CADTH Reimbursement Review

Selinexor (Xpovio)

Sponsor: FORUS Therapeutics Inc.

Therapeutic area: Multiple myeloma

This multi-part report includes:

Clinical Review

Pharmacoeconomic Review

Stakeholder Input

Clinical Review

Executive Summary

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

Introduction

Multiple myeloma (MM) is a plasma cell cancer caused by the growth of cancer cells in the bone marrow. In Canada, more than 3,000 new cases of MM are diagnosed annually, with slightly more cases occurring in men than women.¹ Multiple myeloma is generally incurable, with a median survival for patients of approximately 5 years, and during this time patients can receive 4 or more lines of therapy.² It is a heterogenous condition that typically affects older adults around the age of 65 years, and patient outcomes can depend on many factors, including disease stage, prognostic indicators, and early treatment of symptomatic disease to limit or avoid organ damage.¹ Patients may initially present with symptoms including bone pain, lytic lesions, anemia, fatigue, infections, weight loss, hypercalcemia, and renal dysfunction.¹ Patients may also have cytogenetic abnormalities that can influence the course of their disease, response to therapy, and overall prognosis.

The treatment landscape for MM has changed significantly in the recent past, with the emergence of new therapies in newly diagnosed and relapsed or refractory settings.² Treatment choices for patients depend on whether they are eligible for stem cell transplant. Most patients in Canadian clinical practice will receive a lenalidomide based regimen. At relapse, treatment for patients depends on age, comorbidities, and previous treatments. Most patients will receive a daratumumab containing regimen. Other treatment options as patients continue to progress can include regimens containing carfilzomib, pomalidomide, isatuximab, or belantamab; funding of these regimens varies across Canadian jurisdictions and, in some cases, treatments may only be available through special access programs.

Selinexor is to be administered in combination with bortezomib and dexamethasone. Selinexor is to be administered at a dosage of 100 mg orally once weekly on day 1 of each week. Bortezomib is to be administered at a dosage of 1.3 mg/m² by subcutaneous (SC) injection once weekly on day 1 of each week for 4 weeks followed by 1 week off. Dexamethasone is to be administered at a dosage of 20 mg orally twice weekly on days 1 and 2 of each week.³ Selinexor was submitted to CADTH before Notice of Compliance and was anticipated to be approved by Health Canada on June 2, 2022. Selinexor has not previously been reviewed by CADTH. The sponsor has requested reimbursement of selinexor as per the Health Canada indication.

The objective of this review is to perform a systematic review of the beneficial and harmful effects of selinexor (Xpovio) at 100 mg administered orally in combination with bortezomib and dexamethasone for the treatment of adult patients with MM who have received at least 1 prior therapy.

Stakeholder Perspectives

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

Patient Input

One patient group, Myeloma Canada, provided input on the combination of selinexor plus bortezomib plus dexamethasone (SVd) for the treatment of MM in adult patients. The patient group conducted an online survey that was distributed through email and social media and made available to patients and caregivers across Canada from December 2021 to January 2022. According to Myeloma Canada, patients considered it extremely important to control symptoms of infections, kidney problems, mobility, and neuropathy related to myeloma. Patients also indicated that symptoms significantly affected their abilities to travel, work, exercise, and concentrate. Parking costs, travel costs, drug costs, lost income due to absence from work, and lost income or pensions due to early retirement were described as the most significant financial implications of myeloma treatment. Patients receiving treatment with bortezomib and dexamethasone described fatigue, diarrhea, and nausea as “totally unbearable” side effects. “Tolerable” side effects were anemia and thrombocytopenia. Peripheral neuropathy (PN) was highlighted by many patients as a side effect and an important symptom to control and reduce in severity. Two respondents had experience with SVd through participation in the BOSTON trial; 1 respondent had not relapsed since receiving SVd through the BOSTON trial, while the other relapsed within 3 months and was receiving a different treatment. Nausea was stated to be a “somewhat tolerable” side effect, while diarrhea, PN, and vomiting were “somewhat intolerable.” Other reported side effects included thrombocytopenia, anemia, fatigue, decreased appetite, and weight decrease. One patient indicated the trial regimen was very effective in helping control their myeloma, while the other patient reported that it was somewhat ineffective. Respondents mentioned the following as being important considerations for new treatments: effectiveness of treatment, quality of life, accessibility and portability of treatment, manageable side effects, and access to a supportive and communicative care team. The side effects that patients most frequently ranked as important to avoid when considering new treatments included infections, vomiting, pain, confusion, decreased appetite, and neuropathy. Many patients indicated a preference for an orally administered treatment versus SC injection and infusion. Many respondents indicated that fewer trips to a cancer centre or hospital for treatment would improve their quality of life.

Clinician Input

Input From Clinical Experts Consulted by CADTH

The clinical experts emphasized the need for treatments that improve survival, provide deeper and longer-lasting remissions, and improve disease-related symptoms and complications, such as pain and renal failure. In addition, treatments would have less of a negative impact on patient quality of life and require fewer clinic visits. The clinical experts acknowledged that most patients will relapse with currently available therapies. Treatments are palliative and may prolong patients’ lives, but they do not provide patients with a cure, and patients will eventually become refractory to available treatments. Patients with high-risk cytogenetics and who are ineligible for transplant were stated to be at a particularly high risk of progression and poor outcomes. The experts noted that the side effects of some treatments affect tolerability to treatment and effectiveness. Many treatment options for patients are provided intravenously or subcutaneously at a cancer centre as often as once or twice per week, resulting in significant burdens for patients, caregivers, and treatment centres. Selinexor could be an attractive option for patients as it is administered orally and only once per week, potentially reducing the need for clinic visits. The clinical experts acknowledged that selinexor would not require a significant paradigm shift and that other therapies for MM are available. However, selinexor operates under a mechanism of action that is different to other currently available treatments and may be effective in patients who become resistant to treatments that target other pathways. The clinical experts agreed that other regimens would likely be preferred before using a selinexor based regimen. The toxicity profile of selinexor was also described as different from that of other classes of drugs. The clinical experts agreed that no specific features make a patient a better candidate for selinexor. Patients with pre-existing anorexia, weight loss, or nausea may not be good candidates for this treatment as the side effect profile of selinexor is associated with anorexia, weight loss, or nausea.

A patient’s response to therapy is typically measured through monoclonal protein and serum free light chains; based on these evaluations, a clinically meaningful response would include a sustained partial response (PR) or better. Stable disease may also be considered an acceptable benefit to patients in some cases. Improvements in cancer-related complications, such as anemia, renal failure, hypercalcemia, and tumour-related pain, are also considered when assessing patient responses. In general, meaningful responses to treatment are expected to manifest through improvements in patients’ overall survival (OS) and progression-free survival (PFS). Improvements in quality of life (QoL), myeloma-related symptoms, and treatment toxicity were also stated to be important outcomes when assessing patients’ responses to treatment. Typically, patients may be assessed for response every 4 weeks, although less-frequent monitoring may be warranted if patients demonstrate stable long-term response and predictable and manageable toxicity.

The clinical experts agreed that discontinuing treatment should depend on whether a patient’s disease progresses, which is determined when patients fail to respond to treatment and require a change in their therapy. Significant toxicities or adverse events (AEs) that are not manageable through supportive care or dose modifications were also stated to result in discontinuation of therapy. Both clinical experts acknowledged that administration of therapies requires a specialty hematology or oncology clinic or the equivalent. Physicians with expertise and experience in treating MM, such as hematologists or oncologists, would be treating and monitoring patients. The clinical experts highlighted that changes to the treatment paradigm for MM patients are likely to occur with approval of daratumumab in the first-line setting. Because most patients will likely receive daratumumab plus lenalidomide plus dexamethasone (DRd) as a first-line therapy, these patients will not receive daratumumab-based regimens upon relapse. The next-line option for patients will likely be a combination regimen of bortezomib and another proteosome inhibitor (PI), such as cyclophosphamide plus bortezomib plus dexamethasone (CyBorD) or carfilzomib plus dexamethasone(Kd). Although selinexor could be considered as a second-line option, it may be more likely for SVd to be used in later lines of therapy.

Clinician Group Input

Two groups provided clinician input on the review of SVd for the treatment of adult patients with MM: the Ontario Health–Cancer Care Ontario (OH-CCO) Hematology Drug Advisory Committee (DAC) prepared by 7 physicians and the Canadian Myeloma Research Group (CMRG) prepared by 13 physicians. Both groups generally agreed that improving OS, PFS, disease-related symptoms, and health-related quality of life (HRQoL) are important goals for ideal treatment. The OH-COO group indicated the greatest unmet need for patients currently exists after the second-line setting; patients who failed daratumumab in the second-line setting could have the option to use this regimen in the third line. CMRG described the need for new classes of anti-myeloma drugs to complement available treatments and improve patient convenience (e.g., with oral administration) and toxicity profiles. Both groups agreed that patients who have the greatest unmet need for a selinexor-based regimen are those with relapsed or refractory multiple myeloma (RRMM) who are refractory to immunomodulatory inhibitors, PIs, and anti-CD38 monoclonal antibodies. The CMRG also noted that patients with renal insufficiency and poor risk features (e.g., high-risk cytogenetics, extramedullary disease, or highly proliferative disease) have the greatest unmet need for this therapy.

Selinexor was stated to be currently available through special access programs and was acknowledged by the CMRG to differ from currently available therapies based on route of administration, side-effect profile, and supportive care needs. Therapies such as selinexor that differ from currently available treatment options were stated to be sought after by physicians in Canada after patients progress through funded options and are not yet candidates for palliative care. The OH-CCO DAC expressed uncertainty about the specific placement of SVd in the current treatment paradigm. However, both groups generally agreed that daratumumab- or isatuximab-based treatments would be preferred as second-line regimens before recommending SVd. Both groups agreed that this drug would not affect the treatment sequence employed in current practice.

Both clinician groups acknowledged that eligible patients would be identified by their treating physician or hematologist. The OH-CCO DAC did not specify criteria for patients least suited for treatment, although the CMRG indicated that newly diagnosed patients with MM would be least suitable for treatment with SVd. Both groups indicated that patient responses to treatment would be assessed using conventional myeloma response criteria. A clinically meaningful response to treatment in the setting of advanced disease was stated to include a reduction in measurable disease. Both groups agreed that a patient’s response to treatment would be assessed each cycle or approximately every month. Both groups agreed that discontinuation of treatment would be based on disease progression and toxicity. Both groups agreed that SVd would be administered in outpatient clinics, hematology clinics, and hospitals.

Drug Program Input

The drug programs identified the following jurisdictional implementation issues: relevant comparators, considerations for initiation of therapy, considerations for prescribing of therapy, the funding algorithm, care provision, and system and economic issues. The clinical experts consulted by CADTH weighed evidence from the key study submitted by the sponsor and clinical expertise to provide responses to the drug programs’ implementation questions (Table 4).

