hospital Cancer Centre. Due to the small percentage of patients in the pivotal trial who had previously received the current first-line SOC with CDK4/6 inhibitors, opinions were divided on whether it would be appropriate to offer alpelisib to this patient population. While there was mention that patients had been treated recently through the sponsor’s access program, the submission did not describe the clinicians’ experience with alpelisib.

Drug Program Input

There were several questions from officials with the drug plans regarding patient populations that would be suitable for treatment with alpelisib plus fulvestrant, discontinuation of alpelisib or fulvestrant, and PIK3CA mutation testing. Patients were excluded from the pivotal trial for alpelisib if they had an Eastern Cooperative Oncology Group Performance Status (ECOG PS) score of 2 or more, were receiving LHRH agonist for the induction of ovarian suppression, had inflammatory breast cancer, had symptomatic visceral disease, had received prior chemotherapy in the metastatic setting, had received prior fulvestrant treatment, had uncontrolled central nervous system (CNS) metastases, or had type 1 diabetes or uncontrolled type 2 diabetes. According to the clinical experts consulted by CADTH, patients receiving LHRH agonist for the induction of ovarian suppression would be eligible, while patients in the other groups (aside from those with diabetes) could be considered for eligibility on a case-by-case basis or if they met certain other criteria.

The drug plan representatives also wanted to know if alpelisib could be continued as monotherapy if fulvestrant was discontinued or interrupted. The clinical experts indicated that alpelisib could be continued during an interruption but not after discontinuation. Conversely, the drug plans also wanted to know if patients who had to discontinue alpelisib due to intolerance could continue with single-agent fulvestrant. The clinical experts considered it appropriate to continue these patients on single-agent fulvestrant. In response to a related question, the experts also considered it appropriate to permanently discontinue alpelisib after it had been discontinued for more than 4 weeks due to unresolved toxicity. Another drug plan question was whether it would be appropriate to offer patients on chemotherapy with no evidence of progressive disease or intolerance alpelisib plus fulvestrant. The clinical experts did not consider this appropriate as patients doing well on chemotherapy would not be switched to a different therapy.

With regard to PIK3CA mutation testing, representatives for the drug plans asked which patients should be tested for the PIK3CA mutation and when in the course of treatment this testing should occur. According to the clinical experts consulted by CADTH, patients identified as best suited for alpelisib plus fulvestrant treatment minus the criterion of having an activating mutation in the PIK3CA gene should be tested. (Refer to the preceding section regarding input from clinical experts consulted by CADTH for patients identified as best suited for alpelisib plus fulvestrant treatment.) Testing should be performed at diagnosis of de novo metastatic breast cancer, relapse following treatment for early breast cancer, or progression on first-line therapy for advanced or metastatic breast cancer.

Clinical Evidence

Pivotal Studies and Protocol Selected Studies

Description of Studies

The CADTH systematic review identified 1 relevant study, the SOLAR-1 study. The SOLAR-1 study (N = 572) was a placebo-controlled, double-blind, parallel-group randomized controlled trial (RCT) that randomized patients 1:1 to alpelisib 300 mg daily or matching-administration placebo, in combination with fulvestrant 500 mg on day 1, day 15, and day 29, and every 28 days afterwards. Men and postmenopausal women with hormone receptor–positive, HER2-negative advanced or metastatic breast cancer and previous endocrine therapy were randomized within each of 2 cohorts based on PIK3CA mutation status: the PIK3CA mutant cohort and PIK3CA non-mutant cohort. The primary and key secondary outcomes were PFS and OS in the PIK3CA mutant cohort (N = 341). Endocrine therapy with a CDK4/6 inhibitor was not a part of the SOC at the time the study was conducted (enrolment was from 2015 to 2017) and only 20 patients in the PIK3CA mutant cohort had received prior CDK4/6 inhibitor treatment, meeting the reimbursement criteria requested by the sponsor.

Within the PIK3CA mutant cohort, 20 patients were identified as having prior CDK4/6 inhibitor treatment, according to the randomization stratum. Female patients were included only if they were postmenopausal and were not receiving LHRH agonist for the induction of ovarian suppression. In the subgroup with prior CDK4/6 inhibitor treatment, all patients had an ECOG PS of 0 or 1, most patients were White, and most had secondary endocrine resistance. In the entire cohort, most patients were White, had an ECOG PS of 0 (the remaining having a performance status of 1), had 1 or 2 metastatic sites, had 1 line of prior medication therapy, had no prior hormonal therapy in the metastatic setting, and had secondary endocrine resistance.

Efficacy Results

Efficacy results from the SOLAR-1 study PIK3CA mutant cohort in the subgroup of patients with prior CDK4/6 inhibitor treatment are presented in Table 2. At the final PFS analysis at the June 12, 2018, data cut-off date, median PFS was 5.5 months (95% confidence interval [CI], 1.58 months to 16.76 months) in the alpelisib group and 1.8 months (95% CI, 1.68 months to 3.58 months) in the placebo group. The hazard ratio for the alpelisib group versus the placebo group was 0.48 (95% CI, 0.17 to 1.36).

At the final OS analysis at the April 23, 2020, data cut-off date, median OS was 29.8 months (95% CI, 6.67 months to 38.21 months) in the alpelisib group and 12.9 months (95% CI, 2.46 months to 34.60 months) in the placebo group. The hazard ratio for the alpelisib group versus the placebo group was 0.67 (95% CI, 0.21 to 2.18).

Harms Results

Harms are presented for all patients in the PIK3CA mutant cohort in Table 2. Almost all patients reported at least 1 adverse event (AE) (99.4% in the alpelisib group and 90.6% in the placebo group). Most of the AEs occurring in at least 10% of patients of at least 1 treatment group were more common in the alpelisib group compared with the placebo group. All of the AEs reported by more than 20% of patients in the alpelisib group were also more common in the alpelisib group than in the placebo group: hyperglycemia, diarrhea, nausea, rash, decreased appetite, decreased weight, stomatitis, vomiting, fatigue, and alopecia.

Serious AEs (SAEs) were reported in 39.6% of the alpelisib group and 19.9% of the placebo group. The most common SAEs were hyperglycemia (10.1% in the alpelisib group and none in the placebo group), osteonecrosis of the jaw (3.6% in the alpelisib group and none in the placebo group), stomatitis, acute kidney injury, and rash (2.4% in the alpelisib group and none in the placebo group for the preceding 3 SAEs).

Withdrawals from treatment due to AEs were more common in the alpelisib group (27.2%) versus the placebo group (5.8%). The most common AEs leading to discontinuation were reported in the alpelisib group alone: hyperglycemia (6.5%), rash (4.7%), and diarrhea (3.6%).

On-treatment deaths up to 30 days after the last dose of study treatment occurred in 4.1% of the alpelisib group and 5.8% of the placebo group. The most common cause of on-treatment death was breast cancer (3.6% in the alpelisib group and 4.1% in the placebo group) and other causes of on-treatment death were reported for 1 patient each.

The following notable harms identified in the systematic review protocol occurred in more than 10% of at least 1 treatment group and were more common in the alpelisib group: hyperglycemia, diarrhea, nausea, rash, vomiting, and maculopapular rash.

Critical Appraisal

No relevant conclusions can be drawn regarding PFS and OS in patients treated with alpelisib and fulvestrant versus placebo and fulvestrant because the SOLAR-1 study was not designed to test hypotheses in the subgroup of patients with prior CDK4/6 inhibitor treatment and did not include outcomes in this subgroup in the statistical testing hierarchy. Only the results in this small subgroup can inform comparative efficacy in the patient population targeted by the sponsor’s reimbursement request, since the results in the entire PIK3CA mutant cohort cannot be generalized to the relevant patient population.

Other Relevant Evidence

Description of Studies

There were 2 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. The BYLieve study, a non-comparative cohort study, included 1 cohort of patients treated with alpelisib and fulvestrant that matched the relevant patient population. In a separate observational study, the relevant cohort of the BYLieve study was compared with a database-derived cohort treated with non-alpelisib SOC following propensity score weighting.

Non-Comparative Cohort Study

The BYLieve study assigned patients to 1 of 3 cohorts based on their most recent anticancer therapy. Of the 3 cohorts, cohort A (N = 127) was relevant to the review. Cohort A included patients with hormone receptor–positive, HER2-negative advanced or metastatic breast cancer and a confirmed PIK3CA mutation who had received any CDK4/6 inhibitor plus any AI as their immediate prior treatment. These patients were assigned to receive alpelisib plus fulvestrant at the same dosages as in the SOLAR-1 study. The primary end point in the BYLieve study was the proportion of patients who were alive without disease progression at 6 months by local investigator assessment using the criteria of RECIST Version 1.1 (RECIST 1.1). The outcomes of PFS and OS as well as safety data were also evaluated in the BYLieve study.

Progression and Survival Results

As of the data cut-off date, 61 of 121 (50.4%) patients in the modified full analysis set or mFAS (N = 121) were alive without progressive disease per investigator assessment at 6 months (95% CI, 41.2% to 59.6%). The study met the primary objective for cohort A because the lower bound of the 95% CI was greater than 30%. The median PFS by investigator assessment was 7.3 months (95% CI, 5.6 months to 8.3 months). The PFS rates by investigator assessment at 6 months and 12 months were 54.1% (95% CI, 44.3% to 62.9%) and 27.3% (95% CI, 17.6% to 37.8%), respectively.

The median OS was 17.3 months (95% CI, 17.2 months to 20.7 months). The OS rates at 6 months and 12 months were 91.9% (95% CI, 84.9% to 95.7%) and 75.2% (95% CI, 62.5% to 84.2%), respectively. The sponsor indicated in the BYLieve Clinical Study Report that OS data should be interpreted with caution due to the proportion of patients alive and to ongoing follow-up at the time of the data cut-off date.

Harms Results

Almost all (99.2%) of patients experienced at least 1 treatment-emergent AE. The most common AEs (≥ 20%) were diarrhea (59.8%), hyperglycemia (58.3%), nausea (45.7%), fatigue (29.1%), decreased appetite (28.3%), rash (28.3%), stomatitis (26.8%), and vomiting (23.6%). Overall, 26.0% of patients experienced an SAE. The most common SAEs were hyperglycemia (5.5%), maculopapular rash (3.1%), dyspnea (2.4%), pleural effusion (2.4%), abdominal pain (1.6%), and hematemesis (1.6%). The most common AEs leading to discontinuation of study treatment were rash (3.9%), colitis, hyperglycemia, urticaria, and vomiting (1.6% each). As of the data cut-off date, 7 (5.5%) patients had died during study treatment or within 30 days of the last dose of study drug, and 4 of these on-treatment deaths were attributed to breast cancer.

The following notable harms were reported: hyperglycemia (58.3%), hypersensitivity and anaphylactic reactions (10.2%), diarrhea (59.8%), nausea (45.7%), rash (28.3%), vomiting (23.6%), maculopapular rash (14.2%), pneumonitis (0.8%), and severe cutaneous skin reactions (0.8%).

Critical Appraisal

The BYLieve study is unable to inform the efficacy of alpelisib plus fulvestrant versus a relevant comparator due to its non-comparative study design. There was also no statistical hypothesis testing in the relevant outcomes of interest, PFS and OS.

Observational Study

The observational study compared cohort A from the BYLieve study with a real-world cohort derived from the Flatiron Clinico-Genomic Database. Cohort A from the BYLieve study (N = 120), whose patients received alpelisib plus fulvestrant following treatment with a CDK4/6 inhibitor plus AI, was compared to the Flatiron cohort (N = 95), whose patients received non-alpelisib SOC following treatment with a CDK4/6 inhibitor and non-fulvestrant endocrine therapy. The outcome PFS was compared between the cohorts following weighting of the Flatiron cohort based on propensity scores.

Efficacy Results

Following propensity score weighting to estimate the average treatment effect on those treated, median PFS was 3.7 months (95% CI, 3.1 months to 6.1 months) in the Flatiron cohort and 7.3 months (95% CI, 5.6 months to 8.3 months) in the BYLieve cohort with a P value of 0.040 for the log-rank test. The weighted hazard ratio for PFS in the BYLieve cohort versus the Flatiron cohort was 0.62 (95% CI, 0.44 to 0.85; P = 0.002). The observational study included sensitivity analyses assessing the sensitivity of the results to the form of confounding adjustment — namely, greedy matching and exact matching. The results of those analyses were not meaningfully different from the primary analysis results. No sensitivity analysis was performed on the assumption of no unmeasured confounding.

Harms Results

Harms were not assessed in the observational study.

Critical Appraisal

Overall, there remains a great deal of uncertainty regarding the efficacy of alpelisib in comparison to the SOC due to the inherent limitations of observational data. While the adjustment approaches taken in this study may have adequately balanced on observable prognostic factors categorized as they were, bias in the efficacy estimate due to selection bias, measurement error, unmeasured confounding, and residual confounding cannot be ruled out. No attempts were made to assess or estimate the possible magnitude of such bias.

Conclusions

No conclusions could be drawn from the SOLAR-1 and BYLieve studies regarding the comparative efficacy or effectiveness of alpelisib plus fulvestrant versus any relevant comparator in patients with hormone receptor–positive, HER2-negative, PIK3CA-mutated advanced or metastatic breast cancer after disease progression following an endocrine-based regimen with a CDK4/6 inhibitor. Neither study was designed to draw conclusions on comparative efficacy in this patient population. The sponsor-submitted observational study comparing a cohort from the BYLieve study that received alpelisib plus fulvestrant to a database-derived cohort that received a variety of non-alpelisib SOC therapies reported PFS results in favour of alpelisib. Methodological limitations in the observational study, including a high likelihood of residual confounding that may have led to bias in favour of alpelisib and differences in the methods used to determine PFS, contributed a substantial degree of uncertainty to the effect estimate. Considering the evidence in its entirety, the magnitude of any benefit associated with alpelisib plus fulvestrant in the relevant patient population remains unclear. In the SOLAR-1 and BYLieve studies, alpelisib treatment was associated with hyperglycemia, diarrhea, nausea, rash, decreased appetite, stomatitis, vomiting, and fatigue. Although the AEs reported in the studies can be medically managed, the percentages of patients who discontinued treatment due to AEs in the studies suggest that a large proportion of patients may not be able to remain on treatment with alpelisib due to AEs and/or subsequent side effects from treatments to manage these.

Introduction

Disease Background

Breast cancer is the most commonly diagnosed cancer in Canadian females. According to Canadian data from 2011 to 2015, more than 80% of breast cancers in Canadian females were diagnosed at an early stage (stage I or stage II).³ For the purposes of this review, advanced breast cancer refers to locoregionally advanced breast cancer not amenable to curative therapy. Metastatic breast cancer occurs when cancer spreads to other parts of the body. Most patients with metastatic breast cancer are those with relapse following treatment for early breast cancer as opposed to those with metastatic breast cancer at diagnosis (de novo cases).⁴ Independent of stage, there are 3 main subtypes of breast cancer, based on expression of hormone (estrogen and/or progesterone) receptors and overexpression of HER2. The most common subtype of breast cancer is hormone receptor–positive, HER2-negative breast cancer.