Clinical Evidence

Pivotal Studies and Protocol-Selected Studies

Description of Studies

One multi-centre, phase III, active-controlled, open-label study, BOSTON, was included in this CADTH review. The objective of the BOSTON trial was to compare the efficacy, HRQoL, and safety of SVd to that of bortezomib plus dexamethasone (Vd) in adult patients with RRMM who received 1 to 3 prior anti-MM regimens. Patients were randomized to receive SVd or Vd in a 1:1 ratio and stratified based on prior PI therapy (yes versus no) and number of prior anti-MM regimens (1 versus > 1). Inclusion criteria included adult patients with histologically confirmed MM and measurable disease according to International Myeloma Working Group (IMWG) guidelines who had received between 1 and 3 prior anti-MM regimens. Patients had to have documented evidence of progressive MM on or after their most recent regimen. Patients previously treated with bortezomib or another PI were eligible if certain criteria were met (Table 6). Patients must also have had an Eastern Cooperative Oncology Group Performance Status (ECOG PS) of 2 or lower. Exclusion criteria included previous exposure to selective inhibitor of nuclear export (SINE) compounds, including selinexor, previous malignancies requiring treatment, or evidence of recurring and uncontrolled comorbidities. Patients could not have PN higher than grade 2, or a PN of grade ≥ 2 or higher with pain at baseline, regardless of whether or not they were receiving medication. The primary end point of the BOSTON trial was PFS. Key secondary end points included overall response rate (ORR), incidence of PN events of grade 2 or higher, response rates of very good partial response (VGPR) or better based on independent review committee (IRC) assessment. Other secondary end points included OS, duration of response (DOR), time to next treatment (TTNT), time to response (TTR), and HRQoL.

In general, characteristics across both the SVd and Vd treatment groups were well balanced. The mean age of patients was 65 years (standard deviation [SD] = 9.56) in the SVd group and 67 years (SD = 9.35) in the Vd group, with most patients aged 51 to 64 years (36% in the SVd group versus 31% in the Vd group), 65 to 74 years (39% versus 41%, respectively), or 75 years or older (17% versus 23%, respectively). Fewer patients were between 18 to 50 years of age (8% versus 5%). A slightly greater proportion of males was enrolled in the trial (59% versus 56% in the SVd and Vd groups). Most patients were White (83% versus 80% in the SVd and Vd groups) and not Hispanic or Latino (88% versus 91%), had never smoked (73% versus 74%), and had an ECOG PS of 0 (35% versus 37%) or 1 (54% versus 55%), a mean creatinine clearance at baseline of greater than 60 mL/min (71% versus 66%), and a status of nonfrail at baseline (66% versus 69%). Approximately one-quarter of patients (25% in the SVd group versus 27% in the Vd group) had stage I disease at diagnosis, compared to one-third who were diagnosed with stage II (32% versus 27%, respectively), and one-third with stage III (29% versus 32%). More than half of all patients had a kappa light chain type of active myeloma at baseline (56% versus 61% in the SVd group and Vd group, precatively). The Revised International Staging System (R-ISS) stage at screening was stage I for 29% of patients in the SVd group versus 25% in the Vd group, stage II for 60% of patients in both groups, and 6% and 7%, respectively, for stage III. Approximately half of all patients had a high-risk chromosomal abnormality, with most being 1q21 (41% versus 34% in the SVd and Vd groups, respectively) compared to t(4;14) (11% versus 14%), del(17p)/p53 (11% versus 8%), or t (14;16) (4% versus 5%). The mean number of prior lines of anti-MM therapy was 1.7 in both treatment groups; 51% versus 48% of patients in the SVd and Vd groups, respectively, had 1 prior line of therapy, compared to 33% and 31% of patients with 2 prior lines of anti-MM therapy, and 16% versus 21% of patients with 3 prior lines of anti-MM therapy. Most patients had received prior PI therapy (76% in the SVd group versus 77% in the Vd group). Other treatments patients had been previously exposed included bortezomib (69% versus 70% in the SVd and Vd groups, respectively), lenalidomide (40% versus 37%), carfilzomib (10% in both groups), pomalidomide (6% versus 3%), daratumumab (6% versus 3%), and ixazomib (3% versus 1%). Slightly more patients in the SVd group received a stem cell transplant (39%) compared with the Vd group (30%).

Efficacy Results

Results from the BOSTON trial were reported for 2 data cut-offs. The primary analysis was a pre-specified interim analysis and was based on a data cut-off on February 18, 2020. In agreement with the Data and Safety Monitoring Board (DSMB), the results of this interim analysis were considered final, as the stopping boundary for PFS had been reached. The updated analysis was based on a data cut-off date of February 15, 2021; results of the updated analysis were considered descriptive. The results of the updated analysis were supportive of the primary analysis; the results are not described here but are reported in Table 2.

Table 2: Summary of Key Results From Pivotal and Protocol-Selected Studies

AE = adverse event; CI = confidence interval; DOR = duration of response; IRC = independent review committee; MM = multiple myeloma; NE = not evaluable; NR = not reported; PFS = progression-free survival; PR = partial response; SAE = serious adverse event; SVd = selinexor plus bortezomib plus dexamethasone; TTNT = time to next treatment; TTR = time to response; Vd = bortezomib plus dexamethasone.

Note: Overall survival is calculated from the date of randomization to the date of death. Patients without events were censored at the date of study discontinuation or date of last participating visit, whichever occurred first. DOR is defined for patients with a confirmed PR or better as the duration from the date of first IRC-confirmed PR or better to the date of first IRC-confirmed progressive disease, or death due to any cause, whichever occurred first.

^aStratified for prior proteosome inhibitor therapies, number of prior anti-MM regimens and Revised International Staging System stage at screening.

^bBased on a stratified Cox proportional hazards model with the Efron method of handling ties.

^cP value is considered nominal as results for the updated analysis were not pre-specified or controlled for multiplicity.

^dOverall response rate is the proportion of patients who achieve a partial response or better, before IRC-confirmed progressive disease (PD) or initiating a new MM treatment or crossover.

Source: BOSTON Clinical Study Report.⁴

Overall Survival

At the time of the primary analysis, results for OS were based on a median follow-up time of 17.28 months (95% confidence interval [CI], 16.56 to 19.27) in the SVd group and 17.51 (95% CI, 17.08 to 18.23) in the Vd group. Similar proportions of patients died in the SVd (24.1%) and Vd (30.0%) treatment groups. The median OS was not evaluable [NE] (95% CI, NE to NE) in the SVd group and 24.97 months (95% CI, 23.49 to NE) in the Vd group. The hazard ratio (HR) of death was 0.84 (95% CI, 0.57 to 1.23; 1-sided P = 0.1852, stratified log-rank test). At this point, 75 patients (36%) from the Vd group had crossed over to the SVdX (selinexor plus bortezomib plus dexamethasone after crossover) or SdX (selinexor plus dexamethasone after crossover) groups.

Progression-Free Survival

A median follow-up time of 13.17 months (95% CI, 10.64 to 15.34) was reported for the SVd group and 16.53 months (95% CI, 14.39 to 17.71) for the Vd group. At the primary analysis, a higher proportion of patients in the Vd group experienced a PFS event compared with patients in the SVd group (59.9% versus 41.0%, respectively). The median PFS was longer in the SVd group at 13.93 months (95% CI, 11.73 to NE) compared to 9.46 months (95% CI, 8.11 to 10.78) in the Vd group. An HR of 0.70 (95% CI, 0.53 to 0.93) was reported for PFS, indicating an increase in PFS of 4.47 months and a 30% reduction in risk of disease progression or death in the SVd group compared to the Vd group (1-sided P = 0.0075, stratified log-rank test).

Duration of Response

At the primary analysis, more patients in the SVd group had achieved a PR or better (76.4%) compared with the Vd group (62.3%). The median DOR was 20.27 months (95% CI, 12.55 to NE) in the SVd group compared to 12.88 months (95% CI, 9.26 to 15.77) in the Vd group.

Time to Next Treatment

There were fewer patients in the SVd groups with TTNT events (45.1%) versus the Vd group (65.2%). The median TTNT was longer in the SVd group, at 16.13 months (95% CI, 13.92 to NE), than in the Vd group, at 10.84 months (95% CI, 9.82 to 13.40). There was a longer median treatment-free interval for patients with new MM treatment in the SVd group at 28.0 days (range, 1 to 447) than in the Vd group at 14.0 days (range, 1 to 419).

Time to Treatment Discontinuation

There were no differences between the SVd and Vd treatment groups in patients discontinuing treatment (81.0% versus 82.6%, respectively). The median TTD in the SVd group was 7.10 months (95% CI, 6.44 to 8.54) and 7.95 months (95% CI, 6.80 to 9.23) in the Vd group.

Time to Response

A greater proportion of patients in the SVd group had an IRC-confirmed response of a PR or greater (76.4%) compared with the Vd group (62.3%). The median TTR was numerically shorter in the SVd group, at 1.41 months (95% CI, 1.35 to 1.51) than in the Vd group, at 1.61 months (95% CI, 1.51 to 2.14).

Overall Response Rate

At the primary analysis, 149 patients had an overall response rate (76.4%; 95% CI, 69.8 to 82.2) in the SVd group compared to 129 patients (62.3%; 95% CI, 55.3 to 68.9) in the Vd group. There were no differences in the best overall response of patients between the 2 treatment groups. Most patients achieved a PR (31.8% in the SVd group versus 30.0% in the Vd group), VGPR (27.7% versus 21.7%, respectively), or stable disease (12.8% versus 19.3%, respectively).

Rate of Very Good Partial Responses or Better Responses

At the primary analysis, a VGPR, complete response (CR), or stringent complete response (sCR) was observed in 87 (44.6%) of 195 patients from the SVd group and 67 (32.4%) of 207 patients from the Vd group (odds ratio [OR] = 1.6594; 95% CI, 1.0993 to 2.5049; P = 0.0082).

Health-Related Quality of Life

Patient-Reported Peripheral Neuropathy Measured by EORTC QLQ-CIPN20

Baseline scores for the sensory, motor, and autonomic neuropathy symptoms subscales were similar between the 2 treatment groups. Regarding the sensory and motor subscales, a greater proportion of patients in the Vd group had higher post-baseline scores that showed increases from baseline equal to or greater than 10, 20, 30, 40, or 50 compared with the SVd group, indicating worse symptoms for patients in the Vd group. Regarding the autonomic subscale, a greater proportion of patients in the SVd group had higher post-baseline scores that showed increases from baseline equal to or greater than 10, 20, 30, 40, or 50 compared with the Vd group, indicating worse symptoms for patients in the SVd group. Linear mixed-effect models were also applied to scores on the European Organisation for Research and Treatment of Cancer 20-item Quality of Life Questionnaire for Chemotherapy-Induced Peripheral Neuropathy (EORTC QLQ-CIPN20); a lower mean change from baseline was observed in the SVd group compared to the Vd group for the sensory, motor symptoms, and autonomic subscale, indicating a lower symptom burden in the SVd treatment group. The results of the autonomic symptom score were broken down to its 3 components of blurred vision, difficulty with erection, and dizzy when standing up. The SVd and Vd groups showed similar scores in the dizziness and erectile-function components. The SVd group showed higher scores for blurred vision compared with the Vd group, indicating greater symptom burden. There were no statistically significant differences between the SVd and Vd groups on any of the subscales.