It was estimated that in 2020, there would be 27,200 new cases of breast cancer in Canada, with age-standardized incidence rates of 1.1 per 100,000 males and 128.2 per 100,000 females.⁵ It was also estimated that there would be 5,100 deaths from breast cancer, with age-standardized mortality rates of 0.3 per 100,000 males and 22.0 per 100,000 females.⁵ In an analysis of females diagnosed with breast cancer from 2012 to 2016 in the Ontario Cancer Registry, 64.8% of patients with a known subtype had hormone receptor–positive HER2-negative breast cancer, with an estimated annual incidence rate of 97 to 105 per 100,000 females.⁶ This estimate is in line with an estimate from registries in the US in patients diagnosed from 2010 to 2013 (66.6%).⁷ The percentages of Ontario patients with hormone receptor–positive, HER2-negative breast cancer who had stage III and stage IV disease at diagnosis were 12% and 3.7%, respectively.⁶ Patients with stage IV, hormone receptor–positive, HER2-negative breast cancer at diagnosis in the Ontario study had an estimated median survival of 35.2 months, though percentages of patients with 5-year survival ranged from 24.0% to 94.7%, depending on disease stage at diagnosis.⁶

Standards of Therapy

Although there are no curative treatments for advanced or metastatic breast cancer, there are multiple systemic therapies available for the disease. According to input from clinicians consulted by CADTH for the purpose of this review, standard first-line therapy for advanced or metastatic hormone receptor–positive, HER2-negative breast cancer is a CDK4/6 inhibitor with an AI. An LHRH receptor agonist is also given for ovarian suppression, depending on menopausal status. Exceptions to the use of a CDK4/6 inhibitor might include patients with a very low burden of disease, patients with significant comorbidities or contraindications to CDK4/6 inhibitors, and patients who prefer single-agent endocrine treatment for reasons such as desire for less frequent monitoring or concern about side effects. Patients whose disease recurs while on or shortly after stopping adjuvant AI therapy are considered resistant and are frequently offered CDK4/6 inhibitor with fulvestrant instead of an AI. Upon disease progression, second-line therapy options include a different single-agent AI, a single-agent fulvestrant, an investigational therapy in a clinical trial, chemotherapy (commonly single-agent capecitabine or a taxane), combined everolimus and exemestane, and tamoxifen. Coverage of fulvestrant and everolimus plus exemestane is inconsistent across jurisdictions in Canada. Aggressive disease progression (e.g., disease not responding to first-line therapy or significant visceral metastases) is treated with chemotherapy. Despite the number of options beyond first-line therapy, there is no high-quality evidence for prolongation of survival with these therapies. The SOC outlined by the clinicians consulted by CADTH is consistent with a recent international consensus guideline published by the European School of Oncology (ESO) and the European Society for Medical Oncology (ESMO) for the management of advanced breast cancer.⁸

The clinicians consulted by CADTH for this review described the following treatment goals: improving OS, delaying cancer progression, maintaining or improving quality of life, preventing or improving cancer-related symptoms (e.g., pain, dyspnea, fatigue), maintaining or improving organ function, maintaining patient independence, increasing ability to maintain employment, and delaying the initiation of chemotherapy. Input from patient groups for this review reported similar treatment goals. Patients indicated that they wanted treatments that delayed progression of their disease, prolonged life without sacrificing quality of life, reduced symptoms, and minimized the risk of side effects while stabilizing their disease. Patients were willing to tolerate side effects if treatments improved long-term health outcomes.

Drug

Alpelisib is an oral PI3K inhibitor that, in combination with fulvestrant, is indicated for the treatment of postmenopausal women, and men, with hormone receptor–positive, HER2-negative, PIK3CA-mutated advanced or metastatic breast cancer after disease progression following an endocrine-based regimen. Alpelisib 300 mg is taken daily on a continuous basis along with IM fulvestrant 500 mg on day 1, day 15, and day 29, and every 28 days thereafter. Dose reductions for alpelisib are permitted to 250 mg or 200 mg daily for the management of adverse drug reactions.

The class IA isoforms of PI3K play a role in the control of cell growth, proliferation, metabolism, and migration via the PI3K/protein kinase B/mechanistic target of rapamycin pathway.⁹ There is some evidence to suggest that the presence of mutations in the gene encoding the p110 alpha catalytic subunit of PI3K (PIK3CA) is associated with worse OS and resistance to chemotherapy in patients with metastatic breast cancer.⁹ Alpelisib is the first approved therapy in Canada for patients with PIK3CA-mutated, hormone receptor–positive, HER2-negative advanced or metastatic breast cancer.

Alpelisib has not been previously reviewed by CADTH. The sponsor’s reimbursement request is for alpelisib in combination with fulvestrant for the treatment of postmenopausal women, and men, with hormone receptor–positive, HER2-negative, PIK3CA-mutated advanced or metastatic breast cancer after disease progression following an endocrine-based regimen with a CDK4/6 inhibitor. The reimbursement request differs from the Health Canada indication in that it specifies that patients must have received CDK4/6 inhibitor with a previous endocrine-based regimen.

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 this review: the CBCN, Rethink Breast Cancer, and CanCertainty.

The CBCN is a patient-directed, national health charity that focuses on ensuring the best quality of care for all Canadians affected by breast cancer through the promotion of information, education, and advocacy activities. The CBCN provided information gathered from 2 online surveys conducted in 2017 and 2012. A total of 180 patients living with breast cancer responded to the 2017 survey. The submission only included data from a subset of 90 Canadian respondents with metastatic, hormone receptor–positive, HER2-negative breast cancer. A total of 71 patients and 16 caregivers responded to the 2012 survey. In addition, the CBCN conducted a telephone interview with 1 patient in May 2021.

Rethink Breast Cancer is a registered charity and the organization’s mission is to empower young people worldwide who are concerned about and affected by breast cancer through education, support, and advocacy. Rethink Breast Cancer conducted an online survey from March 31 to April 8, 2021, to inform its submission. A total of 24 postmenopausal women diagnosed with hormone receptor–positive, HER2-negative, advanced or metastatic breast cancer with a PIK3CA mutation completed the survey (4 of the women were from Canada). Six of these patients also participated in telephone interviews.

The CanCertainty Coalition consists of 30 Canadian patient groups, cancer health charities, and caregiver organizations from across the country. It works with oncologists and cancer care professionals to improve the affordability and accessibility of cancer treatment. CanCertainty developed its submission based on published reports relating to breast cancer and oral cancer drugs.

Disease Experience

The CBCN reported that the physical impact of metastatic breast cancer is variable across individuals and has a significant or debilitating impact on patients’ quality of life. In the 2012 survey, the vast majority of patients reported a significant/debilitating or some/moderate impact on their quality of life, due to the symptoms of fatigue, insomnia, and pain. The CBCN also reported that there are multiple social impacts of living with metastatic breast cancer. Metastatic breast cancer restricts an individual’s employment and career, their ability to care for children and dependents, and their ability to be social and meaningfully participate in their community. Most patient respondents also reported restrictions to their ability to exercise, pursue hobbies and interests, and spend time with loved ones. Other experiences identified by patients 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 loss of a parent on children, not knowing what will happen to children, the loss of support of loved ones, as well as marital stress/loss of fidelity and affection from their husband.

Experience With Treatment

Respondents to the Rethink Breast Cancer survey had experience with a variety of treatments, including fulvestrant, letrozole, palbociclib, exemestane, anastrozole, tamoxifen, and capecitabine. All 24 respondents had experience with alpelisib in combination with fulvestrant, and 20 respondents were treated with a CDK4/6 inhibitor before receiving alpelisib. Most respondents had undergone multiple lines of treatment and reported a wide range of outcomes and side effects. The most commonly reported side effects for breast cancer treatments overall were fatigue (100%), diarrhea (83%), loss of appetite (75%), nausea (54%), and headache (46%). Fatigue, diarrhea, and hyperglycemia were identified as the most difficult-to-tolerate side effects of previous treatments. Most respondents (86%) did not report any difficulty accessing treatment. An additional impact, identified in the CBCN submission, is difficulty associated with accessing quality child care during cancer treatment.

A total of 18 respondents to the Rethink Breast Cancer survey matched the requested reimbursement criteria for the present review and the responses from these patients with regard to experience with alpelisib were summarized in the submission. At the time of the survey, 12 of these patients were still receiving alpelisib, 5 stopped receiving it because it did not control their cancer, and 1 stopped receiving it because she could not tolerate the side effects. Compared to other treatments received, most patients indicated that the drug’s side effects were the same (33%) or worse (28%), though some patients (39%) indicated that the side effects were better. Most patients experienced diarrhea (89%), reduced appetite (78%), weight loss (72%), and alopecia (67%) while receiving alpelisib. While not reported as frequently, hyperglycemia was highlighted during patient interviews as being especially hard to manage. Several respondents reported that dose reductions made an important difference in helping them manage the side effects. Some comments from patients regarding the side effects included the following:

“I am tolerating, but it is difficult.”

“Important to find effective ways to manage SE [side effects] right away, especially in the first 4 months when there are so [many] SE that are pretty overwhelming.”

“Piqray worked for almost 18 months and was tough but manageable. I did not have any of the major side effects like blood sugar issues or the rash. I got itchy but that was controlled with antihistamines. I do lose my sense of taste and appetite but that was minor and manageable, although I did lose weight.”

The CBCN conducted a phone interview with a patient from the US who had direct experience with the treatment under review. This patient reported that she was personally satisfied with the treatment’s impact on her metastatic disease. The patient reported that she experienced side effects from alpelisib (hyperglycemia, nausea, and fatigue) but they were manageable. While the patient’s experience was in line with the experiences outlined in the Rethink Breast Cancer submission, the impact of fatigue on productivity and activity levels was highlighted. The patient also indicated that alpelisib was preferable compared to other treatments such as chemotherapy. Some comments from this patient included the following:

“Piqray is actually keeping the cancer under control even better than Ibrance did…. We looked at a couple of other things, my doctor and I, but because I had the mutation, the whole idea of precision medicine and focusing on the weak spots in the cancer specifically, that was why my doctor felt like it would be the best way to go.”

“When I was diagnosed, I had a super high disease load. So I went from so much disease to stability on Ibrance, but there was still a lot of active mets. And now I have one active mets. So it really was effective on the mets.”

“But outside of the hyperglycemia, pretty intense fatigue. For the hyperglycemia, I’m on Jardiance, and that’s kept the hyperglycemia under control. I do get a fair bit of nausea as well, and I’ve got a variety of medications that I take at different times of the day to keep the nausea under control. The fatigue: I drink a lot of coffee, and I’ve had to adjust my activity levels. The fatigue has been something that has been a side effect of every medication I’ve been on, so I feel that that’s a side effect that I’ve become a little bit more able to handle.”

Testing for the PIK3CA mutation was described in the Rethink Breast Cancer submission. Most respondents (79%) had received genomic testing and this was performed by blood test, tumour biopsy, or both. It should be noted that only 4 of the 24 respondents were from Canada. One of the Canadian patients interviewed described wait times of about 3 weeks for a biopsy and 4 weeks for the results, during which she experienced both excitement about having better information and anxiety over not having a treatment plan. The biopsy procedure was described as somewhat painful.

The patient groups reported significant financial challenges associated with treatments for metastatic breast cancer. The financial burden associated with living with advanced breast cancer includes loss of income and substantial costs associated with treatment and disease management. Patients indicated that the cost of medication, the cost of alternative treatments (e.g., massage, physiotherapy) to manage symptoms and side effects, and the time required to travel to treatment had a significant or debilitating impact on their quality of life. Other financial challenges were also reported, including not qualifying for insurance at work, the inability to change employers due to loss of insurance, and the prohibitive cost of new treatment options. CanCertainty noted that reimbursement of oral cancer drugs is not equal across jurisdictions in Canada. As a result, patients who do not have adequate insurance may have to pay out of pocket for medication and/or apply to funding assistance programs, which can take time and delay access to treatment. The groups also considered that for patients who do have private insurance, they may still have copayments, deductibles, and annual or lifetime caps that increase the financial burden on patients and families. Comments from patients included the following:

“Many of the next step treatments are very expensive [and not covered by government programs] and it is a HUGE struggle to get [coverage]. […] When dealing with an incurable disease the last thing you want to have to do is spend time on a letter writing campaign to argue about whether or not you should receive the drugs [recommended by your physician]. At about $1500.00 a week, I don't know many who can afford that.”

“When I turn 65 I will no longer have private insurance. I will not be able to afford the medication I currently take never mind any future medication that I may require.”

“I worry that in the future, a drug that may work for me won’t be accessible to me based on provincial formulary.”

Improved Outcomes

The patient groups indicated that the following measures of effectiveness were most important to respondents: PFS, OS, quality of life, and adverse effects. Reducing symptoms of the disease was also noted as an important outcome. Overall, patients indicated that controlling disease progression was most important to them. Patients indicated that they wanted treatments that delayed progression of their disease, prolonged life without sacrificing quality of life, and minimized the risk of side effects while stabilizing their disease.

The CBCN reported that it is very important for patients to have good quality of life when receiving treatment and the patients it speaks to acknowledge the importance of having the energy to attend their children’s activities and to spend time with family and friends. However, survey results indicated that patients are willing to tolerate side effects for drugs that can improve long-term health outcomes. Patients indicated that any treatment that gives people additional months or years of survival is beneficial. Many patient respondents indicated that treatment side effects such as fatigue, nausea, depression, problems with concentration, diarrhea, hair loss, and insomnia would be acceptable if the treatment extended PFS by approximately 6 months. Comments from patients included the following:

“Risks vs benefits. Some adverse side effects are worth the benefits for short term.”

“I can deal with pretty significant side effects if the outcome of treatment is optimistic.”

“I have a seven-year-old and nine-year-old. I’m not ready to leave them.”

“I will tolerate pretty much anything within reason in order to find and stay on a drug that keeps the tumour burden low.”

In addition to clinical outcomes, patients indicated that affordability and ease of access is important. Patients also expressed the need for personal choice regarding their treatment options. Patients want to be part of the decision-making process regarding which treatments they receive.

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, providing guidance on the potential place in therapy). The following input was provided by 2 clinical specialists with expertise in the diagnosis and management of breast cancer.

Unmet Needs

Current therapies for advanced or metastatic breast cancer beyond the first-line setting have low response rates and have not been shown to improve OS. Chemotherapy options are more poorly tolerated than endocrine therapy, causing nausea, vomiting, alopecia, fatigue, cytopenia, and sometimes dangerous adverse reactions. In addition, many available chemotherapy options are administered intravenously, requiring more hospital visits and reliance on institutions. The presence of PIK3CA gene mutations may be associated with poorer prognosis and resistance to endocrine therapy and alpelisib would be the first treatment available specifically for patients with PIK3CA-mutated cancer.

Place in Therapy

For patients harbouring a PIK3CA mutation, alpelisib would be added to an already established SOC option for the second-line treatment of advanced or metastatic hormone receptor–positive, HER2-negative breast cancer (i.e., fulvestrant). Alpelisib would not be used as a first-line treatment given the strong evidence for the use of CDK4/6 inhibitors with endocrine therapy in that setting. Additionally, alpelisib with fulvestrant would be a preferred treatment for patients with PIK3CA-mutated cancer and good performance status, as opposed to being reserved for patients who are intolerant of other treatments or for whom other treatments are contraindicated. Subsequent lines of therapy would include those previously used after fulvestrant monotherapy, including sequential single-agent chemotherapy drugs or investigational therapies in clinical trials.

If a patient received single-agent AI treatment in the first-line setting, it would be appropriate to recommend fulvestrant with a CDK4/6 inhibitor rather than fulvestrant with alpelisib. There is a lack of evidence from RCTs showing superiority of 1 approach over the other and the side-effect profiles of CDK4/6 inhibitors are favourable overall compared with alpelisib. In patients with life-threatening visceral organ metastases, chemotherapy would be recommended before considering treatment with alpelisib and fulvestrant.

Patient Population

Patients with advanced or metastatic hormone receptor–positive, HER2-negative breast cancer, activating mutations in the PIK3CA gene, good performance status (ECOG PS status of 0 or 1), expected survival of longer than 3 months, and no type 1 diabetes mellitus or uncontrolled type 2 diabetes mellitus would be best suited for treatment with the drug under review. As discussed earlier, patients should have progressed on first-line endocrine therapy and previously received treatment with a CDK4/6 inhibitor in the metastatic setting. Patients must be postmenopausal or rendered functionally postmenopausal. The presence of visceral metastases would not affect a patient’s eligibility.