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

The European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30 (EORTC QLQ-C30) questionnaires were completed at baseline and at least 1 post-baseline time point by 188 patients in the SVd group and 195 patients in the Vd group. The mean baseline scores of patients were similar between the SVd and Vd group for the global health status (GHS) and QoL measure. There were no differences in GHS scores over time between the SVd and Vd groups. There were no statistically significant differences in the domains of the EORTC QLQ-C30 between SVd and Vd treatment groups.

EQ-5D 5-Levels

Baseline scores of patients in the SVd and Vd groups were similar for the EQ-5D 5-Levels (EQ-5D-5L) Visual Analogue Scale (VAS) and there were no differences between treatment groups throughout the trial. No major differences were observed for any other symptom domains.

Harms Results

Adverse Events

The most commonly occurring AEs included thrombocytopenia (60.0% in the SVd group versus 27.0% in the Vd group), nausea (50.2% versus 9.8%, respectively), fatigue (42.1% versus 18.1%), diarrhea (32.3% versus 25.0%), anemia (36.4% versus 23.0%), decreased appetite (35.4% versus 5.4%), PN (32.3% versus 47.1%), decreased weight (26.2% versus 12.3%), asthenia (24.6% versus 13.2%), cataracts (21.5% versus 6.4%), and vomiting (20.5% versus 4.4%). These AEs were all more commonly reported in the SVd group than in the Vd group, except for PN, which occurred more frequently in the Vd group. Other AEs that occurred more frequently in the SVd group included neutropenia (14.9% in the SVd group versus 5.9% in the Vd group), dizziness (12.3% versus 3.9%, respectively), and nasopharyngitis (11.8% versus 4.9%, respectively).

Grade 3 and 4 AEs also occurred more frequently in the SVd group at 79.0% compared to 55.9% of patients in the Vd group. Grade 3 or higher AEs occurred in 85.1% of patients in the SVd group compared to 61.3% of patients in the Vd group. The most commonly occurring grade 3 or higher AEs were thrombocytopenia (39.5% in the SVd group versus 17.2% in the Vd group) and anemia (15.9% versus 10.3%, respectively).

Serious Adverse Events

Serious adverse events (SAEs) were more frequent in the SVd group at 51.8% compared to 37.7% of patients in the Vd group. The most common SAE was pneumonia, which occurred in 11.8% of patients in each treatment group.

Adverse Events Leading to Dose Modifications

AEs leading to dose modifications were more frequent in the SVd group (88.7%) than in the Vd group (76.5%). Specifically, AEs leading to dose reductions or dose interruptions were both more common in the SVd group than in the Vd group (72.3% versus 51.0%, and 85.6% versus 68.1%, respectively).

Mortality

Deaths were reported for 6.2% patients in the SVd group and 5.4% patients in the Vd group. A breakdown of the causes of death is reported in Table 32. The most common cause of death in the SVd group was septic shock (1.5%) and pneumonia (1.0%). The most common cause of death in the Vd group was pneumonia (1.5%).

Notable Harms

Notable harms pre-specified in the CADTH systematic review protocol included pain, anorexia, nausea, gastrointestinal disorders, thrombocytopenia, and neutropenia. The incidence of pain was similar between both treatment groups (2.6% of patients in the SVd group versus 2.0% in the Vd group). No patients reported anorexia. Nausea (50.3% in the SVd group versus 9.8% in the Vd group), gastrointestinal disorders (69.2% versus 44.6%, respectively), thrombocytopenia (60.0% versus 27.0%) and neutropenia (15.9% versus 5.9%) were more common in the SVd group than in the Vd group.

The incidence of grade 2 or higher PN events was a key secondary safety end point of the BOSTON trial. The CADTH systematic review protocol also pre-specified PN as a notable harm. Peripheral neuropathy was less commonly reported in the SVd group than in the Vd group at the primary analysis (21.0% versus 34.3%, respectively). Most events were grade 2. Results at the updated analysis were consistent with the primary analysis.

Critical Appraisal

Two interim analyses were planned for the BOSTON trial. The first interim analysis was for sample size re-adjustment. At the first interim analysis, it was determined that no re-adjustment of sample size would be conducted. The second interim analysis was for an efficacy analysis based on PFS, and would allow for a conclusion of efficacy, and stopping for futility (non-binding). There was agreement between the sponsor and the DSMB to use the second interim analysis as the final analysis for PFS. As more than 75% of the planned PFS events occurred, the DSMB determined that the primary end point of PFS was met at a 1-sided alpha of 0.025, meeting the stopping boundary.

The sponsor conducted an additional analyses of efficacy end points at an updated time point (February 15, 2021). This updated analysis was not pre-specified and was not considered in the statistical analysis plan. All results from the updated analysis should be considered descriptive.

While not unique to the BOSTON trial, it is possible the choice of subsequent therapies affected the efficacy assessments of OS, as analyses for OS included patients who received subsequent therapies. A total of 69 patients in the SVd group and 116 patients in the Vd group received subsequent anticancer therapies. Disproportional differences were noted between treatment groups in the types of subsequent anticancer therapies received, as more patients in the SVd group received lenalidomide, pomalidomide, bortezomib, carfilzomib, and daratumumab compared with the Vd group. In addition, patients in the Vd group were eligible to cross over to receive a selinexor-based regimen. The differences in subsequent therapies are expected to introduce bias in the efficacy analyses of OS and other patient outcomes. However, the direction and extent of the biases are difficult to predict. It is possible that crossing over also affected safety analyses. Patients crossing over to a selinexor-based regimen would have experienced selinexor-related AEs. It is therefore possible that differences between treatment groups in the incidence of selinexor-related AEs are underestimated.

Regarding patient disposition in the BOSTON trial, a greater proportion of patients in the SVd group discontinued treatment due to withdrawal by the patients ||||||||||||||| than in the Vd group ||||||||||||||| The sponsor clarified that patient withdrawal was due to AEs ||||||||||||||| in the SVd group versus ||||||| in the Vd group), logistical reasons ||||||| versus |||||||, poor health or entering hospice care (||||||| versus |||||||, burden of assessments (||||||| versus |||||||, respectively), and IRC-confirmed disease progression (||||||| versus |||||||, respectively). An additional ||||||| patients in the SVd group versus ||||||| patients in the Vd group did not provide any additional information.⁵ Discontinuation due to AEs and/or toxicity were initially reported by 16.9% of patients in the SVd group versus 11.3% of patients in the Vd group. The clarification provided by the sponsor regarding reasons for “withdrawal by the patient” suggests there is additional toxicity related to SVd as an additional ||||||| patients in the SVd group versus ||||||| patients in the Vd group discontinued due to AEs. It is possible that these differences in patient disposition may have affected some efficacy end points, as this imbalance in discontinuations may be a result of informative censoring. Because PFS was the primary end point, it is possible that the analyses were conducted on a population of patients in the SVd group who could better tolerate the investigational treatment. The results of a number of sensitivity analyses conducted by the sponsor continued to support the primary analysis of PFS and favoured treatment with SVd over Vd. However, the sponsor also conducted a sensitivity analysis that considered treatment discontinuation as an event (Table 46); this analysis was the only sensitivity analysis for PFS that did not find a statistically significant improvement in PFS for the SVd group (HR = 0.95; 95% CI, 0.76 to 1.19). The imbalance in patient discontinuations may also have affected other secondary outcomes, specifically TTD. The median TTD was 7.10 months (95% CI, 6.44 to 8.54) in the SVd group and 7.95 months (95% CI, 6.80 to 9.23) in the Vd group (HR = 0.99; 95% CI, 0.79 to 1.23). It was expected that an improvement in PFS would translate to a longer TTD in the investigational therapy group versus the control; however, this was not the case in the BOSTON trial.

The clinical experts consulted by CADTH for this review acknowledged that the eligibility criteria of the BOSTON trial, while similar to those of other clinical trials for MM, were restrictive and likely excluded patients who would be candidates for SVd in clinical practice. For example, the trial excluded patients who had received radiation, chemotherapy immunotherapy, or other anticancer therapy before 2 weeks before receiving study treatment. The eligibility criteria also excluded patients with severe PN, plasma cell leukemia, and comorbidities. Also excluded were those with spinal cord compression, documented systemic light chain amyloidosis, and major surgery less than 4 weeks before beginning the study therapy. In general, exclusion criteria were acknowledged to be restrictive and exclusive of patients who could potentially benefit from treatment with SVd.

The demographic and clinical characteristics of patients randomized in the BOSTON trial were generally considered representative of patients living in Canada, as confirmed through consultation with clinical experts for this review. However, the clinical experts noted that the proportion of patients with previous exposure to lenalidomide was low (39.5% in the SVd group and 37.2% in the Vd group). In Canadian clinical practice, lenalidomide would be administered to most patients as a first-line therapy in a metastatic setting. Therefore, it is expected that nearly all patients living in Canada would have had previous exposure to lenalidomide.

The BOSTON trial was a phase III trial comparing SVd to Vd. The use of Vd as a comparator was not considered appropriate in the current Canadian context. In particular, the dose of bortezomib was differed between the 2 treatment groups. Bortezomib was administered at a dose of 1.3 mg/m² SC on days 1, 8, 15, and 22 of each 35-day cycle in the SVd group. In the Vd group, bortezomib was administered at a dose of 1.3 mg/m² SC on days 1, 4, 8, and 11 of each 21-day cycle for the first 8 cycles; after cycle 8, bortezomib was administered at a dose of 1.3 mg/m² SC on days 1, 8, 15, and 22 of each 25 day cycle. The clinical experts consulted by CADTH confirmed that the twice-weekly dosing in the Vd group is not commonly used in clinical practice. In addition, Vd is not a common regimen administered to patients. The clinical experts confirmed that Vd is often administered to patients as part of a triplet regimen. Overall, the clinical experts agreed that the Vd was not an appropriate comparator in the current Canadian treatment landscape for MM. However, it was acknowledged that enrolment in the BOSTON trial began in 2017, when the standard of care may have been different, and global variation in reimbursement of treatments may have led to the decision to choose Vd as the comparator for the BOSTON trial.

Indirect Comparisons

Description of Studies

Sponsor’s Indirect Treatment Comparison

All 66 studies included patients with RRMM. Most were phase II or III trials, including 19 phase II trials (29%) and 45 phase III trials (68%). Details about trial phase were not reported for 1 study. Another study was a retrospective matched-pair analysis that was included to complete the treatment networks. Of the studies, 50 (76%) were open-label, 12 (18%) were double-blind, and 4 did not report blinding procedures. The median follow-up ranged from |||||||||||||| months (median = |||||||). Sample sizes in treatment groups ranged from ||||||| patients per treatment group (median = |||||||). The median age ranged between |||| and |||| years (median = ||||); the median ages were similar across most trials. The proportion of males across the trials ranged from |||||||| (median = ||||).

Dolph et al. (2021)

A total of 21 studies were included in the network for PFS for second-line treatment, including 14 randomized controlled trials (RCTs) with only second-line patients, 5 studies with a mixed population but in which the majority were second-line patients, and 2 studies in which the majority of patients were in the third line of treatment or later. The 2 studies that included majority third-line or later patients were stated to be necessary to connect dexamethasone with lenalidomide plus dexamethasone (Rd). A total of 24 studies were included in the network for PFS in the third line or later, including 19 studies with outcomes reported exclusively in the third line or later. Four studies included patients in the second line and third line or later, and 1 study included exclusively second-line patients but was necessary to link Vd with bortezomib.

A total of 15 studies were included in the network for OS in the second line; 4 of these studies reported only second-line OS information. A mixed population was enrolled for 9 studies, with the majority of patients being in the second line, and 1 study enrolled primarily patients in the third line or later. A total of 22 studies were included in the network for OS in the third line, including 11 studies that reported outcomes in the third line or later and 10 studies with a mixed population but in which the majority were in the third line or later. A single study reported results exclusively in the second line but was required to connect bortezomib with Vd.

A total of 20 studies were included in the network for objective response rate in the second line, including 12 RCTs reporting outcomes exclusively in the second line. A mixed population was reported in 8 of the studies with a majority of second-line patients. A total of 27 studies were included in the network for ORR in the third line, including 17 that reported outcomes exclusively in the third line or later. A mixed population was reported for 9 studies with the majority of patients being in the third line or later. A single study was included that reported exclusively second-line results but was required to link bortezomib with Vd.

Arcuri et al. (2021)

Six studies included lenalidomide in the control group, with 8 studies including bortezomib in the control group; only 3 studies did not include either of these treatments, and instead included carfilzomib (n = 1) or pomalidomide (n = 2) in the control group. Interventions assessed in the studies included vorinostat (n = 1), panobinostat (n = 1), pomalidomide (n = 1), pegylated doxorubicin (n = 1), cyclophosphamide (n = 1), elotuzumab (n = 1), pembrolizumab (n = 1), autologous stem cell transplantation (n = 1), venetoclax (n = 1), carfilzomib (n = 2), ixazomib (n = 2), daratumumab (n = 3), isatuximab (n = 1), and selinexor (n = 1). Studies included a range of median follow-up of 6 months to 36.8 months. The studies also included patients who had received between 1 and 3 prior therapies. Studies were published between 2007 and 2020. No further assessment of heterogeneity was conducted by the authors.

Botta et al. (2021)

Six phase III RCTs (CASTOR, ENDEAVOUR, OPTIMISM, CANDOR, IKEMA, and BOSTON) were included, representing 1,615 RRMM patients who were previously exposed to lenalidomide and 984 patients who were refractory to lenalidomide.⁶ The authors reported that studies were well balanced for the presence of patients refractory to lenalidomide, who accounted for approximately 70% of patients, except for the CASTOR trial, in which 50% of the patients were refractory to lenalidomide. Studies were also well balanced in terms of exposure to bortezomib, accounting for approximately 65% of patients, except for patients in the IKEMA trial, which had 85% to 93% of patients with previous exposure to bortezomib. The proportions of patients in second-line therapy were well balanced across trials, accounting for approximately 45% of trial patients.⁶ No further assessment of study and patient characteristics was provided.

Efficacy Results

Sponsor’s ITC

Progression-free survival: Regarding the network meta-analysis (NMA) conducted for second-line treatment, compared to Vd, ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| Regarding the NMA conducted for third-line or later treatment, compared to Vd, |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| The remaining regimens, including ||||||||||||||||||||||||||||||||||||||||||. Pairwise comparisons against selinexor ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||

Overall survival: Regarding the NMA conducted in the second-line, there were |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| Regarding the NMA conducted in the third line or later, compared to Vd, ||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| The remaining regimens, ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| Pairwise comparisons against selinexor suggested that |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||

Objective response rate: Regarding the NMA conducted in the second-line, compared to Vd, |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| Pairwise comparisons suggested that ORR was ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| Regarding the NMA conducted in the third line or later, compared to Vd, ||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| Pairwise comparisons suggested that ORR was |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||

Dolph et al. (2021)

Progression-free survival: In the second line, compared to Vd, the greatest benefit was suggested to be from daratumumab plus bortezomib plus dexamethasone (DVd), followed by DRd, and Kd. There were no differences between the remaining treatments of interest, including SVd. In the third line, compared to Vd, treatments that were favoured included DRd, DVd, Kd, and pomalidomide plus bortezomib plus dexamethasone (PVd). There were no differences between the remaining treatments of interest, including SVd. Specific estimates for comparisons were not provided.

Overall survival: There were no differences between treatments in the second line, including SVd. In the third line, DVd, daratumumab plus carfilzomib plus dexamethasone (DKd), and Kd were favoured over Vd. The remaining treatments did not show any differences, including SVd. The treatment effects were not reported.

Objective response rate: In the second line, treatments favoured over Vd included DKd, DVd, and SVd. There were no differences between the remaining interventions of interest. In the third line, DVd, DKd, and Kd were favoured over Vd. The remaining treatments did not show any differences, including SVd. No treatment effects were reported.

Arcuri et al. (2021)

Progression-free survival: There were no differences between selinexor and any of the comparators of interest: carfilzomib (HR = 0.86; 95% CI, 0.50 to 1.48), daratumumab (HR = 0.65, 85% CI, 0.38 to 1.10), high-dose chemotherapy (HR = 1.24; 95% CI, 0.65 to 2.38), isatuximab (HR = 0.85; 95% CI, 0.44 to 1.65), ixazomib (HR = 0.98; 95% CI, 0.55 to 1.75), pomalidomide (HR = 0.97; 95% CI, 0.50 to 1.87). The heterogeneity measured for PFS, as assessed by I², was 64%.

Overall survival: Estimates for comparisons between each treatment were not provided for OS. However, in general, most treatments indicated no difference.

Botta et al. (2021)

Progression-free survival: Results suggested that PFS among patients exposed to lenalidomide was favoured with Isa-Kd (HR = 0.34; 95% CI, 0.18 to 0.64), followed by KDd (HR = 0.36; 95% CI, 0.22 to 0.61), DVd (HR = 0.38; 95% CI, 0.38; 95% CI, 0.26 to 0.56), PVd (HR = 0.61; 95% CI, 0.49 to 0.76), SVd (HR = 0.63, 95%CI, 0.41 to 0.96), and Kd (HR = 0.69; 95% CI, 0.52 to 0.92) compared to Vd.⁶ Among patients who were lenalidomide-refractory, PFS was favoured with treatment with DVd (HR = 0.36; 95% CI, 0.21 to 0.62), followed by KDd (HR = 0.38; 95% CI, 0.21 to 0.68), Isa-Kd (HR = 0.48; 95% CI, 0.25 to 0.92), and PVd (HR = 0.65; 95% CI, 0.50 to 0.84) compared to Vd. No difference was observed between Kd (HR = 0.80; 95% CI, 0.57 to 1.12) and Vd.⁶

Harms Results

Sponsor’s ITC

No analyses for harms were conducted in the sponsor’s ITC.

Dolph et al. (2021)

No analyses for harms were conducted in the ITC conducted by Dolph et al.

Arcuri et al. (2021)

The authors conducted an analysis for SAEs. However, because the analysis for SAEs did not include selinexor, these results are not reported.

Botta et al. (2021)

No analyses for harms were conducted in the ITC conducted by Botta et al.

Critical Appraisal

Sponsor’s ITC

The sponsor included 17 trials in its ITC. There is likely high heterogeneity across study and patient characteristics. Differences in these study and patient characteristics may result in uncertainty in the analyses as the studies may not necessarily be comparable. In addition, the proportion of patients in different lines of therapy may not be similar across treatment groups within studies, and across studies. It is likely that variations in patient characteristics were present in the trials and unaccounted for.

The clinical experts consulted by CADTH for this review also emphasized the importance of considering subgroups of patients who would be exposed to lenalidomide exposed versus refractory to lenalidomide. The sponsor did not conduct any sensitivity analyses to determine the differences in treatment effect for these patient groups. These patient subgroups were highlighted as it is expected that most patients living in Canada will receive a lenalidomide-based regimen in the first line, and that subsequent therapy should consider the patient’s initial response to first-line therapy.

Networks of evidence were separated by line of therapy (second line, third line, or later line), which was considered appropriate given that patients in later lines of therapy tend to have worse outcomes. However, within networks, studies that included a mix of patients in multiple lines of therapy were included in networks in which the majority of patients represented patients in either the second line or later lines of therapy. This may introduce bias as patients in earlier or later lines of therapies can influence each network differently. Patients receiving second-line therapy may overestimate the efficacy of treatments included in studies in the third-line or later networks, while patients receiving later lines of therapy may underestimate the efficacy of treatments included in the second-line networks.

Trials were phase II and III trials, and the earlier-phase trials may not be powered to test hypotheses; inclusion of phase II trials is expected to introduce bias into the NMAs that may not be present in phase III trials, which are typically designed to detect differences between different treatment groups. A retrospective matched-pair analysis, which was required to link bortezomib to Vd (no RCTs were available for this link), was also included. Inclusion of this retrospective study therefore does not satisfy the transitivity assumption of the ITC as all other studies were clinical trials. Because the sponsor considered the connection between bortezomib and Vd to be necessary, it included this retrospective matched-case analysis to allow for comparisons of included regimens. Inclusion of this retrospective matched-case analysis is expected to introduce considerable uncertainty to the NMAs.

Overall, the networks of the NMAs were complex, leading to a high degree of variability. Methodological limitations are likely to have introduced further uncertainty into the analyses. For example, the sponsor did not conduct adjustments for crossover. Crossover to investigational treatment from a control is expected to underestimate the treatment effect observed in that trial and influence the analyses of the ITC. Important effect modifiers were not controlled for. Subgroup analyses were not performed due to small sample sizes. However, the lack of adjustment may introduce bias that can affect treatment comparisons.

Dolph et al. (2021)

The ITC conducted by Dolph et al. was similar to the ITC provided by the sponsor. As the methodology was similar to the ITC provided by the sponsor, the results of both ITCs could be compared. In general, results reported the same or similar conclusions regarding favoured treatments and the efficacy of SVd relative to other interventions. The consistency between these 2 analyses supports the position that the analyses conducted by the sponsor and Dolph et al. are replicable. However, limitations associated with the sponsor’s ITC are linked to the ITC conducted by Dolph et al. Critiques of the sponsor’s ITC reported in a preceding section should also be considered for the ITC published by Dolph et al.