A patient would not be suited for treatment with the drug under review if they have poor performance status, have type 1 diabetes mellitus or uncontrolled type 2 diabetes mellitus, are unable to understand and manage potential toxicities and dosing and monitoring requirements, or are non-compliant with follow-up. As mentioned before, patients with rapidly progressive visceral metastases may be better served by chemotherapy.

Patients with PIK3CA-mutated cancers are identified using liquid biopsy or tissue testing on archival or newly obtained tumour tissue. However, routine PIK3CA mutation testing is not part of the current SOC. PIK3CA mutation testing can be performed in commercial laboratories and academic hospitals in Canada, but it is not routinely funded or accessible. Ideally, testing would be offered to patients with advanced or metastatic hormone receptor–positive, HER2-negative breast cancer who are best suited for treatment with alpelisib (aside from the requirement for activating mutations in the PIK3CA gene). Since prevention of cancer symptoms is an important goal in the treatment of advanced or metastatic breast cancer, the presence of symptoms is not required to initiate treatment in this setting. In the absence of treatment, the disease will invariably progress and cause symptoms.

Some aspects of PIK3CA mutation testing have not been well characterized, such as concordance between methods (next-generation sequencing (NGS) versus polymerase chain reaction (PCR) methods; liquid biopsy versus tissue biopsy), concordance between tissue source (primary tumour versus metastasis), and stability of results over time. Evidence of concordance between liquid biopsy and tissue biopsy results has been summarized in a recent systematic review.¹⁵

Assessing Response to Treatment

Treatment response is monitored using a combination of clinical examination, laboratory evaluation (markers of organ function with or without tumour markers), and radiographic evaluation. Radiographic scans are initially performed at least every 3 months and toxicity or symptom assessments are initially performed every 2 weeks to 4 weeks or as needed. Clinical and laboratory evaluations are performed before each treatment cycle and may detect disease progression ahead of scheduled imaging. Treatment continues as long as the disease is stable or responding on radiographic scans according to the RECIST criteria.

Any of the following would also be considered a clinically meaningful response to treatment: improved survival (overall or progress-free), stabilization or reduction in the frequency or severity of cancer-related symptoms (e.g., pain, dyspnea), improvement in organ function (e.g., bone, liver, lung), maintenance or improvement of performance status and ability to perform activities of daily living, and delay in initiation of chemotherapy (especially IV chemotherapy).

Discontinuing Treatment

Treatment with alpelisib and fulvestrant should be discontinued if there is disease progression, intolerable or dangerous toxicity (especially uncontrollable hyperglycemia), or an event or development of a comorbidity that adversely impacts performance status or survival (e.g., stroke). Patient preference would also dictate treatment discontinuation.

Prescribing Conditions

Treatment with alpelisib and fulvestrant would be prescribed by medical oncologists or associated team physicians with expertise in cancer therapies and toxicity management. Patients would be treated on an outpatient basis under medical oncology supervision, which can include help from family physicians and/or nurse practitioners with additional training in oncology. Fulvestrant injections would be administered in a hospital outpatient clinic or family doctor’s office.

Clinician Group Input

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

One clinician group submission was received from 6 clinicians with the Breast Medical Oncology group at the Ottawa Hospital Cancer Centre, an academic teaching hospital centre in Ontario. The group offers routine SOC treatments and access to promising treatments in phase I to phase III clinical trials. The group serves a large referral base from the Champlain Local Health Integration Network in Ontario.

Unmet Needs

The most important treatment goals in this disease setting are to maintain or improve quality of life compared with currently available treatments, delay the progression of cancer, improve or maintain organ function, reduce cancer symptoms, and allow patients to be treated with oral therapy with a view to allowing patients to remain gainfully employed and independent, and preventing institutionalization.

Current therapies have the following shortcomings in terms of achieving treatment goals: disappointing responses to second-line therapy and later therapies and no improvement in OS, patients becoming more rapidly refractory to current therapies, and chemotherapy options that are poorly tolerated due to numerous side effects and sometimes dangerous adverse reactions.

Place in Therapy

Alpelisib with fulvestrant, the latter of which is an already established SOC second-line therapy, would be a paradigm-changing SOC option in the second line for patients with a PIK3CA mutation. It would replace fulvestrant monotherapy and lines of therapy following alpelisib with fulvestrant would be the same as those previously used following fulvestrant monotherapy. The improvement in PFS found in the SOLAR-1 study of 5 months would be considered worthwhile by patients and clinicians.

It would be appropriate to use standard first-line therapy, which includes CDK4/6 inhibitors, before using alpelisib plus fulvestrant. Chemotherapy would be recommended beforehand in patients with life-threatening visceral organ metastases.

Patient Population

The patients best suited for treatment with alpelisib and fulvestrant are those meeting the eligibility criteria for the SOLAR-1 study and with activating mutations of PIK3CA. Patients should also have ECOG PS of 0 to 1 with expected survival of more than 3 months, should not have diabetes mellitus type 1 or uncontrolled type 2 (glycated hemoglobin of greater than 6.4%), and should be able to understand and comply with the specific safety, monitoring, and side-effect management issues associated with the drug. Patients would be eligible with or without visceral metastases, with brain metastases if controlled or treated, and with ovarian function suppression to achieve postmenopausal state if not already postmenopausal. Patients with an ECOG PS of 2 would potentially be considered for treatment with alpelisib and fulvestrant if they were fit with aggressive disease but unfit for chemotherapy. Alpelisib would uniquely address the needs of patients with relevant PIK3CA mutations progressing after standard first-line therapy. Aside from activating PIK3CA mutations, patient subgroups and other clinical factors cannot be used to select patients who are likely to derive the greatest benefit from treatment with alpelisib and fulvestrant.

Patients least suitable for treatment with alpelisib and fulvestrant are those with ECOG PS of 2 to 4 and those with diabetes mellitus type 1 or uncontrolled type 2. One of the clinicians felt that, without evidence, patients who had first-line CDK4/6 inhibitor treatment would not be suitable for treatment with alpelisib, while most of the clinicians felt that access and funding for alpelisib should not be withheld. There was mention of further trials under way for assessing the efficacy of alpelisib after first-line treatment with CDK4/6 inhibitors.

Access to molecular testing for PIK3CA mutation status would be required to identify patients for treatment. Testing is not challenging but is not routinely funded or accessible currently in Canada. It can be done by commercial laboratories or in academic hospital laboratories if funded.

Assessing Response to Treatment

Response to treatment is determined based on symptoms, laboratory markers, and radiographic scans and tumour measurements. Scans are usually initially performed at least every 3 months and treatment is continued if disease is stable or responding radiographically according to RECIST criteria. Any of the following would be considered a clinically meaningful response to treatment: reduction in the frequency or severity of symptoms, improvement of organ function, stabilization of symptoms, and maintenance or improvement of performance status. Toxicity or symptom assessments would be performed every 2 weeks to 4 weeks early in treatment or as needed.

Discontinuing Treatment

The following factors should be considered when deciding to discontinue treatment: disease progression, intolerable or dangerous toxicity (especially uncontrolled grade 3 or grade 4 hyperglycemia, rash, or diarrhea), and patient preference or refusal.

Prescribing Conditions

Treatment should only be prescribed by certified medical oncologists or associated team physicians with expertise in cancer therapies and toxicity management. Treatment would take place in the community setting (for alpelisib), and in hospital outpatient clinics or family doctors’ offices (for fulvestrant).

Additional Considerations

Substantial discordance of opinion was noted regarding treatment of patients with prior first-line CDK4/6 inhibitor therapy. Some clinicians were of the opinion that the benefits of alpelisib in this population were uncertain and the toxicities with alpelisib were significant, with 1 clinician noting that capecitabine would likely yield a better therapeutic index. Other clinicians noted the lack of recent advances in this niche and were of the opinion that the benefits of alpelisib are commensurate with patient values and that the toxicities, though substantial, are predictable and manageable by medical oncologists.

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 alpelisib 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 indirect evidence selected from the literature that met the selection criteria specified in the review. The third section includes 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 50 mg, 150 mg, and 200 mg alpelisib tablets in combination with fulvestrant for the treatment of postmenopausal women, and men, with hormone receptor–positive, HER2-negative, PIK3CA-mutated advanced or metastatic breast cancer after disease progression following an endocrine-based regimen with a CDK4/6 inhibitor

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.¹⁶

Published literature was identified by searching the following bibliographic databases: MEDLINE All (1946‒) via Ovid and Embase (1974‒) via Ovid. The search strategy comprised both controlled vocabulary, such as the US National Library of Medicine’s MeSH (Medical Subject Headings), and keywords. The main search concepts were Piqray or alpelisib. Clinical trials registries were searched: the US National Institutes of Health’s ClinicalTrials.gov, the WHO’s International Clinical Trials Registry Platform 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 May 19, 2021. Regular alerts updated the search until the meeting of the CADTH pan-Canadian Oncology Drug Review Expert Committee 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.¹⁷ Included in this search were the websites of regulatory agencies (the 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.

A focused literature search for network meta-analyses dealing with breast cancer was run in MEDLINE All (1946‒) on May 19, 2021. No limits were applied to the search.

Findings From the Literature

A total of 1 study was identified from the literature for inclusion in the systematic review (Figure 1). The included study is 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 relevant study, the SOLAR-1 study, was selected for inclusion in the CADTH systematic review. The SOLAR-1 study was a pivotal study provided in the sponsor’s submission to CADTH and Health Canada and was also identified in the CADTH systematic literature search. The SOLAR-1 study was a placebo-controlled, double-blind, parallel-group RCT with a primary objective of determining whether treatment with alpelisib and fulvestrant prolongs PFS compared with placebo and fulvestrant for patients with PIK3CA-mutated advanced breast cancer. The key secondary objective was to compare OS between the same groups as stated in the primary objective.

The SOLAR-1 study (N = 572; enrolment from 2015 to 2017) randomized patients 1:1 to alpelisib 300 mg daily or matching placebo, in combination with fulvestrant 500 mg on day 1, day 15, and day 29, and every 28 days afterwards. Patients were randomized within each of 2 cohorts based on the results of real-time PCR tumour tissue testing for mutations in exon 7, exon 9, and exon 20 of the PIK3CA gene: the PIK3CA mutant cohort and the PIK3CA non-mutant cohort. Randomization was stratified according to whether patients had received prior CDK4/6 inhibitor treatment (the total number of patients in the “yes” category was to be no more than 30% of the overall study population) and whether they had lung and/or liver metastases. The study was conducted in North America (including 9 sites in Canada), South America, Europe, Asia, and Australia. The data cut-off date for the analysis of the primary end point in the SOLAR-1 study was June 12, 2018. The final OS analysis was performed using a data cut-off date of April 23, 2020.

During the first 14 days of the 35-day screening phase, a tumour sample from archived tissue or a new biopsy was collected and sent to a sponsor-designated laboratory for PIK3CA mutation testing before randomization.

In the original study protocol, the primary and key secondary end points were to be evaluated in both the PIK3CA mutant and non-mutant cohorts. Following the first protocol amendment, PFS and OS in the PIK3CA non-mutant cohort became secondary end points.

At the sponsor’s request that the reimbursement criteria focus on patients with prior treatment with an endocrine-based regimen with a CDK4/6 inhibitor, the present systematic review reports efficacy results only for the stratum of patients in the PIK3CA mutant cohort who had received prior CDK4/6 inhibitor treatment (N = 20). The BYLieve study — a phase II, non-comparative trial with 1 cohort (N = 127) meeting the systematic review criteria for population, intervention, and outcomes — is summarized in the Other Relevant Evidence section of this report.

Populations

Inclusion and Exclusion Criteria

The main inclusion and exclusion criteria for the SOLAR-1 study are provided in Table 6. Women were eligible if they were postmenopausal, with the definition of postmenopausal excluding menopause induced by ovarian suppression. Patients had to have radiological or objective evidence of recurrence or progression during or after AI therapy, and it was not specified that this therapy had to be in the advanced or metastatic setting. While patients who were endocrine sensitive were initially enrolled, a protocol amendment later clarified that patients without previous treatment for metastatic disease had to have relapsed with progression while at or within 12 months of completion of adjuvant or neoadjuvant endocrine therapy. Patients who had relapsed with progression while at or within 12 months of completion of adjuvant endocrine therapy and then relapsed with progression after 1 line of endocrine therapy for metastatic disease were not included. Patients were excluded if they had received prior treatment with chemotherapy other than adjuvant or neoadjuvant chemotherapy, fulvestrant, or any PI3K, mammalian target of rapamycin, or protein kinase B inhibitor. Patients had to have an ECOG PS of 0 or 1. Patients with any of the following were excluded: patients with diabetes mellitus type 1 or uncontrolled diabetes mellitus type 2 or hyperglycemia, patients with inflammatory breast cancer, and patients with symptomatic visceral disease or any disease burden making the patient ineligible for endocrine therapy.

Baseline Characteristics

The primary end point in the SOLAR-1 study was PFS in the entire PIK3CA mutant cohort (N = 341) and results in patients in that cohort with previous CDK4/6 inhibitor treatment (N = 20) were only available in subgroup analyses. Since summaries of patient disposition, treatment exposure, and protocol deviations were not available for the relevant patient population for the present review, these data are presented for the entire PIK3CA mutant cohort. Baseline characteristics are presented for both the entire PIK3CA cohort and the subgroup with prior CDK4/6 inhibitor treatment; data for certain characteristics (disease stage at study entry, lines of prior medication therapy, and lines of prior hormonal therapy in metastatic setting) were not available for the subgroup.

Detailed information on key baseline characteristics for the full analysis set (FAS) in the PIK3CA mutant cohort in the SOLAR-1 study and in the subgroup with prior CDK4/6 inhibitor treatment are provided in Table 7. Overall, in the entire PIK3CA mutant cohort, baseline characteristics were balanced between the treatment groups, though there were some imbalances with regard to metastatic sites. Within the subgroup with prior CDK4/6 inhibitor treatment, there were imbalances in the alpelisib group versus the placebo group in ECOG PS (44.4% versus 72.7% with ECOG PS of 0), numbers of sites of metastases (44.5% versus 18.2% with 1 site and no patients versus 18.2% with 3 sites), and progesterone receptor status (55.6% versus 81.8% with positive status).

Within the entire PIK3CA mutant cohort, there was an approximately even split between patients who had received 1 line or no lines of endocrine therapy in the metastatic setting, with less than 3% having received 2 previous lines. This information was not reported for the subgroup with prior CDK4/6 inhibitor treatment and it was not clear how many of those patients had received 2 previous lines of endocrine therapy. In the subgroup, most patients (66.7% and 72.7% within each group) were categorized as having secondary endocrine resistance, while some were categorized under primary endocrine resistance (33.3% and 18.2%) or categorized as endocrine sensitive (11.8% and 11.0%).

The numbers of patients with prior CDK4/6 inhibitor treatment, as identified for randomization stratification with the interactive response technology, were 9 in the alpelisib group and 11 in the placebo group. These patients were the ones included in subgroup analyses for the prior CDK4/6 inhibitor treatment subgroup and it is these analyses that are presented in the current report. However, according to clinical data collected in case report forms and reported in summaries of prior antineoplastic therapy, the numbers of patients with prior CDK4/6 inhibitor therapy (those with prior abemaciclib, palbociclib, or ribociclib therapy) were 11 in the alpelisib group and 8 in the placebo group. Finally, a concordance table in the SOLAR-1 Interim Clinical Study Report comparing numbers of patients within each group according to each source of data showed 12 patients and 11 patients in the alpelisib and fulvestrant groups, respectively, according to the clinical data. According to the concordance table, all 9 patients in the alpelisib group in the randomization stratum were identified as having prior CDK4/6 inhibitor therapy in the clinical data. In contrast, 10 of 11 patients in the placebo groups were common between the randomization stratum and the clinical data. The reasons for these differences are unclear.

Interventions

Patients were assigned based on tissue testing to the PIK3CA mutant cohort or PIK3CA non-mutant cohort following confirmation of eligibility. Patients were randomized and allocated by an interactive response technology provider, with patients randomized in each cohort 1:1 to the alpelisib or placebo group using previous treatment with CDK4/6 inhibitor(s) (yes or no) and the presence of lung and/or liver metastasis (yes or no) as stratification factors. Patients and investigators were blinded to PIK3CA mutation status throughout the study. Patients, investigators, and local radiologists were blinded to treatment assignment throughout the study. The sponsor’s clinical trial team was blinded to treatment assignment until the database was locked for the main PFS analysis, subsequent to the June 12, 2018, data cut-off date. An independent data monitoring committee reviewed safety and efficacy data during the study and an independent statistical group, external to the sponsor and not involved in study conduct, prepared the data monitoring committee reports. Interim PFS analyses were performed by an independent statistician who was also not involved in study conduct.

Patients in the alpelisib group received alpelisib as 50 mg and 200 mg film-coated tablets and patients in the placebo group received placebo tablets that were identical to alpelisib medication in appearance and packaging. The starting dosage for alpelisib and placebo was 300 mg taken orally once daily, with instructions to take the medication within 1 hour after a meal at approximately the same time each day. Patients were given an adequate supply at each study visit, instructions for administration, and medication diaries. Treatment adherence was assessed using pill counts and information provided by the patients and/or caregiver. Fulvestrant 500 mg was administered intramuscularly on day 1 and day 15 of cycle 1 and on day 1 (± 3 days) of each subsequent 28-day cycle.

A maximum of 2 dose reductions of 50 mg each was permitted for alpelisib and placebo (with no re-escalation permitted) to manage intolerance of study medication, while no dose adjustments were permitted for fulvestrant. Guidance for dose modifications for alpelisib and placebo was outlined in the study protocol. Patients were required to permanently discontinue treatment with alpelisib or placebo if they required a dose delay of more than 28 days. Patients were required to permanently discontinue treatment with fulvestrant if it was withheld for more than 35 days since a planned injection. If 1 study medication was discontinued, patients could continue treatment with the other study medication.

Patients taking alpelisib or placebo in combination with fulvestrant were not permitted to take medications with a known risk for torsades de pointes, other investigational and antineoplastic therapies, or herbal preparation and/or medications and dietary supplements (except for vitamins). However, medications required to treat AEs, medications to manage cancer symptoms and concurrent diseases, and supportive care agents were permitted.

Patients could voluntarily withdraw from either or both study treatments, and an investigator could discontinue either or both study treatments if continuation was considered to be detrimental to the patient’s well-being. Patients taking only 1 of the 2 study medications continued with study visits. Patients who discontinued both medications were to attend an end-of-treatment visit within 14 days of study treatment discontinuation and were followed up for safety data until 30 days after the last dose.

Outcomes

A list of efficacy end points identified in the CADTH review protocol that were assessed in the SOLAR-1 study is provided in Table 8. These end points are further summarized as follows. Although health-related quality of life (HRQoL) was assessed in the SOLAR-1 study using the EQ-5D Five-Level (EQ-5D-5L) questionnaire and the European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30 (EORTC QLQ-C30), there were no analyses of those outcomes in the subgroup with prior CDK4/6 inhibitor treatment.

Efficacy

Progression-Free Survival

PFS was based on local radiology of tumour assessments according to the RECIST 1.1 guideline.²³ Tumour assessments were also reviewed centrally by a blinded independent review committee using an audit-based approach to provide supportive analyses. Imaging assessments used the same imaging modality used at baseline and were performed every 8 weeks (± 7 days) during the first 18 months following randomization and every 12 weeks (± 7 days) thereafter until disease progression, death, withdrawal of consent, loss to follow-up, a patient or guardian decision, or end of treatment. Assessments continued after the start of new antineoplastic therapy. If there was clinical suspicion of disease progression, physical examination and imaging assessment were performed promptly. Baseline imaging assessments included CT or MRI of the chest, abdomen, and pelvis (and of the brain if clinically indicated), a whole body bone scan, localized bone CT, MRI, or X-ray for lesions identified on a whole body bone scan not visible on a chest, abdomen, or pelvis CT or MRI, colour photography for skin lesions, and CT or MRI of other metastatic sites if clinically indicated. Complete physical examinations were performed during screening (within 14 days of randomization), on day 8 and day 15 of cycle 1, on day 1 and day 15 of cycle 2, and on day 1 of each cycle thereafter.

Overall Survival

Following the discontinuation of study treatment and tumour assessments, survival follow-up continued until the final number of OS events was reached (178) or the study was stopped for other reasons. Survival follow-up was conducted through clinical visits or telephone calls every 12 weeks (± 1 week) until death, loss to follow-up, or withdrawal of consent for survival follow-up.

Adverse Events

AEs that began or worsened after patient consent were recorded at each study visit until at least 30 days following discontinuation of study treatment. AEs were detected by non-directive questioning at visits, when volunteered by patients between visits, or through physical examination, laboratory test, or other assessments.

Statistical Analysis

The primary outcome of the study was PFS in the FAS in the PIK3CA mutant cohort based on local radiology assessment. PFS was defined as the time from the date of randomization to the date of first documented disease progression or death due to any cause. Patients were censored at the last tumour assessment if they did not have an event or if they had an event that occurred after at least 2 missing tumour assessments. If a patient was missing the baseline tumour assessment, that patient was censored at the date of randomization for progressive disease events. The survival distribution was estimated using Kaplan-Meier analysis.

A 1-sided 2.0% significance level was used for rejecting the null hypothesis that the log-hazard ratio was 0 or more (or the hazard ratio was ≥ 1) for the alpelisib group versus the placebo group. A stratified log-rank test, with strata based on prior CDK4/6 inhibitor treatment status and the presence of lung and/or liver metastases, was used to compare PFS between treatment groups. The hazard ratio for PFS and its 95% CI were estimated using a stratified Cox proportional hazards model, using the aforementioned strata.

Haybittle-Peto boundaries were used to control for overall type I error rate in a 3-look group sequential design with an interim futility analysis planned after observing approximately 97 PFS events (40% of expected events) and an interim efficacy analysis for superiority after approximately 185 PFS events (76% of expected events). A final PFS analysis was to be performed after approximately 243 events. Based on the planned numbers of events, the significance levels for the interim and final PFS efficacy analyses were 0.0001 and 0.0199, respectively.

OS in the PIK3CA mutant cohort was the key secondary end point and was compared between treatment groups using the same methods as for PFS, though a separate Lan-DeMets (O’Brien-Fleming) alpha spending function was used to control type I error rate with an overall alpha of 2.5%. OS could be tested at the interim PFS efficacy analysis, but only if PFS was statistically significant. If OS was not significant at the interim analysis, a second OS analysis was performed at approximately 85% of expected deaths. If OS was not tested at the interim PFS efficacy analysis, it was tested at the final PFS analysis instead. A final OS analysis (if previously OS analyses were not significant) was planned at approximately 178 deaths if the final PFS was significant.

It was calculated that a total of 243 PFS events would be required for 83.8% power to reject the null hypothesis, assuming a true hazard ratio of 0.6 or an increase in median PFS from 7.0 months to 11.67 months. A total of 340 patients randomized in the PIK3CA mutant cohort would be needed, based on 40% of patients having a PIK3CA mutation; enrolment rates of 12 patients, 35 patients, and 59 patients per month during the first 6 months, the next 6 months, and the remainder of the study, respectively; and a 10% loss to PFS follow-up. With 340 patients randomized and a 5% loss to follow-up for OS, it was estimated that 178 deaths would occur by the OS final analysis and that with a true hazard ratio of 0.67, there would be 72% power to reject the null hypothesis for the key secondary end point.

Subgroup Analyses

Pre-specified subgroup analyses for the primary and key secondary end points were planned if statistical significance was reached. The same analysis methods that were used for the primary end point were applied to each subgroup. The subgroup analyses did not account for multiplicity and were not part of the sample size considerations for the study. In the analyses for the subgroup of patients with prior CDK4/6 inhibitor treatment, a Cox proportional hazards model was used and was stratified by the presence of lung and/or liver metastases.

Sensitivity Analyses

Sensitivity analyses for the primary end point were planned if statistical significance was reached. In the subgroup of patients with prior CDK4/6 inhibitor treatment, the primary end point analysis was to be repeated in the per-protocol (PP) set, using different censoring methods (an actual event approach, a backdating approach, and the approach of censoring patients after a new antineoplastic therapy), and using stratification factors based on the clinical database as opposed to those used in the interactive response technology. In the actual date censoring approach, events were included even if they occurred following 2 or more missing tumour assessments. In the backdating approach, events occurring after at least 1 missing assessment were assigned the date of the next scheduled assessment.

Analysis Populations

All patients randomized to study treatment were included in the FAS and were analyzed according to the cohort, treatment, and strata assigned during randomization. Patients in the PP set were those from the FAS without certain protocol deviations that were specified in the statistical analysis plan. The safety analysis set included all patients who received any study treatment, with patients analyzed according to the study treatment received.

Results

Patient Disposition

Detailed patient disposition for the SOLAR-1 study in the entire PIK3CA mutant cohort for data cut-off dates representing the final PFS analysis (June 12, 2018, data cut-off date) and final OS analysis (April 23, 2020, data cut-off date) is presented in Table 10. The percentages of patients discontinuing from the study were 4.1% and 2.9% in the alpelisib and placebo groups, respectively, at the final PFS analysis and 5.9% and 2.9% at the final OS analysis. The most common reason for study discontinuation, progressive disease, was more common in the alpelisib group versus the placebo group (2.4% versus 0.6% at the final PFS analysis and 3.0% and 0.6% at the final OS analysis). Patient disposition information within the subgroup of interest — patients with prior CDK4/6 inhibitor treatment — was unavailable.

The most common protocol deviations at the final OS analysis in the SOLAR-1 study are presented in Table 11. The most common protocol deviation was use of prohibited concomitant medication (20.7% in the alpelisib group and 14.0% in the placebo group), though in most cases, the medication was 1 prohibited for safety reasons (i.e., those with known risk for torsades de pointes).

The reasons for the exclusion of patients from the PP set for the primary end point analysis in the PIK3CA mutant cohort were as follows: the patient relapsed at or within 12 months from completion of adjuvant or neoadjuvant endocrine therapy and subsequently progressed after 1 line of endocrine therapy or progressed on more than 1 line of endocrine therapy for metastatic disease, the patient was not postmenopausal, there was prohibited concomitant medication at baseline, the criteria for measurable disease or lytic bone lesion was not met, the criteria for prior antineoplastic therapy was not met, breast cancer type (HR status) was not met, or no study treatment was taken. Aside from the first reason, no more than 3.0% of patients in a group were excluded for each reason. In total, 14.8% and 14.0% of patients in the alpelisib and placebo groups were excluded from the PP set, respectively. Protocol deviation information within the subgroup of interest, patients with prior CDK4/6 inhibitor treatment, was unavailable.

Exposure to Study Treatments

Details on the duration of treatment exposure and study treatment discontinuations at the time of the final OS analysis (April 23, 2020, data cut-off date) are provided in Table 12. Most patients had discontinued alpelisib or placebo at the final PFS analysis (82.2% for alpelisib and 84.2% for placebo; these values are not shown in Table 12) and at the final OS analysis (91.7% for alpelisib and 97.7% for placebo). At the final OS analysis, treatment discontinuation due to progressive disease was less common in the alpelisib group compared with the placebo group (47.9% versus 78.9%) while discontinuation due to AE was more common in the alpelisib group than in the placebo group (26.6% versus 5.8%). The percentages of patients discontinuing fulvestrant were similar to those for alpelisib and placebo, though discontinuation due to AE was less common (3.6% for alpelisib and 1.8% for placebo). Treatment exposure information within the subgroup of interest, patients with prior CDK4/6 inhibitor treatment, was unavailable.

Follow-up duration and reasons for censoring patients for the primary end point analysis and final OS analysis are presented in Table 13. Median follow-up duration for PFS at the primary end point analysis (June 12, 2018, data cut-off date) was 20.2 months and 19.9 months in the alpelisib group and placebo group, respectively, in the overall PIK3CA mutant cohort. In the subgroup with prior CDK4/6 inhibitor treatment, the median follow-up duration was similar (19.0 months in the alpelisib group and 18.0 months in the placebo group). At the final OS analysis (April 23, 2020, data cut-off date), the mean follow-up duration was 42.6 months in the alpelisib group and 42.3 months in the placebo group, with similar median follow-up durations in the subgroup with prior CDK4/6 inhibitor treatment (41.4 months and 40.4 months).

Efficacy

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

Progression-Free Survival

The results for PFS in patients with prior CDK4/6 inhibitor treatment in the PIK3CA mutant cohort are presented in Table 14. At the primary PFS analysis at the June 12, 2018, data cut-off date, median PFS was 5.5 months (95% CI, 1.58 months to 16.76 months) in the alpelisib group and 1.8 months (95% CI, 1.68 months to 3.58 months) in the placebo group. The hazard ratio for the alpelisib group versus the placebo group was 0.48 (95% CI, 0.17 to 1.36). The results were identical between the primary PFS analysis (June 12, 2018, data cut-off) and the final PFS analysis (April 23, 2020, data cut-off date). Conducting sensitivity analyses in the PP set, using different censoring methods, and using stratification factors based on the clinical database yielded results consistent with the primary analysis.

Kaplan-Meier curves for PFS in the 2 treatment groups in patients with prior CDK4/6 inhibitor treatment are shown in Figure 2.

Figure 2: Progression-Free Survival in Patients with Prior CDK4/6 Inhibitor Treatment

CDK4/6 = cyclin-dependent kinase 4 and 6; CI = confidence interval; fulv = fulvestrant.

Source: SOLAR-1 study interim Clinical Study Report (2018).¹

Overall Survival

The results for OS in patients with prior CDK4/6 inhibitor treatment in the PIK3CA mutant cohort are presented in Table 15. At the final OS analysis at the April 23, 2020, data cut-off date, the median OS was 29.8 months (95% CI, 6.67 months to 38.21 months) in the alpelisib group and 12.9 months (95% CI, 2.46 months to 34.60 months) in the placebo group. The hazard ratio for the alpelisib group versus the placebo group was 0.67 (95% CI, 0.21 to 2.18).

Kaplan-Meier curves for OS in the 2 treatment groups in patients with prior CDK4/6 inhibitor treatment are shown in Figure 3.

Figure 3: Overall Survival in Patients With Prior CDK4/6 Inhibitor Treatment

CDK4/6 = cyclin-dependent kinase 4 and 6; CI = confidence interval; fulv = fulvestrant.

Source: SOLAR-1 study final Clinical Study Report (2020).²

Health-Related Quality of Life

Although HRQoL was assessed in the SOLAR-1 study using the EQ-5D-5L questionnaire and the EORTC QLQ-C30, analysis in the subgroup of patients with prior CDK4/6 inhibitor treatment was not available for HRQoL outcomes. Results for the index score of the EQ-5D-5L for the entire PIK3CA mutant cohort are presented in Appendix 3 to provide context for the CADTH Pharmacoeconomic Review Report.