The authors conducted an additional NMA including only Vd-containing regimens. This was preferred methodologically as it did not rely on a retrospective study to link treatments and allowed for comparisons between regimens with 1 shared common anchor; in this case all regimens were compared to Vd. The authors also stated that this analysis was highly relevant as lenalidomide is used in most patients as a first-line option and would not likely be used in later lines. Therefore, lenalidomide-based regimens are likely not important comparators in the second and later lines. The clinical experts consulted by CADTH for this review supported this statement and agreed that lenalidomide-based regimens would not be competing with other regimens in the second or later lines as they would most likely be used in the first line. However, the authors did not report specific results, and it is unclear exactly which interventions were favoured over the others.

The authors also reported that the CASTOR study, which was included in some networks, incorporated 2 trial characteristics that were not consistent with usual clinical practice and magnified the effect of daratumumab in the study. Specifically, the CASTOR study administered bortezomib twice weekly even though most clinicians administer bortezomib once weekly, and the trial required that bortezomib be discontinued after 24 weeks in both the DVd and Vd treatment groups, resulting in treatment with daratumumab to be compared to no treatment after 24 weeks. The clinical experts consulted by CADTH for this review confirmed that bortezomib is often administered beyond 24 weeks (or 8 cycles) for patients who can tolerate and respond well to treatment. The CADTH team agreed that this is likely to have amplified the treatment effects of daratumumab, and biased results that did support most daratumumab-based regimens in the NMAs.

Arcuri et al. (2021)

There is likely high variation in patient characteristics across the trials, which is likely to have introduced biases and result in considerable uncertainty in the analyses. The authors did not report a thorough assessment of heterogeneity. However, variations were reported across trial characteristics. Studies were published between 2007 and 2020; treatment practices of 2007 are likely not the same as current treatment practices, and the patient groups being compared are likely not the same as new therapies that have been introduced that alter the treatment pathways for patients and their outcomes. Differences in treatment duration were not accounted for in the analyses. The authors acknowledged that prolonged treatment duration may lead to increased PFS and higher rates of near-complete or complete responses. It is possible that effect modifiers that could affect efficacy analyses may be present but were unaccounted for. For example, the authors included patients across multiple lines of therapies. The clinical experts consulted by CADTH for this review confirmed that patients in later lines of therapy likely will experience poorer outcomes, and differences in patients across different lines of therapy may under- or overestimate treatment effects.

The authors connected studies through a common comparator group of either Rd or Vd based on the assumption that these 2 treatments are equally effective. This allowed the authors to create a single control group, a shorter path for indirect comparisons, and greater power to detect differences. However, 3 studies that did not include either Rd or Vd as a comparator and instead included Kd or pomalidomide plus dexamethasone (Pd), were also incorporated into this comparison group. The authors conducted a sensitivity analyses that separated the control group into 2 categories: 1 group included lenalidomide- and pomalidomide-based regimens, and another included bortezomib-based regimens. The authors concluded that the 2 treatments were equivalent, which further supported their decision to group these categories together. The clinical experts consulted by CADTH for this review did not agree with the assumption that Rd and Vd were equally effective treatments. In addition, the clinical experts also disagreed that Kd and Pd were equally effective treatments; however, they acknowledged that use of Pd would occur after treatment with Rd, and that Pd would be expected to be less efficacious for patients as it is used in a later line in patients previously treated with an immunomodulatory drug. The CADTH team therefore considered comparisons conducted in this ITC to be inappropriate, as data for treatments that are not considered equivalent were combined to create connections between regimens.

In general, details of the methodology used by the authors for the ITC were sparse. It is not possible to provide a full appraisal of these methods. The authors did not report on whether they adjusted for crossover in the trials, although it is unlikely. Treatment crossover could have biased efficacy analyses of these trials. However, it was reported that the authors conducted NMAs with fixed effects, unless the I² values were greater than 40%, in which case random-effects models were used. The I² value of the NMA for PFS was 64%, which indicates that a random-effects model was used. The analyses of OS and SAEs were reported to have an I² value of 0; however, a random-effects model was used for the analysis of OS. The use of random effects was considered appropriate given the number of comparators and the high amount of heterogeneity; however, without an assessment of model convergence and consistency, it is not completely possible to know which model was best for these analyses.

Conclusions

One multinational, sponsor-funded, open-label RCT, BOSTON, was included in the CADTH review. SVd demonstrated statistically significant and clinically meaningful improvement in PFS compared to Vd in a population of patients with MM who had received 1 to 3 prior lines of therapy. At the time of the analysis, median OS was not reached; however, other secondary end points (e.g., ORR, DOR, TTR, and TTNT) were supportive of the primary end point of PFS, demonstrating improved efficacy with SVd over Vd. An updated analysis continued to support the improved PFS of SVd over Vd, although these results were considered descriptive. The comparator in the BOSTON trial, Vd, was not considered appropriate for the current Canadian treatment landscape due to changes in standard of care. Four ITCs, including 1 submitted by the sponsor and 3 published ITCs, compared the efficacy of SVd to other relevant comparators (i.e., DVd, DRd, Kd, PVd, CyBorD, and Isa-Pd). The ITCs were congruent with direct evidence from the BOSTON trial that found improved PFS and ORR with treatment SVd over Vd. However, the ITCs also suggested that other regimens, such as daratumumab-based regimens, may be preferred over SVd, and this suggestion was supported by the clinical experts. However, the methodological limitations and heterogeneity across patients included in the ITCs limit the ability to draw firm conclusions. The collected HRQoL data suggest that there were no differences between patients in the SVd and Vd treatment groups; however, HRQoL data also highlight impacts on patient vision in the SVd group, although this finding should be interpreted with caution given the exploratory nature of the analysis. Detrimental effects on patient vision were also observed in harms data, which indicate an increase in cataracts in the SVd group. Notable harms that occurred more frequently in the SVd group included nausea, gastrointestinal disorders, thrombocytopenia. and neutropenia. In general, AEs related to SVd were described by the clinical experts consulted by CADTH for this review as manageable.

Introduction

Disease Background

Multiple myeloma is a plasma cell cancer caused by the growth of cancer cells in the bone marrow. In Canada, more than 3,000 new cases of MM are diagnosed annually, with slightly more cases occurring in men than women.¹ While new therapies have been introduced that can improve a patient’s OS and PFS, MM remains an incurable condition.² Some estimates suggest that the median survival for patients with MM is approximately 5 years, and during this time patients can receive 4 or more lines of therapy.² Multiple myeloma is a heterogenous condition that typically affects older adults around the age of 65 years, and patient’s outcomes can depend on many factors, including disease stage, prognostic indicators, and early treatment of symptomatic disease to limit or avoid organ damage.¹ Typically when MM is suspected clinically, patients are tested for the presence of M proteins, although a small proportion of patients (approximately 2%) may present without any evidence of M protein.⁷

The clinical experts consulted by CADTH for this review highlighted a particular need to prevent skeletal damage due to the disease. Bone disease is 1 of the main causes of morbidity for patients, and can be detected using imaging techniques, such as MRI or PET and CT scans.⁷ The most common site of pain related to bone pain is the lumbar spine. Patients may also initially present with lytic lesions, anemia, fatigue, infections, weight loss, hypercalcemia, and/or renal dysfunction.¹ Patients may also have cytogenetic abnormalities that can influence the course of their disease, response to therapy, and overall prognosis. Cytogenetic abnormalities can include t(4;14), t(14;16), t(14;20), del(17p), or gain(1q), which can be detected using fluorescence in situ hybridization (FISH) technology. The R-ISS system is used to classify the stage of disease for patients diagnosed with MM; the combined elements of tumour burden and disease biology (e.g., the presence of high-risk cytogenetic abnormalities) are used to create a prognostic index to assist in clinical care and comparison of data from clinical trials.⁷

Standards of Therapy

The treatment landscape for MM has changed significantly in recent years with the emergence of new therapies in newly diagnosed and relapsed or refractory settings.²

According to the clinical experts, initial therapy depends on whether patients are eligible for transplant at diagnosis. Initial treatment for patients who are eligible was stated to include induction therapy for 4 months with cyclophosphamide, bortezomib, and CyBorD. Other treatment regimens include lenalidomide plus bortezomib plus dexamethasone (RVd) for high-risk patients or those who fail to response to initial therapy when funding is available. Other regimens containing daratumumab and carfilzomib have recently been described but are rarely used as they are not currently funded for induction. Following induction therapy, the clinical experts stated that patients undergo stem cell collection with growth factors with or without high-dose cyclophosphamide. Following this, patients undergo treatment with melphalan followed by a stem cell transplant and then consolidation therapy with RVd in some jurisdictions (depending on local practices and funding). Patients continue with lenalidomide maintenance therapy until disease progression; this was stated to be the standard across all Canadian jurisdictions. The clinical experts also acknowledged that some patients with high-risk cytogenetics confirmed by fluorescence in situ hybridization (FISH), such as t(4:14), t(14:16), 1q gain(1q), and del(17p), will be offered tandem transplants.

For patients who are not eligible for transplant, RVd was described as the most commonly used treatment; other treatment options include Rd and CyBorD, although CyBorD has been used less frequently in current practice since the approvals of lenalidomide-based regimens. Daratumumab-based regimens such as DRd or daratumumab plus bortezomib plus melphalan plus prednisone (DVMP) were expected to be used more frequently due to recent approvals from CADTH and funding approvals across jurisdictions. The clinical experts indicated that they expected that daratumumab-based regimens would be the preferred front-line option, with DRd as the most likely choice.

Both clinical experts agreed that most patients would be started on IV pamidronate or zoledronic acid (Zometa) to prevent bone-related AEs.

At relapse, treatment for patients was stated to depend on patient factors, including age, comorbidities, and previous treatments. In the second-line treatment setting, a second transplant may be an option for transplant-eligible patients. Although this was stated not to be a common approach because of the available alternatives, patients with long responses to the first transplant will often be considered for a second transplant if their age and comorbidities are not contraindications. The clinical experts stated that most patients will receive a daratumumab-containing regimen, likely DRd or DVd; patients who are refractory to lenalidomide would usually receive DVd while those who previously received bortezomib would receive DRd. Current treatment practices suggest using bortezomib or another PI after treatment with lenalidomide.

Regimens containing carfilzomib were acknowledged to be available to patients in the second line, although these regimens are typically reserved for relapse after daratumumab-based regimens in the third line or after. Pomalidomide-based regimens were stated to be considered in the third or fourth line of therapy. Isatuximab-based regimens were stated to be another option for patients, especially for those who are not eligible for daratumumab-based regimens, but these are currently not funded. The clinical experts stated that isatuximab-based regimens would likely not be effective for patients who progress on a daratumumab-based regimen. Belantamab is another option that could be available to patients; however, this treatment is only available through special access and is not used frequently.