Harms

Only those harms identified in the review protocol are reported as follows. Results for harms are presented for the entire PIK3CA mutant cohort (this aligns with the Health Canada–approved indication for alpelisib). The CADTH review team did not request data from the sponsor on AEs in the subgroup with prior CDK4/6 inhibitor treatment, since AE data were expected to be more informative in the entire PIK3CA mutant cohort than in the subgroup and results for harms were not expected by the clinicians consulted by CADTH to differ between patients who had or had not previously received CDK4/6 inhibitor treatment. See Table 16 for detailed harms data at the final OS analysis (April 23, 2020, data cut-off date).

Adverse Events

Almost all patients reported at least 1 AE (99.4% in the alpelisib group and 90.6% in the placebo group). Most of the AEs occurring in at least 10% of patients of at least 1 treatment group were more common in the alpelisib group compared with the placebo group. All of the AEs reported by more than 20% of patients in the alpelisib group were also more common in the alpelisib group than in the placebo group: hyperglycemia, diarrhea, nausea, rash, decreased appetite, decreased weight, stomatitis, vomiting, fatigue, and alopecia. The only AE reported by more than 20% of patients in the placebo group was nausea (20.5%).

Medications or therapies for AEs were required for 96.4% of patients in the alpelisib group and 63.8% of patients in the placebo group. All of the AEs requiring medications or therapies occurring in at least 10% of 1 treatment group were more common in the alpelisib group compared with the placebo group and were as follows: hyperglycemia, diarrhea, rash, nausea, stomatitis, maculopapular rash, urinary tract infection, and mucosal inflammation.

Serious Adverse Events

SAEs were reported in 39.6% of patients in the alpelisib group and 19.9% of patients in the placebo group. The most common SAEs were hyperglycemia (10.1% in the alpelisib group and none in the placebo group), osteonecrosis of the jaw (3.6% in the alpelisib group and none in the placebo group), stomatitis, acute kidney injury, and rash (2.4% in the alpelisib group and none in the placebo group for the preceding 3 SAEs).

Withdrawals Due to Adverse Events

Withdrawals from treatment due to AEs were more common in the alpelisib group (27.2%) versus the placebo group (5.8%). The most common AEs leading to discontinuation were reported in the alpelisib group alone: hyperglycemia (6.5%), rash (4.7%), and diarrhea (3.6%).

Mortality

On-treatment deaths were defined as those occurring within 30 days of the last dose of study treatment. The most common cause of on-treatment death was breast cancer (3.6% in the alpelisib group and 4.1% in the placebo group). Other causes of on-treatment death were reported for 1 patient each: cardiorespiratory arrest in the alpelisib group and gastrointestinal hemorrhage, pneumonia, and septic shock in the placebo group.

Notable Harms

The following notable harms identified in the systematic review protocol occurred in more than 10% of patients of at least 1 treatment group and were more common in the alpelisib group: hyperglycemia, diarrhea, nausea, rash, vomiting, and maculopapular rash. The following notable harms occurred in 3.6% or less of each treatment group: hypersensitivity, anaphylactic reaction, drug hypersensitivity, increased glycated hemoglobin, increased blood glucose, pneumonitis, erythema multiforme, glucose urine present, Stevens-Johnson syndrome, diabetes mellitus, ketoacidosis, and type 2 diabetes mellitus.

Critical Appraisal

Internal Validity

The methods for randomization, treatment allocation, and maintenance of blinding to treatment assignment were appropriate in the SOLAR-1 study. However, the differences in side-effect profiles between the treatment groups may have resulted in treatment unblinding in both patients and investigators. For example, hyperglycemia was reported in 66.9% of patients in the alpelisib group and 8.8% of patients in the placebo group, while 26.6% of patients in the alpelisib group discontinued treatment due to AEs compared with 5.8% of patients in the placebo group. Since PFS was assessed objectively (with the radiologist remaining blinded to treatment assignment) and there was regular follow-up for PFS and OS, the risk of bias from potential treatment unblinding was low for determining these outcomes. However, the bias caused by potential differences in care, increased monitoring, or missing visits is unclear.

There were discrepancies among different approaches for identifying the subgroup of patients in the PIK3CA cohort with prior CDK4/6 inhibitor treatment, with the approach in the main analysis being the use of the randomization stratum (9 patients and 11 patients in the alpelisib group and placebo group, respectively). Although there was a sensitivity analysis of PFS performed in the subgroup of patients with prior CDK4/6 inhibitor treatment as identified by the clinical data, the numbers of patients in this sensitivity analysis (12 in the alpelisib group and 11 in the placebo group) did not match those reported in the summary of prior antineoplastic therapy (11 in the alpelisib group and 8 in the placebo group). It is uncertain if there were patients in the main analysis and sensitivity analysis who hadn’t received prior CDK4/6 inhibitor treatment and if this was more common in 1 group than the other. It is also possible that a number of patients who had received prior CDK4/6 inhibitor treatment were not included in the main analysis and sensitivity analysis. The potential impact on the results from differing approaches in identifying patients with prior CDK4/6 inhibitor treatment for analysis is unclear.

Within the whole PIK3CA mutant cohort, some imbalances in baseline characteristics were noted; however, the clinical experts consulted by CADTH for this review did not consider them likely to impact the efficacy and safety results. Within the subgroup with prior CDK4/6 inhibitor treatment, there was a lower percentage of patients in the alpelisib group with an ECOG PS of 0 (44.5% in the alpelisib group versus 72.7% in the placebo group), with potential bias in favour of the placebo group.

Despite the high percentages of patients discontinuing treatment in both groups, the percentages of patients discontinuing from the study were lower than 6% in each group. The percentages of patients discontinuing the study were higher in the alpelisib group versus the placebo group (4.1% versus 2.9% at the primary PFS analysis and 5.9% versus 2.9% at the final OS analysis). Similar trends were noted in percentages of patients discontinuing the study due to progressive disease. There was also a higher percentage of patients in the alpelisib group censored from the primary PFS analysis due to withdrawn consent than in the placebo group (5.9% versus 2.9%). The potential direction of bias is unclear for these imbalances in discontinuation and censoring, although the low percentages in each category (less than 6%) mean that any impact on the efficacy results is likely to be minor in the PIK3CA mutant cohort. However, patient disposition was not available for the subgroup with prior CDK4/6 inhibitor treatment and it is unclear what impact study discontinuations may have had on results in that subgroup.

There were some analytical issues of note in the OS analysis. The percentages of patients lost to follow-up were 10.7% and 9.9% in the alpelisib group and placebo group, respectively, and the potential impact of this missing data is uncertain. There were also no PP analyses conducted for OS in the subgroup with prior CDK4/6 inhibitor treatment. In addition, it is unclear from visual inspection of the OS Kaplan-Meier curves in Figure 3 whether the proportional hazards assumption was met. Additionally, no consideration was made for informative censoring via adjustment in the Cox model in the PFS and OS analyses, nor was the small sample size considered, as the well-known statistical properties of the Cox model likely do not hold in such a sample size.

Regardless of the aforementioned issues with internal validity in the SOLAR-1 study, the main limitation of the study that precludes any assessment of comparative efficacy in the population with prior CDK4/6 inhibitor treatment is that the study was not designed to test hypotheses in that population. The 2 outcomes in the testing hierarchy were PFS and OS in the entire PIK3CA mutant cohort and the study sample size considerations were based on those outcomes. As described in the SOLAR-1 Interim Clinical Study Report, “Analyses of efficacy in subgroups were intended to explore the intrinsic consistency of any treatment effect.” In the absence of pre-specified subgroup analyses within the statistical testing hierarchy, no conclusions can be drawn about efficacy in any of the subgroups.

External Validity

As with internal validity, the main limitation of the SOLAR-1 study in terms of external validity is that the objectives of the study do not align with the relevant population for this review. The clinical experts consulted by CADTH for this review indicated that the results for the entire PIK3CA cohort cannot be used to inform the comparative efficacy of alpelisib specifically in patients with prior CDK4/6 inhibitor treatment. This uncertainty, combined with the inability to draw conclusions in the subgroup with prior CDK4/6 inhibitor treatment, means that there is insufficient evidence to support the efficacy of alpelisib and fulvestrant compared with fulvestrant alone in the relevant population for this review. It is important to note that the population targeted by the sponsor’s reimbursement request largely represents the population that would be considered for alpelisib and fulvestrant therapy in Canadian clinical practice, since an endocrine regimen with a CDK4/6 inhibitor is the current SOC for first-line treatment for advanced and metastatic breast cancer. Another issue with the study population is that the SOLAR-1 study included patients with prior endocrine therapy, regardless of setting, and 52% of patients in the PIK3CA mutant cohort had not received endocrine therapy in the metastatic setting. In contrast, patients in Canada who are suited for treatment with alpelisib and fulvestrant will have received endocrine therapy in the metastatic setting, according to the clinical experts consulted by CADTH.

Female patients receiving an LHRH agonist for induction of ovarian suppression were excluded from the SOLAR-1 study, but the clinical experts consulted by CADTH considered it reasonable to generalize the results to this patient population. Although the PIK3CA mutant cohort had 1 male patient and the results were not considered to be generalizable to male patients, the clinical experts would not exclude male patients from treatment with alpelisib and fulvestrant in practice.

Aside from the aforementioned issues, the baseline characteristics, including ECOG PS, sites of metastases, and endocrine resistance status (primary resistance versus secondary resistance), appeared to be in line with patients treated in the Canadian setting, according to the clinical experts consulted by CADTH, although the clinical experts noted that patients with an ECOG PS of 2 may be considered for treatment with alpelisib plus fulvestrant in practice.

According to the study protocol, patients who discontinued treatment with 1 study drug (alpelisib/placebo or fulvestrant) could continue to receive the other study drug. The clinical experts consulted by CADTH for this review indicated that alpelisib monotherapy would not be an option upon discontinuation of fulvestrant.

The frequency of assessments of disease progression and patient follow-up in the study were similar to what would be found in Canadian clinical practice, according to the clinical experts consulted by CADTH, though in practice most patients would have imaging assessments every 12 weeks from the start.

Indirect Evidence

CADTH performed a literature review to identify any relevant indirect comparisons. A focused literature search for network meta-analyses dealing with breast cancer was run in MEDLINE All (1946–) on May 19, 2021. No limits were applied to the search. The search yielded 253 results whose abstracts were reviewed for relevance. Of those, 3 were identified as potentially relevant and, upon full-text review, were deemed not relevant because they did not perform any analyses restricted to the patient population of interest.^24-26 An additional result²⁷ from a clinical literature search alert was excluded upon full-text review for the same reason. Therefore, no indirect evidence was available for this review.

Other Relevant Evidence

This section includes 2 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. The patient population in the SOLAR-1 study did not align with the patient population relevant to this review, those with prior CDK4/6 inhibitor treatment, and conclusions on the comparative efficacy of alpelisib and fulvestrant versus fulvestrant alone could not be drawn in that population. The BYLieve study, a non-comparative cohort study, included 1 cohort of patients treated with alpelisib and fulvestrant that matched the relevant patient population. In a separate observational study, the relevant cohort of the BYLieve study was compared with a database-derived cohort treated with non-alpelisib SOC following propensity score weighting.

Non-Comparative Cohort Study

One sponsor-submitted, non-comparative cohort study is summarized to provide additional evidence regarding the efficacy and safety of alpelisib plus fulvestrant in patients with PIK3CA-mutated, hormone receptor–positive, HER2-negative advanced breast cancer who have been previously treated with a CDK4/6 inhibitor in combination with an endocrine-based therapy.

Methods

The BYLieve study is an ongoing phase II, multi-centre, open-label, 3-cohort, non-comparative study. The objective of the study, as stated in the BYLieve Clinical Study Report, is to assess the efficacy and safety of alpelisib plus endocrine therapy (either fulvestrant or letrozole) in premenopausal and postmenopausal women, and men, with PIK3CA-mutant, hormone receptor–positive, HER2-negative advanced breast cancer who have progressed on or after prior treatments. Gonadal suppression had to be achieved with either goserelin or leuprolide in men and premenopausal women. Patients were assigned to a cohort based on most recent previous therapy (AI or fulvestrant) as indicated in the following:

• Cohort A: Patients who received any CDK4/6 inhibitor plus any AI as immediate prior treatment will receive alpelisib plus fulvestrant.

• Cohort B: Patients who received any CDK4/6 inhibitor plus fulvestrant as immediate prior treatment will receive alpelisib plus letrozole.

• Cohort C: Patients who received systemic chemotherapy or endocrine therapy (as monotherapy or in combination with targeted treatment, except CDK4/6 inhibitor plus AI) as immediate prior treatment will receive alpelisib plus fulvestrant. In this cohort, endocrine therapy included letrozole, fulvestrant, and CDK4/6 inhibitor plus fulvestrant.

The BYLieve study planned to enrol 112 patients in each cohort (total N = 336). As of the data cut-off date, a total of 209 patients were enrolled: 127 patients in cohort A, 81 patients in cohort B, and 1 patient in cohort C. The BYLieve study enrolled patients from 114 multinational study locations (including 3 sites in Canada).

The overall design of the BYLieve study is summarized in Figure 4. The BYLieve study included a pre-screening phase and a screening phase. In the pre-screening phase, patients were tested for eligible PIK3CA mutations. Patients who tested positive for the PIK3CA mutation were then offered the opportunity to sign the main study informed consent form to enter the screening phase. The pre-screening and screening phases were to be completed within 21 days. In the treatment phase, patients were treated until disease progression, intolerable toxicity, or until 18 months after the last enrolled patient’s first treatment. Treatment crossover between cohorts was not permitted. Patients who discontinued study treatment due to disease progression, death, loss to follow-up, or withdrawal of consent for efficacy follow-up were not followed for efficacy, though those who discontinued due to disease progression were followed for safety for 30 days. Patients who discontinued study treatment for reasons other than disease progression, death, loss to follow-up, or withdrawal of consent for efficacy follow-up entered the post-treatment follow-up phase, which included 30 days of safety follow-up and efficacy follow-up. For post-treatment efficacy follow-up, tumour response was assessed locally via CT or MRI per RECIST 1.1 every 12 weeks until disease progression, death, withdrawal of consent, loss to follow-up, a patient or guardian decision, or the end of the study. For the survival follow-up period following end-of-treatment follow-up, all patients were followed for survival status after progression every 12 weeks regardless of treatment discontinuation reason.

Figure 4: BYLieve Study Design — Overview

^a All patients were followed for survival status (after progression) every 12 weeks regardless of the treatment discontinuation reason (except in cases of consent being withdrawn, death, or a patient being lost to follow-up) until death, loss to follow-up, or withdrawal of consent for survival follow-up or the end of the study.

Source: BYLieve Clinical Study Report.³²

The primary outcome of the BYLieve study is the proportion of patients who were alive without disease progression at 6 months based on local investigator assessment using RECIST 1.1 in each cohort. Secondary outcomes included OS and PFS by local investigator assessment.

This report includes the primary end point, OS, PFS, and harms data for cohort A alone because this is the only cohort that aligns with the intervention under review and the reimbursement criteria requested by the sponsor. Cohort B does not align with the intervention under review because patients received alpelisib in combination with letrozole. Cohort C does not align with the reimbursement request because patients were not previously treated with an endocrine-based regimen with a CDK4/6 inhibitor.