The treatment practices described by the clinical experts also align with recommended regimens by National Comprehensive Cancer Network (NCCN) guidelines.⁸

Drug

Selinexor is a reversible covalent selective inhibitor of nuclear export (SINE) compound that blocks the exportin 1 (XPO1) protein, which is a nuclear transport protein that transports cargo proteins within the cell. Selinexor mediates the nuclear transport of many cargo proteins, including cargo proteins associated with the growth of oncogenic proteins as well as tumour suppressor proteins. Inhibition of XPO1 by selinexor leads to reductions in cancer cells. When combined with bortezomib and dexamethasone, the resulting SVd regimen demonstrates antitumour activity, including in in vivo models resistant to PIs.³ Selinexor is indicated to be administered in combination with bortezomib and dexamethasone. Selinexor is to be administered at a dosage of 100 mg orally once weekly on day 1 of each week. Bortezomib is to be administered at a dosage of 1.3 mg/m² SC once weekly on day 1 of each week for 4 weeks followed by 1 week off. Dexamethasone is to be administered at a dosage of 20 mg orally twice weekly on days 1 and 2 of each week.³

Selinexor was submitted to CADTH pre-Notice of Compliance and is anticipated to be approved by Health Canada on June 2, 2022. Selinexor has not previously been reviewed by CADTH. The sponsor has requested reimbursement of selinexor as per the Health Canada indication.

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

One patient group, Myeloma Canada, provided input on the combination of selinexor with bortezomib and dexamethasone for the treatment of MM in adult patients. Myeloma Canada is a charitable organization with a mission of improving the lives of all patients living in Canada affected by myeloma. The patient group conducted an online survey of patients and caregivers across Canada from December 2021 to January 2022. The survey was distributed through email and social media. Overall, 254 patients and 1 caregiver responded to the survey (N = 255); respondents represented people living in Canada from Alberta, British Columbia, Manitoba, New Brunswick, Newfoundland and Labrador, Nova Scotia, Ontario, Quebec, and Saskatchewan. Three respondents from France were also included. The following is a summary of key input from the perspective of the patient group.

Disease Experience

According to Myeloma Canada, patients considered it extremely important to control symptoms of infections, kidney problems, mobility, and neuropathy related to myeloma. Patients also indicated that symptoms significantly impacted their abilities to travel, work, exercise, and concentrate. Infections, kidney problems, mobility, and neuropathy are among the symptoms that patients report as extremely important to control. While half of respondents stated they did not need the support of a caregiver or family member to help manage disease or treatment-related symptoms, 39% stated that they did, and 4% indicated that they did but were unable to access the support they need. Parking costs, travel costs, drug costs, lost income due to absence from work, and lost income or pensions due to early retirement were identified as the most significant financial implications due to myeloma treatment.

Experience With Treatment

Of patients who had or were currently receiving treatment with bortezomib and dexamethasone, more than half indicated that, overall, the side effects were either bearable (7 of 23 responses) or somewhat bearable (7 of 23 responses). Side effects considered “totally unbearable” by most (19 of 23) patients included fatigue, diarrhea, and nausea, while the most “tolerable” side effects were anemia and thrombocytopenia. Peripheral neuropathy was highlighted by many (19 of 23) patients as a side effect and an important symptom to control (7 of 19) and to reduce the severity (7 of 19).

Two respondents had experience with SVd through participation in the BOSTON trial; 1 had not relapsed since receiving SVd through the BOSTON trial, while the other relapsed within 3 months and was receiving a different treatment. One patient indicated the trial regimen was very effective in helping control their myeloma, while the other patient reported that it was somewhat ineffective. Nausea was stated to be a “somewhat tolerable” side effect, while diarrhea, PN, and vomiting were “somewhat intolerable.” Other reported side effects included thrombocytopenia, anemia, fatigue, decreased appetite, and weight decrease. One respondent indicated that SVd had met their expectations in treating their myeloma but mentioned that it was “too soon to tell” whether myeloma treatment had improved their overall QoL.

Improved Outcomes

Respondents most frequently mentioned effectiveness of treatment, QoL, accessibility and portability of treatment, manageable side effects, and access to a supportive and communicative care team as important when considering any treatment for myeloma. The side effects that patients most frequently ranked as important to avoid when considering new treatments included infections, vomiting, pain, confusion, decreased appetite, and neuropathy. Many patients indicated a preference for orally administered treatments, while SC injection and infusion were less preferred. Many respondents indicated that fewer trips to a cancer centre or hospital for treatment would positively affect their QoL.

Myeloma Canada additionally described that access to health care services and treatment options for MM patients varies across the country and that a treatment option that would minimize patients’ time in hospital would be welcome. The patient group added that the treatment under review would fulfill an unmet need in patients requiring a fourth line of therapy as 1 of the only options for them.

Clinician Input

Input From Clinical Experts Consulted by CADTH

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

Unmet Needs

The clinical experts highlighted the need for treatments that improve survival, provide deeper and longer-lasting remissions, and improve disease-related symptoms and complications, such as pain and renal failure. In addition, treatments would have less of a negative impact on patient QoL and require fewer clinic visits.

The clinical experts acknowledged that most patients will relapse with currently available therapies. Treatments are palliative and may prolong patients’ lives, but they do not provide a cure. Patients will eventually become refractory to available treatments. The clinical experts also stated that remission beyond the second-line setting is becoming brief. Patients with high-risk cytogenetics were stated to be at particularly high risk of progression and poor outcomes as they may be more likely to fail to respond to initial therapy or have a brief initial remission. The clinical experts also highlighted unmet needs among patients who are not eligible for transplant as they have a poorer prognosis compared to patients who are eligible. The side effects of some treatments were stated to affect tolerability to treatment and effectiveness. Many treatment options for patients are provided intravenously or subcutaneously at a cancer centre as often as once or twice per week, resulting in a significant burden on patients, caregivers, and treatment centres.

Place in Therapy

The clinical experts acknowledged that selinexor would not require a significant paradigm shift and that other therapies to treat MM are available. However, selinexor operates through a mechanism of action that differs from other treatments currently available for MM, presenting an opportunity to be effective in patients who become resistant to treatments that target other pathways. It is therefore likely that selinexor would be used as a later line of therapy. The clinical experts agreed that other regimens are preferred before using a regimen based on selinexor. The toxicity profile of selinexor was also stated to be different from other classes of drugs. Selinexor could be an attractive option for patients as it is administered orally and only once per week, potentially reducing the need for clinic visits.

Patient Population

The clinical experts agreed that there are no specific features that would make a patient a better candidate for selinexor. In general, patients with high-risk features (i.e., high-risk cytogenetics), poor performance status, and those who have received many prior lines of therapy will be unlikely to respond to other therapies, including selinexor. Patients with pre-existing anorexia, weight loss, or nausea may not be good candidates for this treatment as the side effects of selinexor include anorexia, weight loss and nausea.

Assessing Response to Treatment

A patient’s response to therapy is typically measured in terms of monoclonal protein and serum free light chains; based on these evaluations, a clinically meaningful response would include a sustained PR or better. Stable disease may also be considered an acceptable benefit to patients in some cases. Improvements in cancer-related complications, such as anemia, renal failure, hypercalcemia, and tumour-related pain, are also considered when assessing a patient’s response. Imaging of bones can assess response, but this was described as a less-sensitive outcome that may not always be helpful. In general, meaningful responses to treatment are expected to translate through improvements in patient OS and PFS. Improvements in QoL, myeloma-related symptoms, and treatment toxicity were also considered important outcomes when assessing patient’s response to treatment.

Typically, patients may be assessed for response every 4 weeks. Less-frequent monitoring of patients (i.e., every 2 or 3 months) may be warranted if patients demonstrate a stable long-term response and predictable and manageable toxicity.

Discontinuing Treatment

The clinical experts agreed that discontinuing treatment should depend on whether a patient’s disease progresses. When patients fail to respond to treatment, a change in therapy is required. Disease progression can be determined by measuring levels of serum free light chains and paraprotein, and sometimes with worsening hemoglobin, renal function, and bone imaging. Significant toxicities or AEs that cannot be managed through supportive care or dose modifications were also described as indications warranting discontinuation of therapy.

Prescribing Conditions

Treatment will need to be prescribed by a hemato-oncologist. Administration of therapies will require a specialty hematology or oncology clinic or equivalent. Physicians with expertise and experience in treating MM, such as a hematologist or oncologist, would be treating and monitoring patients. Patients may be monitored by other care providers, such as clinical associates, general practitioners in oncology, and nurse practitioners, particularly when a hematologist or oncologist is not routinely available onsite; however, these staff are typically under the supervision of a treating hematologist or oncologist.

Additional Considerations

The clinical experts emphasized that changes to the treatment paradigm for MM patients are likely to occur with approvals of daratumumab in the first-line setting. Because most patients will likely receive DRd as a first-line therapy, these patients will not receive daratumumab-based regimens upon relapse. The next-line option for patients would likely be a combination regimen that includes bortezomib and another PI (e.g., CyBorD or Kd). Selinexor in combination with bortezomib and dexamethasone could be considered as a second-line option; however, it may be more likely for SVd to be used in later lines of therapy.

Clinician Group Input

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

Two groups provided clinician input on the review of SVd for the treatment of adult patients with MM. The OH-CCO Hematology DAC submission was prepared by 7 physicians and the CMRG submission was prepared by 13 physicians.

OH-CCO provides timely evidence-based clinical and health system guidance on drug-related issues in support of CCO’s mandate, including the Provincial Drug Reimbursement Programs (PDRP) and the Systemic Treatment Program. The CMRG is a charitable organization whose membership consists of researchers in Canada who aim to develop better treatments that extend the life of melanoma patients, improve the QoL of those living with myeloma and related disorders, and find a cure for these diseases and other plasma cell disorders.

Unmet Needs

Both patient groups generally agreed that improving OS, PFS, disease-related symptoms, and HRQoL are important goals for ideal treatment. The CMRG mentioned that new treatments that reduce the burden of myeloma-related organ damage and symptoms (e.g., anemia, ongoing lytic bone destruction, renal failure, and hypercalcemia) would be ideal for patients. The CMRG also pointed out that, although CR is ideal, an incomplete response (e.g., minimal response or stable disease) may also afford symptom control or improvement. Treatments that control the disease and prevent worsening of patient’s symptoms (e.g., renal damage causing dialysis) were stated to be helpful for patients and prevent degeneration of performance status, leading to less resource utilization in the treatment of MM complications and reduced caregiver burden. The CMRG highlighted the need to prevent further lytic skeletal destruction; while complications related to skeletal damage can be arrested, fractures stabilized, and pain controlled, lytic lesions do not fully heal. Overall, affected bones remain at risk for future fractures and are further weakened from cortical structure.