Inclusion and Exclusion Criteria

The BYLieve study included men and premenopausal or postmenopausal women (≥ 18 years of age) with advanced (locoregionally recurrent or metastatic) hormone receptor–positive and HER2-negative breast cancer not amenable to curative therapy and a confirmed PIK3CA mutation determined by local or central laboratory testing of tumour tissue or plasma. The inclusion of premenopausal women who were receiving concomitant ovarian suppression with an LHRH agonist was allowed in the first protocol amendment. Patients were required to have measurable disease per RECIST 1.1 criteria or at least 1 predominantly lytic bone lesion present. To be enrolled in cohort A, patients must have had documented tumour progression on or after treatment with a CDK4/6 inhibitor in combination with an AI (the last treatment regimen before study entry), an ECOG PS of 2 or less, fasting plasma glucose of 140 mg/dL (7.7 mmol/L) or less, and glycated hemoglobin of 6.4% or less. Patients were excluded if they had an established diagnosis of diabetes mellitus type 1 or uncontrolled type 2 diabetes. Patients could have 2 or fewer previous anticancer therapies and no more than 1 previous chemotherapy regimen in the advanced or metastatic setting. Patients with CNS involvement were excluded unless they met the following criteria: they were at least 4 weeks from prior therapy completion (including radiation and/or surgery) to starting the study treatment; they had a clinically stable CNS tumour at the time of screening, untreated or without evidence of progression for at least 4 weeks after treatment as determined by clinical examination and brain imaging (MRI or CT) during the screening period; and they were on a stable low dose of steroids for 2 weeks before initiating study treatment.

Baseline Characteristics

The baseline characteristics of patients enrolled in cohort A of the BYLieve study are summarized in Table 18. The mean age of patients was 56.7 years. All patients (100%) were female. Most patients were postmenopausal (78.0%), were Caucasian (63.8%), had stage IV disease at study entry (97.6%), had invasive ductal carcinoma (68.5%), had secondary endocrine resistance (59.8%), and had an ECOG PS of 0 (62.2%). The mean time from most recent recurrence and/or relapse was 2.2 months.

Interventions

Patients were assigned to receive alpelisib plus fulvestrant (cohort A) if their immediate prior treatment was a CDK4/6 inhibitor plus any AI. Patients received 300 mg alpelisib orally once per day and 500 mg fulvestrant intramuscularly on day 1 of each 28-day cycle and on day 15 of cycle 1. Treatment continued until disease progression, unacceptable toxicity, death, or discontinuation from study treatment due to any other reason. For patients unable to tolerate alpelisib due to AEs, a maximum of 2 dose reductions of alpelisib were allowed (the first dose reduction was to 250 mg per day while the second dose reduction was to 200 mg per day). Men and premenopausal women also received goserelin (an injectable subcutaneous implant, 3.6 mg, every 28 days) or leuprolide (injectable IM depot, 7.5 mg, every 28 days) as study treatments.

Concomitant medications required to treat AEs, manage cancer symptoms, and treat concurrent diseases were allowed, as were supportive care agents (e.g., pain medications, antiemetics, antidiarrheals). Oral antidiabetics were permitted to treat patients who developed hyperglycemia during the study. Gastric protection agents, corticosteroids, and palliative radiotherapy (i.e., local radiotherapy for analgesic purposes or for lytic lesions at risk of fracture) were also permitted. Anticoagulants and bisphosphonates could be used with caution during study treatment. Stronger inducers of CYP3A4 were prohibited.

Outcomes

The outcomes identified in the CADTH systematic review protocol that were included in the BYLieve study design are summarized in Table 19.

The primary end point of the BYLieve study was defined as the proportion of patients who were alive without disease progression at 6 months by local investigator assessment using RECIST 1.1 criteria. In the original study protocol, tumour response was assessed locally via CT or MRI per RECIST 1.1 at screening, every 8 weeks in the first 6 months, and every 12 weeks thereafter.²⁸ After approval of a protocol amendment (January 30, 2019), assessments occurred every 12 weeks per SOC throughout the study until disease progression, death, withdrawal of consent, loss to follow-up, a patient or guardian decision, or the end of the study. If there was clinical suspicion of disease progression, physical examination and imaging assessment were performed promptly.

Safety was evaluated by collecting data on AEs at every visit as well as by physical examination, vital signs, performance status evaluation, electrocardiogram, cardiac imaging, and laboratory evaluations for hematology and biochemistry, including glucose monitoring.

Statistical Analysis

In the BYLieve study protocol, it was planned that the primary analysis for each cohort would be performed 6 months after the last patient enrolled in the cohort had started treatment. Two interim analyses were added as protocol amendments. The first interim analysis was planned to be performed after at least 20 patients receiving alpelisib plus fulvestrant (cohort A) had received at least 6 months of follow-up. The second interim analysis was planned to be performed when approximately 170 patients (50% of enrolled patients) had been treated in the study (regardless of cohort) and had received at least 6 months of follow-up. At the time of the interim analyses, preliminary disease progression, survival, and safety data were analyzed in a descriptive fashion. The primary end point was not analyzed, and no statistical hypothesis was tested. Another protocol amendment allowed analyses of the primary end point for each cohort independent of the other cohorts. Following this update, the primary analysis for each cohort was planned to be performed 6 months after the start of treatment by the last patient. The final analysis for cohort A with testing of the statistical hypothesis for the primary end point has been completed.

Analysis Sets

The FAS was defined as all patients to whom study treatment was assigned and who had received at least 1 dose of study treatment. Patients were analyzed according to the treatment to which they were assigned. The FAS was used for all baseline and demographic summaries, patient disposition, and biomarker analysis.

The mFAS included all patients of the FAS population who had PIK3CA mutation confirmed by a Novartis-designated laboratory; patients were analyzed according to the treatment to which they were assigned. The mFAS was designated as the primary population for the analysis of progression and survival end points.

The PP set included a subset of the patients in mFAS who were compliant with the protocol requirements. The PP set was used for sensitivity analyses of the primary end point.

The safety analysis set included patients who had received at least 1 dose of any component of the study treatment. Patients were analyzed according to the study treatment received.

Primary End Point Analysis

The proportion of patients alive without disease progression after 6 months was presented with a 95% CI that was determined using the Clopper and Pearson exact method.³⁰ Per the statistical analysis plan, the null hypothesis would be rejected if the lower bound of the 95% CI was greater than 30%.

The study protocol defined 6 months as 24 weeks, plus or minus 1 week; therefore, tumour assessments between week 23 and week 25 were considered for the primary analysis. For the primary end point analysis, patients who progressed, died, or discontinued the study before 6 months were counted as a “failure.”

The mFAS is the primary analysis population for the primary end point. It was planned that analysis of the primary end point would also be performed on the FAS and possibly be repeated on the PP set as supportive analyses. As per protocol, the primary analysis for each cohort was to be performed 6 months after the last patient had started treatment or discontinued early in the respective cohort. The primary end point was planned to be analyzed only at the final analyses of cohort A; hence, the sponsor determined that the end point did not require adjustment for multiplicity.

Sample Size Calculation

The sample size was based on an exact binomial test for a single proportion to test the null hypothesis that the proportion of patients who are alive without progression at 6 months is less than or equal to 0.30. With a 1-sided 2.5% level of significance (2-sided 95% CI), a total sample size of 112 patients in each cohort was required to have a power of at least 90% when the true proportion of patients who are alive without progression at 6 months is greater than or equal to 0.45.

Secondary End Point Analysis

Overall Survival

OS was estimated using the Kaplan-Meier method. The median OS, the OS rate at 6 months, and the OS rate at 12 months with 95% CIs were reported. If a patient was not known to have died, OS was censored at the last known date that the patient was alive.

Progression-Free Survival

PFS was estimated using the Kaplan-Meier method. The median PFS, the PFS rate at 6 months, and the PFS rate at 12 months with 95% CIs were reported. If no PFS event was observed before the cut-off date, PFS was censored. The censoring date was the date of the last adequate tumour assessment before the data cut-off date. If a PFS event was observed after 2 or more missing or non-adequate tumour assessments, then PFS was censored at the date of the last adequate tumour assessment. If a PFS event was observed after a single missing or non-adequate tumour assessment, the actual date of the event was used.

Patient Disposition

A total of 127 patients were enrolled into cohort A of the BYLieve study and 100% of those patients were treated. As of the data cut-off date, 26.0% of patients had ongoing study treatment and 74.0% had discontinued treatment. The most common reasons for discontinuing study treatment were progressive disease (50.4%) and AEs (14.2%).

A total of 82 (64.6%) patients discontinued study treatment but did not enter the post-treatment efficacy follow-up phase of the study. These patients did not enter the efficacy follow-up phase per the study protocol because they discontinued study treatment due to documented disease progression, death, loss to follow-up, or withdrawal of consent.

Protocol Deviations

Overall, 45.7% of patients in cohort A had protocol deviations. The most frequently reported protocol deviations were use of a prohibited concomitant medication (18.1%) and study treatment deviation (13.4%). The most common study treatment deviation was fulvestrant dose 500 mg not administered (10.2%).

Exposure to study treatments in cohort A of the BYLieve study is summarized in Table 22. The median duration of exposure to alpelisib was 5.1 months (N = 127); the median duration of exposure to fulvestrant was 6.5 months (N = 126). The median (interquartile range) relative dose intensity for alpelisib was 89.9% (75.0% to 100%).

PP premenopausal women were eligible to participate in the BYLieve study if they were receiving concomitant ovarian suppression with a LHRH agonist. A total of 16 patients were exposed to goserelin and the median duration of exposure was 8.3 months (range = 1 month to 17 months). Two patients were exposed to leuprolide: 1 for less than 2 months and 1 for 7 months.

Dose reductions and interruptions in cohort A of the BYLieve study are summarized in Table 23. Per the study protocol, a maximum of 2 dose reductions of alpelisib (first dose reduction to 250 mg per day; second dose reduction to 200 mg per day) were allowed for each patient, after which the patient was discontinued. Overall, 51.2% of patients required at least 1 dose reduction during treatment with alpelisib. The most common reason for a dose reduction was an AE (37.8%) and the median duration of dose reduction was 35 days. Overall, 63.8% of patients required at least 1 dose interruption during treatment with alpelisib. The most common reason for dose interruption was an AE (59.8%).

Progression and Survival Outcomes

Progression and survival outcomes for cohort A of the BYLieve study (alpelisib plus fulvestrant in patients who were previously treated with any CDK4/6 inhibitor plus any AI) are summarized in Table 24. At the time of the data cut-off date (December 17, 2019), the median duration of follow-up in cohort A was 11.7 months.

The CADTH review protocol identified PFS, OS, and HRQoL as the main efficacy outcomes of interest. Results for these, along with the primary outcome used in the BYLieve study, are the focus of this summary.

Primary Outcome

As of the data cut-off date, 61 of 121 (50.4%) patients in the mFAS were alive without progressive disease per investigator assessment at 6 months (95% CI, 41.2% to 59.6%). The study met the primary objective for cohort A because the lower bound of the 95% CI was greater than 30%.

A supportive analysis was done in the PP set, which showed 56 of 106 (52.8%) patients were alive without progressive disease at 6 months.

Progression-Free Survival

A total of 72 (59.5%) PFS events were observed as of the data cut-off date, comprising 66 (54.5%) progressive disease events and 6 (5.0%) deaths. The Kaplan-Meier plot of PFS by investigator assessment is depicted in Figure 5. The median PFS by investigator assessment was 7.3 months (95% CI, 5.6 months to 8.3 months). The PFS rates by investigator assessment at 6 months and 12 months were 54.1% (95% CI, 44.3% to 62.9%) and 27.3% (95% CI, 17.6% to 37.8%), respectively.

Overall, 40.5% of patients were censored in the analysis of PFS. The most common reason for censoring was that the follow-up of the patient was ongoing without an event (31.4%). No patients were censored due to loss to follow-up.

Figure 5: Kaplan-Meier Plot of Progression-Free Survival by Local Investigator Assessment — Cohort A of BYLieve Study, mFAS

No. = number.

Source: BYLieve Clinical Study Report.³²

Overall Survival

As of the data cut-off date, 25 (20.7%) deaths were reported in the mFAS. The Kaplan-Meier plot of OS is depicted in Figure 6. The median OS was 17.3 months (95% CI, 17.2 months to 20.7 months). The OS rates at 6 months and 12 months were 91.9% (95% CI, 84.9% to 95.7%) and 75.2% (95% CI, 62.5% to 84.2%), respectively.

Overall, 79.3% of patients were censored in the OS analysis. Of these, 55.4% were alive and 24.0% were lost to follow-up. The sponsor indicated in the BYLieve Clinical Study Report that OS data should be interpreted with caution due to the proportion of patients alive and ongoing follow-up at the time of the data cut-off date.

Figure 6: Kaplan-Meier Plot of Overall Survival — Cohort A of BYLieve Study, mFAS

mFAS = modified full analysis set; no. = number.

Source: BYLieve Clinical Study Report.³²

Health-Related Quality of Life

Data on HRQoL were not included in the BYLieve study.

Harms

Only those harms identified in the CADTH review protocol are reported as follows. See Table 25 for detailed harms data for cohort A of the BYLieve study.

Adverse Events

Almost all patients (99.2%) experienced at least 1 treatment-emergent AE. The most common AEs (≥ 20%) were diarrhea (59.8%), hyperglycemia (58.3%), nausea (45.7%), fatigue (29.1%), decreased appetite (28.3%), rash (28.3%), stomatitis (26.8%), and vomiting (23.6%).

A total of 82 (64.6%) patients experienced an AE leading to dose adjustment or interruption. The most common AEs requiring a dose adjustment or interruption were hyperglycemia (29.1%), rash (12.6%), maculopapular rash (9.4%), and diarrhea (7.9%).

The majority of patients (94.5%) experienced an AE requiring additional therapy. The most common AEs requiring additional therapy were hyperglycemia (43.3%), diarrhea (26.0%), nausea (19.7%), rash (19.7%), stomatitis (18.1%), and maculopapular rash (11.8%).

Serious Adverse Events

Overall, 26.0% of patients experienced an SAE. The most common SAEs were hyperglycemia (5.5%), maculopapular rash (3.1%), dyspnea (2.4%), pleural effusion (2.4%), abdominal pain (1.6%), and hematemesis (1.6%).

Withdrawals Due to Adverse Events

The most common AEs leading to discontinuation of study treatment were rash (3.9%) and colitis, hyperglycemia, urticaria, and vomiting (1.6% each).

Deaths

As of the data cut-off date, 25 (20.7%) patients had died. Seven (5.5%) patients had died during study treatment or within 30 days of the last dose of study drug, and 4 of these on-treatment deaths were attributed to breast cancer.

Notable Harms

Overall, 74 (58.3%) patients experienced hyperglycemia. Thirteen (10.2%) patients experienced hypersensitivity and anaphylactic reactions. Diarrhea, nausea, and vomiting were reported in 59.8%, 45.7%, and 23.6% of patients, respectively. Pneumonitis and severe cutaneous skin reactions were experienced by 1 (0.8%) patient each. Overall, 36 (28.3%) patients experienced rash and 18 (14.2%) patients experienced maculopapular rash.

Critical Appraisal

Internal Validity

The BYLieve study is a prospective, non-comparative cohort study. Patients were assigned to cohorts based on their most recently received anticancer treatment (i.e., patients were assigned to cohort A and thus received alpelisib plus fulvestrant if their most recent treatment was any CDK4/6 inhibitor plus any AI). Due to the absence of a comparator group, no conclusions can be made regarding the efficacy and safety of alpelisib plus fulvestrant in the post-CDK4/6 inhibitor setting compared to other treatment options based on the BYLieve study data.

The BYLieve study was open label and, thus, patients and investigators were aware of the treatment status. Open-label studies are prone to various biases (e.g., patient selection bias, reporting bias). Furthermore, disease progression was assessed by the local investigator for the primary outcome and PFS. The open-label administration of alpelisib in combination with fulvestrant may have biased the reporting of these end points. Progressive disease was determined by the investigator using RECIST 1.1, which is commonly used in clinical trials.

The BYLieve study did not include all enrolled patients (i.e., the FAS) in the primary analyses. The primary analysis population was the mFAS, which included all enrolled patients who had PIK3CA mutation confirmed by a Novartis-designated laboratory (n = 121, 95.3%).