The OH-COO group indicated the greatest unmet need for patients currently exists after the second-line setting; patients who failed daratumumab in the second-line setting could have the option to use this regimen in the third line. The CMRG expressed the need for new classes of anti-myeloma drugs to complement available treatments and improve patients’ convenience (e.g., with oral administration) and toxicity profiles.

Both groups agreed that patients who have the greatest unmet need for a selinexor-based regimen are those with RRMM who are refractory to immunomodulatory inhibitors, PIs, and anti-CD38 monoclonal antibodies. The CMRG also specified patients with renal insufficiency and poor risk features (e.g., high-risk cytogenetics, extramedullary disease, or highly proliferative disease) have the greatest unmet need for this therapy; the needs of patients who are earlier in their disease course and are refractory to lenalidomide and/or daratumumab were also highlighted.

Place in Therapy

According to the CMRG, patients with MM who are younger, fit, and eligible for transplant are initially treated with triplet bortezomib-based regimen induction, followed by autologous stem cell transplant, followed by maintenance with lenalidomide until disease progression. For patients ineligible for transplant, a lenalidomide-based regimen (e.g., Rd, RVd, or DRd) would typically be used with lenalidomide and continued until disease progression. The CMRG also stated that a small proportion of patients who are not eligible for transplant may receive a bortezomib regimen (e.g., bortezomib plus melphalan plus prednisone [VMP] or CyBorD) as a “fixed duration” regimen for approximately 9 cycles. Most newly diagnosed patients experience their first relapse on lenalidomide and are considered refractory to lenalidomide.

The OH-CCO group indicated most patients in the second line would receive daratumumab or isatuximab. The CMRG agreed that daratumumab would be offered in the second line, specifying that, depending on the initial regimen, it may be paired with either Vd or Rd.

As stated by CMRG, third-line therapy is based on either carfilzomib with dexamethasone and/or cyclophosphamide, pomalidomide with dexamethasone with or without cyclophosphamide, or dexamethasone with or without bortezomib. Carfilzomib was considered the most suitable for patients without cardiac comorbidities.

Therapies beyond the third-line treatment setting include treatments provided through clinical trials, special access programs, or palliative therapies. The CMRG noted that clinical trials almost exclusively require prior exposure to PIs, immunomodulatory drugs and anti-CD38 monoclonal antibodies. Selinexor was stated to be currently available through special access programs and was acknowledged by the CMRG to differ compared with currently available therapies based on the route of administration, side-effect profile, and supportive care needs. According to the clinical groups, physicians in Canada are seeking therapies such as selinexor that differ from currently available treatment options for patients who progress through funded options but are not yet candidates for palliative care.

The OH-CCO DAC expressed uncertainty about the specific placement of SVd into the current treatment paradigm. However, both groups generally agreed that daratumumab- or isatuximab-based regimens would be preferred in the second line before recommending SVd. Both groups agreed that this drug would not affect the treatment sequence employed in current practice.

Patient Population

Both clinician groups acknowledged that eligible patients would be identified by their treating physician or hematologist. The OH-COO DAC did not specify criteria for patients who would be least suited for treatment, although the CMRG indicated that newly diagnosed patients with MM would be least suitable for treatment with SVd.

Assessing Response to Treatment

Both groups indicated that patient responses to treatment would be assessed using conventional myeloma response criteria. Specifically, responses could be assessed using regular blood and urine examinations to measure M protein, quantitative immunoglobulins, free light chains and immunofixation, and imaging and/or bone marrow biopsies for patients with oligosecretory or non-secretory disease.

The clinical experts suggested that a clinically meaningful response to treatment in the setting of advanced disease includes a minimum 50% reduction in measurable disease (i.e., M protein or reduction in oligosecretory disease lesions). The CMRG emphasized that, depending on the severity of myeloma-related organ damage, a lesser response might be acceptable if the disease stopped progressing. Both groups agreed that patient responses to treatment would be assessed each cycle or approximately every month.

Discontinuing Treatment

Both groups agreed that discontinuation of treatment would be based on disease progression and toxicity.

Prescribing Conditions

Both groups agreed that SVd would be administered in outpatient clinics, hematology clinics, and hospitals.

Additional Considerations

The CMRG highlighted the route of administration of selinexor is advantageous for patients, as oral therapies are more convenient. In addition, the weekly administration of bortezomib as part of the regimen was acknowledged as another advantage, as the incidence of PN was stated to be lower than the biweekly dose typically used with bortezomib. The OH-CCO DAC additionally indicated that inclusion of bortezomib introduces uncertainty to this regimen, as SVd cannot be given to patients who are triple-refractory because of the need to be bortezomib-sensitive.

Drug Program Input

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

Table 4: Summary of Drug Plan Input and Clinical Expert Response

pERC = CADTH pan-Canadian Oncology Drug Review Expert Review Committee; ITC = indirect treatment comparison; PAG = Provincial Advisory Group; PI = proteosome inhibitor; SVd = selinexor plus bortezomib plus dexamethasone; Vd = bortezomib plus dexamethasone.

Clinical Evidence

The clinical evidence included in the review of selinexor 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 studies 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 sponsor-submitted long-term extension studies and additional relevant studies that were considered to address important gaps in the evidence included in the systematic review.

Systematic Review (Pivotal and Protocol-Selected Studies)

Objectives

To perform a systematic review of the beneficial and harmful effects of selinexor (Xpovio) at 100 mg administered orally in combination with Vd for the treatment of adult patients with MM who have received at least 1 prior therapy.

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 those considered to be important to patients, clinicians, and drug plans.

Table 5: Inclusion Criteria for the Systematic Review

AE = adverse event; DOR = duration of response; HRQoL = health-related quality of life; IMWG = International Myeloma Working Group; ORR = overall response rate; OS = overall survival; PFS = progression-free survival; SAE = serious adverse event; TTNT = time to next treatment; TTP = time to progression; TTR = time to response; WDAE = withdrawal due to adverse event.

^aLenalidomide-based regimens were considered for inclusion as comparators but were ultimately excluded on the basis that they would not be competing with selinexor in combination with bortezomib and dexamethasone, and instead used in sequence.

^bThis regimen is not currently funded across Canadian jurisdictions.

^cThis regimen has received a positive funding recommendation from CADTH.

^dOutcomes that were identified as being of particular importance to patients in the input received by CADTH from patient groups.

The literature search was performed by an information specialist using a peer-reviewed search strategy. Two CADTH clinical reviewers independently selected studies for inclusion in the review based on titles and abstracts, according to the pre-determined 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.

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 tool.¹²

Published literature was identified by searching the following bibliographic databases: MEDLINE All (1946‒) via Ovid and Embase (1974‒) via Ovid. All Ovid searches were run simultaneously as a multi-file search. Duplicates were removed using Ovid deduplication for multi-file searches, followed by manual deduplication in EndNote. The search strategy comprised both controlled vocabulary, such as the National Library of Medicine’s MeSH (Medical Subject Headings), and keywords. The main search concept was selinexor. Clinical trials registries searched included the US National Institutes of Health’s clinicaltrials.gov, WHO’s International Clinical Trials Registry Platform (ICTRP) search portal, Health Canada’s Clinical Trials Database, and the European Union Clinical Trials Register.

No filters were applied to limit the retrieval by study type. Retrieval was not limited by publication date or by language. Conference abstracts were excluded from the search results. Refer to Appendix 1 for the detailed search strategies.

The initial search was completed on February 15, 2022. Regular alerts updated the search until the meeting of the CADTH pan-Canadian Oncology Drug Review Expert Committee (pERC) on April 13, 2022.

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.¹³ Included in this search were the websites of regulatory agencies (FDA and European Medicines Agency). Google was used to search for additional internet-based materials. Refer to Appendix 1 for more information on the grey literature search strategy.

Findings From the Literature

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

Figure 1: Flow Diagram for Inclusion and Exclusion of Studies

Table 6: Details of the Included Study

C1D1 = cycle 1 day 1; DOR = duration of response; ECOG PS = Eastern Cooperative Oncology Group Performance Status; EORTC QLC-CIPN20 = European Organisation for Research and Treatment of Cancer 20-item Quality of Life Questionnaire for Chemotherapy-Induced Peripheral Neuropathy; EORTC QLQ-C30 = European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30; EQ-5D-5L = EQ-5D 5-Levels; FLC = free light chain; HRQoL = health-related quality of life; IgA = immunoglobin A; IMWG = International Myeloma Working Group; IRC = independent review committee; MM = multiple myeloma; ORR = overall response rate; ORR1 = overall response rate for patients treated with selinexor plus bortezomib plus dexamethasone after crossover; OS = overall survival; PFS = progression-free survival; PFS1 = progression-free survival for patients treated with selinexor plus bortezomib plus dexamethasone after crossover; PFS2 = progression-free survival for patients who received treatment after treatment with selinexor plus bortezomib plus dexamethasone, bortezomib plus dexamethasone, or selinexor plus bortezomib plus dexamethasone after crossover; PI = proteosome inhibitor; PN = peripheral neuropathy; PR = partial response; R-ISS = Revised International Staging System; SC = subcutaneous; SINE = selective inhibitor of nuclear export; TTNT = time to next treatment; TTR = time to response.

Note: Two additional reports were included.^15,16

^aSymptomatic ischemia.

Source: BOSTON Clinical Study Report.⁴

Description of Studies

One multi-centre, phase III, active-controlled open-label study, BOSTON, was included in this CADTH review. The objective of the BOSTON trial was to compare the efficacy, HRQoL and safety of SVd with those of Vd in adult patients with RRMM who received 1 to 3 prior anti-MM regimens. BOSTON was conducted at 123 sites across 21 countries including Canada. Patients were randomized to receive SVd or Vd in a 1:1 ratio and stratified based on prior PI therapy (yes versus no), number of prior anti-MM regimens (1 versus > 1). Patients were randomized within each of 4 geographic regions. At screening, patients’ R-ISS stage was assessed (R-ISS stage III versus R-ISS stage I or II); patients for whom R-ISS staging was not possible due to unavailable data for chromosomal abnormalities and serum lactate dehydrogenase were assigned to the R-ISS category corresponding to their ISS stage. Evaluations for MM were conducted at baseline, on day 1 of each cycle every 3 weeks through to week 37 to identify patients who progress quickly, and then every 5 weeks thereafter for the remainder of the study regardless of patient cycle length. This ensured comparable PFS data for both the SVd and Vd treatment groups. Patients who progressed, based on assessment by IRC, while receiving Vd were eligible to cross over to receive either SVd (SVdX) or selinexor plus dexamethasone (SdX) for those who were intolerant to bortezomib. Patients who received a selinexor-based regimen after crossover underwent MM evaluations every 5 weeks. Patient responses were graded according to IMWG response criteria.¹⁷ Per IMWG, response may be confirmed if the patient fails to provide a 24-hour urine sample after screening activities occur. To confirm a response, 2 consecutive assessments were needed. All MM disease assessments were required to be performed at the study visits pre-specified in the study protocol.