For the primary end point analysis, patients who progressed, died, or discontinued the study before 6 months had passed were counted as failures. The primary analysis for cohort A was performed as planned (i.e., 6 months after the last patient had started treatment or discontinued early). Since the primary end point was tested for statistical significance in 1 analysis, adjustment for multiplicity was not required. At the planned analysis, the study met its primary end point (i.e., the lower bound of the 95% CI for the proportion of patients alive without progression after 6 months was > 30%). The sponsor considered a proportion of 30% to be a clinically meaningful threshold for the primary end point of the study, but the rationale for this threshold is unclear. This is not a commonly used end point, or 1 that was considered of particular importance by the patient group or the clinical experts consulted by CADTH. Thus, the validity of the primary end point is unknown.

The secondary end points in the BYLieve study — OS and PFS — were estimated using the Kaplan-Meier method. The sponsor indicated that the OS data should be interpreted with caution due to the proportion of patients alive and having ongoing follow-up at the time of the data cut-off date. CADTH reviewers agree that the data are difficult to interpret and as a result, no conclusions can be made regarding the effect of alpelisib in combination with fulvestrant on OS. Furthermore, 24.0% of patients were censored due to being lost to follow-up in the analysis of OS. This loss to follow-up could lead to significant bias in the estimation of OS. In the analysis of PFS, 0 patients were lost to follow-up. Again, due to the lack of any comparator, the results of the PFS or OS analyses do not provide evidence of the effects of alpelisib in combination with fulvestrant on these outcomes.

The median duration of follow-up at the primary analysis of cohort A was less than 1 year and most (64.6%) patients did not enter the post-treatment follow-up phase per the study protocol (i.e., patients only entered the post-treatment follow-up phase if they did not discontinue study treatment due to documented disease progression, death, loss to follow-up, or withdrawal of consent for efficacy follow-up).

External Validity

The BYLieve study was an international, multi-centre study that included 3 sites in Canada. The majority of patients enrolled in cohort A of the BYLieve study were postmenopausal women, were Caucasian, had stage IV disease at study entry, had invasive ductal carcinoma, had secondary endocrine resistance, and had an ECOG PS of 0. The mean age of patients was 56.7 years and the mean time from most recent recurrence and/or relapse was 2.2 months. The clinical experts consulted by CADTH indicated that the baseline characteristics of the study patients were generally consistent with the Canadian patient population. The BYLieve study only enrolled women; thus, the study does not provide evidence on the use of alpelisib in combination with fulvestrant in men. However, the clinical experts consulted by CADTH considered it reasonable to generalize the results to the male population.

For the primary outcome analysis, the sponsor considered a proportion of 30% who had not progressed, died, or discontinued at 6 months to be a clinically meaningful threshold for the primary end point of the study. The clinical experts consulted by CADTH indicated that they would expect fewer than 30% to experience disease progression or death at 6 months after starting the second-line treatments that are currently used in standard care (e.g., fulvestrant alone). As a result, the 30% threshold used in the BYLieve study may not be clinically relevant in the Canadian treatment setting.

The intervention used in cohort A of the BYLieve study aligns with the reimbursement request and intended use of alpelisib in Canadian patients, per the clinical experts consulted by CADTH.

The patient groups that provided input for this review identified OS and PFS as important outcomes, which were secondary end points in the BYLieve study. Patients also identified HRQoL as an important outcome, but this was not captured in the BYLieve study.

The median duration of follow-up was less than 1 year as of the data cut-off date. As a result, the BYLieve study does not provide long-term data on alpelisib in combination with fulvestrant.

Observational Study

The sponsor-submitted observational study, which has been published,³¹ compared 1 of the cohorts of the BYLieve study to a set of patients who were identified from electronic medical records as provided by the Flatiron Clinico-Genomic Database. The study was not randomized, and propensity score weighting was used in an attempt to adjust for confounding.

Methods

Objectives

The primary objective of the study was to evaluate and compare PFS among patients who received alpelisib and fulvestrant in the BYLieve trial (cohort A) to patients who received a non-alpelisib SOC therapy in the Flatiron Clinico-Genomic Database, all of whom had hormone receptor–positive, HER2-negative, advanced breast cancer harbouring a PIK3CA mutation and had disease progression during or after treatment with CDK4/6 inhibitors with endocrine therapy. The secondary objective was to evaluate and compare the rate of patients who remained progression-free at 6 months as measured by a milestone PFS analysis between the same groups of patients.

Patient Selection Criteria and Selection Process

Key selection criteria of the 2 cohorts included in the study are summarized in Table 26, in addition to the treatments and outcomes. The index date was defined as the date when the treatment under consideration began. The Flatiron data included data collected between January 1, 2011, and June 30, 2019, and the index date was required to be on or before January 1, 2019, to allow for at least 6 months of potential follow-up. Most of the inclusion and exclusion criteria that were used in the BYLieve study were not possible to apply in the Flatiron database because the necessary information was not available in electronic medical records. The following patient characteristics were available for both cohorts: sex, age, race (White versus non-White), ECOG PS, cancer stage at diagnosis, cancer stage at the index date, site of metastasis (bone lesion only, lung and/or liver), number of sites of metastases (< 3 and ≥ 3), number of prior lines of treatment in the advanced setting, duration between diagnosis and the index date, and the index SOC regimen.

The attrition table for patients in the Flatiron cohort is presented in Table 27.

Propensity Score Development

The covariates in Table 28, described in the sponsor’s report³³ as clinically relevant prognostic variables, summarize the patient characteristics that were used to balance the cohorts for effect estimation. In the primary analysis, these covariates were used as predictors in a logistic regression model with treatment with alpelisib plus fulvestrant as the binary outcome to estimate the conditional probability of receiving that treatment for each individual in the observational study.

Statistical Methods

Three methods of balancing were used: propensity score estimation³⁴ followed by weighting by the odds of treatment with alpelisib plus fulvestrant,³⁵ 1:1 greedy nearest-neighbour matching based on the same propensity scores using a caliper of 0.2 times the standard deviation of the logit of the propensity scores, and 1:1 exact matching based on exact matches of the covariates (not propensity score based³⁶). The first approach was considered the primary analysis while the latter 2 were considered sensitivity analyses. Missing data were not imputed.

Following weighting and/or matching, the median PFS, the PFS survival curves, and the 6-month probability of PFS were estimated using the Kaplan-Meier method with Hall-Wellner confidence bands.³⁷ The survival curves were compared using the log-rank test, and Cox regression was used to estimate hazard ratios. Bootstrapping was used in the weighted analysis with 200 replicates to estimate the standard errors.³⁸ It is unclear whether the weights were re-estimated at each bootstrap sample. Patients in the BYLieve cohort were given weights of 1 while patients in the Flatiron cohort were given weights of the odds of treatment propensity score divided by (1 minus propensity score), where propensity score is the propensity score of treatment. Balance was assessed by comparing standardized mean differences of the baseline variables.

Results

Patient Characteristics

Table 29 shows the characteristics of the cohorts before and after propensity score weighting. Patients in the BYLieve cohort received a weight of 1, so the study characteristics were the same before and after weighting. The clinical experts consulted by CADTH for the review agreed that before weighting, patients in the Flatiron cohort appeared to be older and sicker, on average, than those in the BYLieve cohort. The characteristics that were used in the propensity score development were balanced after weighting, but many other important measured characteristics were not balanced after weighting — notably, ECOG PS at baseline, and stage at initial diagnosis. In the Flatiron cohort, the SOC included a wide variety of treatments, the most common of which were capecitabine and fulvestrant.

Efficacy

Figure 7 shows the results of the primary analysis, the weighted Kaplan-Meier estimated survival curves of progression for the BYLieve cohort and the propensity score odds–weighted Flatiron cohort, along with the result of the adjusted log-rank test (P = 0.004).

Figure 7: Post-Weight Progression-Free Survival (Primary Analysis)

PFS = progression-free survival.

Source: Sponsor-submitted Observational Study Report (2020).³³

Table 30 shows the summary statistics corresponding to the primary survival analysis. The median survival estimated by the weighted Kaplan-Meier method in the Flatiron cohort was 3.7 months (95% CI, 3.1 months to 6.1 months) compared to 7.3 months (95% CI, 5.6 months to 8.3 months) in the BYLieve cohort. The estimated hazard ratio estimated by weighted Cox regression comparing the BYLieve cohort to the Flatiron cohort was 0.62 (95% CI, 0.44 to 0.85; P = 0.002) where the P value was calculated by the Wald test using bootstrap standard errors. Due to the weighting method (propensity score odds), these represent estimates of the average treatment effect on those treated under the assumptions that the propensity score model was correctly specified, that there were no confounders other than the ones specified in Table 28, and that those covariates were measured without error.³⁵

Sensitivity Analyses

The observational study included sensitivity analyses assessing the sensitivity of the results to the form of confounding adjustment — namely, greedy matching and exact matching. The results of those analyses were not meaningfully different from the primary analysis results. No sensitivity analysis to the assumption of no unmeasured confounding was done.

Critical Appraisal

Internal Validity

The sponsor submitted an observational study comparing patients exposed to alpelisib with patients exposed to non-alpelisib SOC using a cohort from the BYLieve study combined with data from the Flatiron health database. Differences between the cohorts were accounted for using a propensity score weighted analysis based on a set of clinically relevant prognostic variables. The main findings were not sensitive to the choice of adjustment method when compared to 2 types of matching methods. There are many weaknesses of this study that raise concerns regarding the reliability and validity of these estimates of efficacy between alpelisib and SOC. There was no apparent pre-specified and registered protocol for this study, and no power or sample size calculations were reported. Hence, the statistical inference obtained from this study has low reliability and validity.

The 2 cohorts are innately different, making them harder to compare and requiring confounding and selection bias adjustment. The BYLieve study was a multinational cohort study that recruited patients from all around the world, but the Flatiron database is based in the US. Data collection methods differ dramatically between carefully planned cohort studies and electronic health records that are not designed specifically for research purposes.

Confounding and selection bias adjustment is likely inadequate for several reasons. The set of covariates used for propensity score modelling was insufficient to control for known prognostic factors, including number and type of prior lines of therapy, ECOG PS, sex, and disease stage at index date. In the SOLAR-1 study, ECOG PS of 0 versus 1 was a highly significant prognostic variable, but was not included in the propensity score model due to the large amount of missing data in the Flatiron cohort.

There was a planned exclusion of patients from the Flatiron cohort who were treated with alpelisib. Due to this exclusion, the study treatment and the cohort definition were completely confounded. As such, it is unknown whether the differences observed in the study are attributable to the treatment or to the inherent differences in selection in the Flatiron database versus controlled clinical studies; this source of confounding cannot by accounted for by the weighting and/or matching approach. The patient disposition generally indicates that the Flatiron cohort is older and has more severe disease, on average. Hence, this source of confounding is a critical weakness of the study since it raises concerns for the validity of any causal inference from the study. No sensitivity analyses were done to address this limitation.

Finally, the justification provided by the sponsor for the exclusion of events that occurred within 14 days of starting treatment was given as follows: “This criterion was implemented based on Flatiron’s analytic guidance which suggests that progression or death observed within 14 days of treatment initiation may not be directly associated with the effectiveness of treatment.”³³ No data to support this guidance were provided, but the clinical experts consulted by CADTH thought that this was reasonable. Furthermore, it is unclear whether patients with progression events that occurred within 14 days were excluded or if those events were simply not counted. The BYLieve study was a multinational cohort study that recruited patients from all around the world, but the Flatiron database is exclusively based in the US. Data collection methods differ dramatically between carefully planned cohort studies and electronic health records that are not designed specifically for research purposes.

Overall, there remains a great deal of uncertainty regarding the efficacy of alpelisib in comparison to SOC, due to the inherent limitations of observational data. While the adjustment approaches taken in this study may have adequately balanced on observable prognostic factors categorized as they were, bias in the efficacy estimate due to selection bias, measurement error, unmeasured confounding, and residual confounding cannot be ruled out. No attempts were made to assess or estimate the possible magnitude of such bias.

External Validity

The external validity of the results should be the same as that of the BYLieve cohort in that the results are weighted to match that study. However, the same limitations in the ability to weight properly given the covariates used and the innate differences in the cohorts remain. As the Flatiron cohort includes only US patients, it is likely that practices and patients are different from the Canadian population and are, in general, sicker and receiving a lower SOC.

Discussion

Summary of Available Evidence

The CADTH systematic review identified 1 relevant study, the SOLAR-1 study. The SOLAR-1 study (N = 572) was a placebo-controlled, double-blind, parallel-group RCT that randomized patients 1:1 to alpelisib 300 mg daily or matching placebo, in combination with fulvestrant 500 mg on day 1, day 15, and day 29, and every 28 days afterwards. Men and postmenopausal women with hormone receptor–positive, HER2-negative advanced or metastatic breast cancer and previous endocrine therapy were randomized within each of 2 cohorts based on PIK3CA mutation status: the PIK3CA mutant cohort and PIK3CA non-mutant cohort. The primary and key secondary outcomes were PFS and OS in the PIK3CA mutant cohort (N = 341). Endocrine therapy with a CDK4/6 inhibitor was not a part of the SOC at the time the study was conducted (enrolment was from 2015 to 2017) and only 20 patients in the PIK3CA mutant cohort were included in the subgroup analyses of patients who had received prior CDK4/6 inhibitor treatment and therefore met the reimbursement criteria requested by the sponsor.

Included in the sponsor’s submission was the non-comparative BYLieve study and an observational study comparing cohort A in the BYLieve study with a real-world, database-derived cohort. The BYLieve study assigned patients to 1 of 3 cohorts based on their most recent anticancer therapy. Of the 3 cohorts, cohort A (N = 127) was identified by the CADTH review team as relevant to the review; it included patients with hormone receptor–positive, HER2-negative advanced or metastatic breast cancer and a confirmed PIK3CA mutation and who had received any CDK4/6 inhibitor plus any AI as their immediate prior treatment. These patients were assigned to receive alpelisib plus fulvestrant at the same dosages as in the SOLAR-1 study. The primary end point in the BYLieve study was the proportion of patients who were alive without disease progression at 6 months by local investigator assessment using RECIST 1.1 criteria. The outcomes of PFS and OS as well as safety data were also evaluated in the BYLieve study. The observational study compared cohort A from the BYLieve study to a database-derived cohort that received non-alpelisib SOC immediately following treatment with a CDK4/6 inhibitor and non-fulvestrant endocrine therapy. The outcome PFS was compared between the cohorts following weighting of the database-derived cohort based on propensity scores.

Interpretation of Results

Efficacy

No relevant conclusions can be drawn regarding PFS and OS in patients treated with alpelisib and fulvestrant versus placebo and fulvestrant because the SOLAR-1 study was not designed to test hypotheses in the subgroup of patients with prior CDK4/6 inhibitor treatment and the results of the analyses were not statistically significant. The only outcomes in the testing hierarchy were PFS and OS in the entire PIK3CA mutant cohort and the study sample size considerations were based on those outcomes. Analyses of efficacy in subgroups, of which the subgroup of patients with prior CDK4/6 inhibitor treatment was 1, was for the purpose of examining the consistency of any treatment effect found in the entire cohort across the subgroups rather than demonstrating efficacy in any of the subgroups. In the absence of pre-specified subgroup analyses within the statistical testing hierarchy, no conclusions can be drawn about efficacy in any of the subgroups. As well, different approaches for identifying patients with prior CDK4/6 inhibitor therapy resulted in different numbers of patients being identified as such within the PIK3CA mutant cohort; it is unclear whether all the patients in the stratum with prior CDK4/6 inhibitor treatment did, in fact, receive that treatment.

The clinical experts consulted by CADTH for this review agreed that the results from the entire PIK3CA mutant cohort cannot be generalized to the patient population under review — namely, patients who have received a prior endocrine regimen with a CDK4/6 inhibitor. The design of the SOLAR-1 study reveals that efficacy results were not expected to be consistent between patients with and without prior CDK4/6 inhibitor treatment. Randomization was stratified according to previous CDK4/6 inhibitor usage, with the following rationale given in the SOLAR-1 study interim Clinical Study Report: “Prior CDK4/6 inhibitor use was selected since treatment with a CDK4/6 inhibitor in combination with letrozole has shown a strong PFS advantage over single-agent letrozole; therefore, prior treatment with a CDK4/6 inhibitor may impact the outcome of subsequent treatment with fulvestrant and/or PI3K inhibitors compared to CDK4/6 inhibitor naïve subjects.”¹ Reviews of alpelisib from regulatory agencies did not contain any conclusions regarding the efficacy of alpelisib with fulvestrant in patients previously treated with a CDK4/6 inhibitor. Despite approving the use of alpelisib in patients with progression on or following an endocrine-based regimen, the FDA stated in its statistical and clinical evaluation of alpelisib that, based on the SOLAR-1 study, “no conclusions may be drawn regarding the efficacy of alpelisib plus fulvestrant in the CDK 4/6 inhibitor-pretreated population.”²⁰ The European Medicines Agency made a similar statement in its review, based on the SOLAR-1 and BYLieve studies, and restricted the approved indication to patients previously treated with endocrine therapy as a monotherapy.²¹

Of note, the sponsor has indicated that a new phase III trial will be conducted for alpelisib plus fulvestrant in patients with prior CDK4/6 inhibitor treatment. The trial will be a double-blind, placebo-controlled RCT in men and postmenopausal women with hormone receptor–positive, HER2-negative, PIK3CA-mutated advanced breast cancer who have progressed on or after treatment with an AI and a CDK4/6 inhibitor. As stated by the sponsor, the planned study population aligns with the population targeted by the reimbursement request. The first patient visit was expected to occur in October of 2021, with the first interpretable results for the trial planned for the third quarter of 2024.

While the BYLieve study did include an entire cohort of patients from the patient population under review, as with the SOLAR-1 study, it is also unable to inform the efficacy of alpelisib plus fulvestrant versus a relevant comparator (i.e., single-agent fulvestrant, single-agent chemotherapy, or everolimus plus exemestane) due to its non-comparative study design. Patients were assigned to 1 of 3 cohorts depending on the prior anticancer therapy they had received, as opposed to being randomized to treatment groups. Although the study met its primary end point of the lower bound of the 95% CI for proportion of patients alive without progression after 6 months being greater than 0.30, the clinical experts consulted by CADTH indicated that they would expect less than 30% of the patients to experience disease progression or death within 6 months after starting the second-line treatments that are currently used in standard care (e.g., fulvestrant alone). Therefore, the 30% threshold used in the BYLieve study may not be clinically significant in the Canadian treatment setting. Aside from the primary end point, there was no statistical hypothesis testing in the BYLieve study and no conclusions could be drawn regarding the outcomes of interest in the CADTH systematic review protocol, PFS and OS.

The sponsor-submitted observational study comparing cohort A from the BYLieve study to the Flatiron database-derived, real-world cohort had several key limitations that are of concern for drawing conclusions about the comparative efficacy of alpelisib plus fulvestrant versus any relevant comparator. Following propensity score–based weighting of the cohorts, PFS was 3.7 months (95% CI, 3.1 months to 6.1 months) in the Flatiron cohort and 7.3 months (95% CI, 5.6 months to 8.3 months) in the BYLieve cohort. The hazard ratio for disease progression or death for the BYLieve cohort versus the Flatiron cohort was 0.62 (95% CI, 0.44 to 0.85). There remains a great deal of uncertainty regarding the efficacy of alpelisib in comparison to SOC due to the inherent limitations of observational data. While the adjustment approaches taken in this study may have adequately balanced on observable prognostic factors categorized as they were, bias in the efficacy estimate due to selection bias, measurement error, unmeasured confounding, and residual confounding cannot be ruled out. No attempts were made to assess or estimate the possible magnitude of such bias. Given these limitations, the comparative efficacy of alpelisib plus fulvestrant versus the SOC identified in the Flatiron cohort remains unclear.

Input from patient groups and clinicians emphasized that, in addition to OS and PFS, quality of life is a very important consideration for patients. Outcomes assessing HRQoL were not analyzed for the subgroup in the SOLAR-1 study and were not included in the BYLieve study. Even if analyses of HRQoL outcomes for the subgroup of patients with prior CDK4/6 inhibitor treatment in the SOLAR-1 study were available, they would be subject to the same limitations as PFS and OS in that subgroup. In addition, potential treatment unblinding due to the obvious differences in AE profiles between treatment groups could have led to bias in the EQ-5D-5L and EORTC QLQ-C30 outcomes.

Harms

The safety profile of alpelisib, as informed by the entire PIK3CA cohort in the SOLAR-1 study (corresponding with the Health Canada–approved indication for alpelisib) and cohort A in the BYLieve study, was consistent between studies. In the SOLAR-1 study, the most common AEs (those reported in at least 20% in either treatment group) were also more common in the alpelisib group versus the placebo group: hyperglycemia, diarrhea, nausea, rash, decreased appetite, decreased weight, stomatitis, vomiting, fatigue, and alopecia. The most commonly reported AEs in the BYLieve study in cohort A were similar, except for the absence of decreased weight and alopecia. According to the clinical experts consulted by CADTH for this review, all of the AEs identified in the clinical studies can be managed, aside from osteonecrosis of the jaw. The clinical experts identified hyperglycemia, rash, and diarrhea as the main AEs and indicated that rash can be prevented with prophylactic antihistamine, while hyperglycemia and diarrhea are more challenging to manage. However, the ability of patients to manage the AEs depends on their tolerance for side effects from additional treatments. Aside from osteonecrosis of the jaw, discontinuing treatment with alpelisib may be necessary due to medical reasons such as significant dehydration due to diarrhea or elevation of glucose levels to the point of requiring visits to the emergency department. The most common serious AE in both studies was hyperglycemia, with all other serious AEs occurring in less than 4% of any treatment group. Hyperglycemia, rash, and diarrhea were the most common reasons for discontinuing alpelisib due to an AE in the SOLAR-1 study, while rash was the most common reason in the BYLieve study. The percentage of patients who discontinued treatment due to an AE was notably higher in the alpelisib group than in the placebo group in the SOLAR-1 study; for the alpelisib groups, it was 27.2% in the SOLAR-1 study and 20.5% in the BYLieve study. The clinical experts consulted by CADTH did not have enough experience in treating patients with alpelisib to comment on which estimate was more reflective of clinical practice in Canada but indicated that both seemed reasonable. Safety outcomes were not assessed in the observational study.

According to input from patient groups, minimizing the risk of side effects was identified as being important to patients. The commonly reported AEs from the SOLAR-1 and BYLieve studies were in line with those reported by patients with experience with alpelisib treatment. The patient groups indicated that patients who had taken alpelisib experienced substantial difficulties with the side effects, including dealing with intense fatigue or being overwhelmed by the number of side effects early on. Hyperglycemia was commonly identified as being particularly hard to manage. Despite these difficulties, the patient group submissions also made it clear that overall, patients found the side effects manageable and were willing to tolerate them for the sake of a treatment that would improve long-term health outcomes.

Conclusions

No conclusions could be drawn from the SOLAR-1 and BYLieve studies regarding the comparative efficacy or effectiveness of alpelisib plus fulvestrant versus any relevant comparator in patients with hormone receptor–positive, HER2-negative, PIK3CA-mutated advanced or metastatic breast cancer after disease progression following an endocrine-based regimen with a CDK4/6 inhibitor. Neither study was designed to draw conclusions on comparative efficacy in this patient population. The sponsor-submitted observational study comparing a cohort from the BYLieve study that received alpelisib plus fulvestrant to a database-derived cohort that received a variety of non-alpelisib SOC therapies reported PFS results in favour of alpelisib. Methodological limitations in the observational study, including a high likelihood of residual confounding that may have led to bias in favour of alpelisib and differences in the methods used to determine PFS, contributed a substantial degree of uncertainty to the effect estimate. Considering the evidence in its entirety, the magnitude of any benefit associated with alpelisib plus fulvestrant in the relevant patient population remains unclear. In the SOLAR-1 and BYLieve studies, alpelisib treatment was associated with hyperglycemia, diarrhea, nausea, rash, decreased appetite, stomatitis, vomiting, and fatigue. Although the AEs reported in the studies can be medically managed, the percentages of patients who discontinued treatment due to an AE in the studies suggest that a large proportion of patients may not be able to remain on treatment with alpelisib due to AEs and/or subsequent side effects from treatments to manage these.

References

1.Interim Clinical Study Report: CBYL719C2301. A phase III randomized double-blind, placebo controlled study of alpelisib in combination with fulvestrant for men and postmenopausal women with hormone receptor positive, HER2-negative advanced breast cancer which progressed on or after aromatase inhibitor treatment [internal sponsor's report]. Basel (CH): Novartis; 2018.

2.Clinical Study Report: CBYL719C2301. SOLAR-1: A phase III randomized double-blind, placebo controlled study of alpelisib in combination with fulvestrant for men and postmenopausal women with hormone receptor positive, HER2-negative advanced breast cancer which progressed on or after aromatase inhibitor treatment. Clinical Study Report - final overall survival analysis and update of efficacy and safety data [internal additional sponsor's report]. Basel (CH): Novartis; 2020.

3.Canadian Cancer Statistics Advisory Committee. Canadian Cancer Statistics 2018. A special 2018 report on cancer incidence by stage. Toronto (ON): Canadian Cancer Society; 2018: www.cancer.ca/Canadian-Cancer-Statistics-2018-EN. Accessed 2021 Jun 11.

4.Mariotto AB, Etzioni R, Hurlbert M, Penberthy L, Mayer M. Estimation of the number of women living with metastatic breast cancer in the United States. Cancer Epidemiol Biomarkers Prev. 2017;26(6):809-815. PubMed

5.Brenner DR, Weir HK, Demers AA, et al. Projected estimates of cancer in Canada in 2020. Can Med Assoc J. 2020;192(9):E199-E205. PubMed

6.Seung SJ, Traore AN, Pourmirza B, Fathers KE, Coombes M, Jerzak KJ. A population-based analysis of breast cancer incidence and survival by subtype in Ontario women. Curr Oncol. 2020;27(2):e191-e198. PubMed

7.Howlader N, Cronin KA, Kurian AW, Andridge R. Differences in breast cancer survival by molecular subtypes in the United States. Cancer Epidemiol Biomarkers Prev. 2018;27(6):619-626. PubMed

8.Cardoso F, Paluch-Shimon S, Senkus E, et al. 5th ESO-ESMO international consensus guidelines for advanced breast cancer (ABC 5). Ann Oncol. 2020;31(12):1623-1649. PubMed

9.Mosele F, Stefanovska B, Lusque A, et al. Outcome and molecular landscape of patients with PIK3CA-mutated metastatic breast cancer. Ann Oncol. 2020;31(3):377-386. PubMed

10.PrPiqray® (alpelisib): 50 mg, 150 mg and 200 mg tablets [product monograph]. Dorval (QC): Novartis Pharmaceuticals Canada; 2020 Mar 11.

11.PrFaslodex® (fulvestrant): 50 mg/mL pre-filled syringe for intramuscular injection [product monograph]. Mississauga (ON): AstraZeneca Canada Inc.; 2019 Nov 15: https://pdf.hres.ca/dpd_pm/00053975.PDF. Accessed 2021 Jun 8.

12.PrXeloda® (capecitabine): 150 mg and 500 mg tablets [product monograph]. Mississauga (ON): Hoffmann-La Roche Limited; 2021 Apr 19: https://pdf.hres.ca/dpd_pm/00060777.PDF. Accessed 2021 Jun 8.

13.PrAfinitor® (everolimus): 2.5 mg, 5 mg and 10 mg, tablets [product monograph]. Dorval (QC): Novartis Pharmaceuticals Canada Inc.; 2021 May 5: https://pdf.hres.ca/dpd_pm/00060902.PDF. Accessed 2021 Jun 8.

14.PrAromasin® (exemestane): 25 mg tablets [product monograph]. Kirkland (QC): Pfizer Canada Inc.; 2018 Mar 6: https://pdf.hres.ca/dpd_pm/00044156.PDF. Accessed 2021 Jun 8.

15.Anderson EJ, Mollon LE, Dean JL, et al. A systematic review of the prevalence and diagnostic workup of PIK3CA mutations in HR+/HER2- metastatic breast cancer. Int J Breast Cancer. 2020:3759179. PubMed

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

17.Grey matters: a practical tool for searching health-related grey literature. Ottawa (ON): CADTH; 2019: https://www.cadth.ca/grey-matters. Accessed 2021 May 7.

18.Andre F, Ciruelos E, Rubovszky G, et al. Alpelisib for PIK3CA-mutated, hormone receptor-positive advanced breast cancer. N Engl J Med. 2019;380(20):1929-1940. PubMed

19.Andre F, Ciruelos EM, Juric D, et al. Alpelisib plus fulvestrant for PIK3CA-mutated, hormone receptor-positive, human epidermal growth factor receptor-2-negative advanced breast cancer: final overall survival results from SOLAR-1. Ann Oncol. 2021;32(2):208-217. PubMed

20.Center for Drug Evaluation Research. Multidiscipline review. Piqray (alpelisib) oral. Novartis Pharmaceuticals. Application No.:212526. Approval date: 05/24/2019 (FDA approval package). Silver Spring (MD): U.S. Food and Drug Administration (FDA); 2019: https://www.accessdata.fda.gov/drugsatfda_docs/nda/2019/212526Orig1s000MultidisciplineR.pdf. Accessed 2021 Jun 17.

21.Committee for Medicinal Products for Human Use. Assessment report: Piqray (alpelisib). (European public assessment report). Amsterdam (NL): European Medicines Agency; 2020 May 28: https://www.ema.europa.eu/documents/assessment-report/piqray-epar-public-assessment-report_en.pdf. Accessed 2021 Jun 17.

22.Novartis Pharmaceuticals Inc. response to 2021 June 23 DRR request for additional information regarding Piqray DRR review. Additional SOLAR-1 study data in subgroup with prior CDK4/6 inhibitor use [internal additional sponsor's information]. Basel (CH): Novartis Pharmaceuticals; 2021.

23.Eisenhauer EA, Therasse P, Bogaerts J, et al. New response evaluation criteria in solid tumours: revised RECIST guideline (version 1.1). Eur J Cancer. 2009;45(2):228-247. PubMed

24.Giuliano M, Schettini F, Rognoni C, et al. Endocrine treatment versus chemotherapy in postmenopausal women with hormone receptor-positive, HER2-negative, metastatic breast cancer: a systematic review and network meta-analysis. Lancet Oncol. 2019;20(10):1360-1369. PubMed

25.Han Y, Wang J, Wang Z, Xu B. Comparative efficacy and safety of CDK4/6 and PI3K/AKT/mTOR inhibitors in women with hormone receptor-positive, HER2-negative metastatic breast cancer: a systematic review and network meta-analysis. Curr Probl Cancer. 2020;44(6):100606. PubMed

26.Wang S, Liu M, Lian S, et al. Which is the most appropriate PI3K inhibitor for breast cancer patients with or without PIK3CA status mutant? A systematic review and network meta-analysis. Biomed Res Int. 2020:7451576. PubMed

27.Leung JH, Leung HW, Wang SY, Huang SS, Chan AL. Efficacy and safety of CDK4/6 and PI3K/AKT/mTOR inhibitors as second-line treatment in postmenopausal patients with hormone receptor-positive, HER-2-negative metastatic breast cancer: a network meta-analysis. Expert Opin Drug Saf. 2021;28. PubMed