The end-of-study period was when the last patient completed their last survival follow-up. Completion of follow-up for the last patient was to occur when the last patient in the study was followed for up to 5 years after patients received their last dose of study treatment (including the assigned study therapy as well as selinexor-based regimens after crossover), withdrew their consent, were withdrawn from the study by the investigator, died, or had been lost to follow-up, whichever occurred first.

Populations

Inclusion and Exclusion Criteria

A complete summary of eligibility criteria for the BOSTON trial is supplied in Table 6. Briefly, inclusion criteria included adult patients with histologically confirmed MM with measurable disease per IMWG guidelines and who had received between 1 and 3 prior anti-MM regimens. Patients had to have documented evidence of progressive MM on or after their most recent regimen. Patients previously treated with bortezomib or another PI were eligible if certain criteria were met (Table 6). Patients must also have had an ECOG PS of no more than 2. Exclusion criteria included previous exposure to SINE compounds, including selinexor, previous malignancies requiring treatment, or evidence of recurring, and uncontrolled comorbidities. Patients could not have a PN greater than grade 2, or a PN of grade 2 or higher with pain at baseline, regardless of whether or not they were receiving medication.

Baseline Characteristics

A summary of baseline characteristics for all patients in the BOSTON trial is provided in Table 7. In general, characteristics across both the SVd and Vd treatment groups were well balanced. The mean age of patients was 65 years (SD = 9.56) in the SVd group and 67 years (SD = 9.35) in the Vd group, with most patients being aged 51 to 64 years (36% in the SVd group versus 31% in the Vd group), 65 to 74 years (39% versus 41%, respectively), or 75 years or older (17% versus 23%); fewer patients were between 18 and 50 years (8% versus 5%). There was a slightly greater proportion of males enrolled in the trial (59% versus 56% in the SVd and Vd groups, respectively). Most patients were White (83% versus 80% in the SVd and Vd groups, respectively) and not Hispanic or Latino (88% versus 91%), never smokers (73% versus 74%), and had an ECOG PS of 0 (35% versus 37%) or 1 (54% versus 55%), a mean creatinine clearance at baseline of greater than 60 mL/min (71% versus 66%), and a status of nonfrail at baseline (66% versus 69%).

Regarding disease stage, approximately one-quarter of patients (25% in the SVd group versus 27% in the Vd group) were diagnosed with stage I disease at diagnosis, compared to one-third who were diagnosed with stage II (32% versus 27%, respectively), and one-third with stage III (29% versus 32%, respectively). More than half of all patients had kappa light chain type of the active myeloma at baseline (56% versus 61% in the SVd group and Vd group, respectively). The R-ISS stage at screening was stage I for 29% of patients in the SVd group versus 25% in the Vd group, stage II for 60% of patients in both groups, and 6% and 7%, respectively, for stage III. Approximately half of patients had a high-risk chromosomal abnormality, with most being 1q21 (41% versus 34% in the SVd and Vd groups, respectively) compared to t(4;14) (11% versus 14%), del(17p)/p53 (11% versus 8%), or t(14;16) (4% versus 5%). The mean number of prior lines of anti-MM therapy was 1.7 in both treatment groups; 51% versus 48% of patients in the SVd and Vd groups, respectively, had 1 prior line of therapy, compared to 33% and 31% of patients with 2 prior lines of anti-MM therapy, and 16% versus 21% of patients with 3 prior lines of anti-MM therapy.

Most patients had received prior PI therapy (76% in the SVd group versus 77% in the Vd group). Other treatments to which patients had been previously exposed included bortezomib (69% versus 70% in the SVd and Vd groups, respectively), lenalidomide (40% versus 37%), carfilzomib (10% in both groups), pomalidomide (6% versus 3%), daratumumab (6% versus 3%), and ixazomib (3% versus 1%). Slightly more patients in the SVd group received a stem cell transplant (39%) than in the Vd group (30%).

A summary of baseline characteristics for patients in the crossover populations (SVdX and SdX) are reported in Table 8. There were no differences in baseline characteristics between the SVdX and SdX groups. Baseline characteristics also generally matched those of the total population. However, the mean number of prior lines of anti-MM therapies was greater in the crossover population (mean prior anti-MM therapies = 3); this is expected as these patients would have received treatment in the BOSTON trial as well, which would increase the average number of therapies for this group compared to patients who entered the BOSTON trial at baseline.

Table 7: Summary of Baseline Characteristics (Intention-to-Treat Population)

ECOG = Eastern Cooperative Oncology Group; MM = multiple myeloma; PI = proteosome inhibitor; R-ISS = Revised International Staging System; SD = standard deviation; SVd = selinexor plus bortezomib plus dexamethasone; Vd = bortezomib plus dexamethasone.

Source: BOSTON Clinical Study Report.⁴

Table 8: Baseline Characteristics by Crossover Population

MM = multiple myeloma; SD = standard deviation; SdX = selinexor plus dexamethasone after crossover; SVdX = selinexor plus bortezomib plus dexamethasone after crossover.

Source: BOSTON Clinical Study Report.⁴

Interventions

Patients in the intervention group received SVd. Each treatment in this regimen was administered in the following doses:

• Selinexor administered at 100 mg (5 tablets of 20 mg each) orally on days 1, 8, 15, 22, and 29 of each cycle. Each cycle was 35 days.

• Bortezomib administered at a dose of 1.3 mg/m² SC on days 1, 8, 15, and 22 of each cycle. Each cycle was 35 days.

• Dexamethasone administered at a dose of 20 mg orally on days 1, 2, 8, 9, 15, 16, 22, 23, 29, and 30. Each cycle was 35 days.

Selinexor tablets were administered orally with at least 120 mL (4 ounces) of water and could be taken with or without food once daily. The doses of the selinexor-based combination were determined based on results from a phase II trial (STOMP).

Patients in the comparator group received Vd. Each treatment in this regimen was administered in the following doses:

• Bortezomib administered at a dose of 1.3 mg/m² SC on days 1, 4, 8, and 11 of each cycle for the first 8 cycles. Each cycle consisted of 21 days; after cycle 8, bortezomib was administered at a dose of 1.3 mg/m² SC on days 1, 8, 15, and 22 of each cycle; each cycle was 25 days.¹⁸

• Dexamethasone administered at a dose of 20 mg rally on days 1, 2, 4, 5, 8, 9, 11, and 12 of each cycle for the first 8 cycles; each cycle consisted of 21 days; after cycle 8, dexamethasone was administered at a dose of 20 mg orally on days 1, 2, 8, 9, 15, 16, 22, 23, 29, and 30 of each cycle; each cycle was 25 days.¹⁸

Where possible, each study treatment (selinexor, bortezomib, and dexamethasone), was generally taken 1 to 2 hours apart. Dexamethasone was to be administered at least an hour before selinexor and bortezomib. Disease progression was required to be confirmed by IRC, unless patients were medically contraindicated; an exception was made for patients randomized to the Vd group who terminated bortezomib treatment before IRC-confirmed disease progression if the decision to stop bortezomib treatment was due to significant toxicities, such as PN, and all treatment measures addressing these toxicities were exhausted and documented before bortezomib termination. The decision to terminate bortezomib early for these patients was discussed and approved by a medical monitor to allow crossover to SdX after progression was confirmed by IRC.

Treatment could be discontinued if it was decided by the investigator or patient to discontinue, or due to pregnancy, unacceptable AEs or toxicity that could not be managed by supportive care, withdrawal of consent, death, or the study sponsor’s decision to terminate the study. After patients experienced disease progression, patients in the SVd group completed their end-of-treatment visit and were followed for survival. Patients in the Vd group either crossed over or discontinued study treatment, completed their end-of-treatment visit, and were followed for survival.

Crossover

After disease progression, per IRC assessment, patients in the Vd group were eligible to cross over to receive SVd (referred to as SVdX) or selinexor plus dexamethasone (SdX) if they were intolerant to bortezomib. The following procedures were put in place to prevent premature crossover:

• Investigators assessed for disease progression according to IMWG criteria, which included repeated testing if disease progression was based on serum and/or urine M protein, quantitative immunoglobulins for immunoglobin A or D, or serum free light chain (FLC). Disease progression could also be based on new or enlarging plasmacytoma(s) or bone lesion(s) or on other symptoms and signs of clinical progression that met the IMQG criteria.

• All cases of disease progression were confirmed by an IRC before crossover.

• Crossover was not permitted based purely on disease progression assessed by the investigator that did not meet any IMWG criteria and which could not be verified by IRC (e.g., deteriorating performance status).

• Crossover was not permitted if dosing of bortezomib was terminated before disease progression was confirmed by an IRC, unless termination of bortezomib was due to significant toxicities (i.e., PN), and all treatment measures addressing these toxicities were exhausted and documented before bortezomib termination. Early termination of bortezomib was discussed and approved by the sponsor’s medical monitor to allow crossover to SdX after progression was confirmed by an IRC.

• Investigator-assessed presumptive events of progressive disease that were not confirmed by the IRC had their PFS censored at the time of treatment discontinuation.

Patients receiving SVdX would return to cycle 1 for SVd treatments at the doses of the SVd group, and undergo MM evaluations every 5 weeks. Patients receiving SdX returned to cycle 1 for treatment with selinexor plus dexamethasone and underwent MM evaluations every 5 weeks. Doses for the SdX group were the same as the doses for selinexor and dexamethasone in the SVd group.¹⁸

Dose Escalation

Patients being treated with a selinexor-based regimen (SVd, SVdX, or SdX) could have been considered for dose escalation for selinexor if they met the following criteria: did not achieve at least a PR within the first 2 cycles, were tolerating SVd well at dose level 0, and did not have any AEs of grade 2 or higher related to study treatment at the time of dose escalation.¹⁸

For cycles 3 and beyond, selinexor could be increased to a fixed oral dosage of 60 mg twice weekly during weeks 1 through 5 of each cycle. Dexamethasone (20 mg) was given twice weekly on the same days as selinexor.¹⁸ A total of 45 patients (23.1%) in the SVd group underwent dose escalation with selinexor.¹⁹

Patients could also have undergone dose escalation with Vd. Ten patients (5.1%) in the SVd group and 6 patients (2.9%) in the Vd group underwent dose escalation to bortezomib. Six patients (3.1%) in the SVd group and 9 patients (4.4%) in the Vd group underwent dose escalation to dexamethasone.

Dose Modifications

Dose modifications were permitted for selinexor, bortezomib, and dexamethasone to manage tolerability.

Dose reductions were conducted based on a pre-specified dose-adjustment guide for selinexor related to AEs, including fatigue, anorexia or weight loss, acute nausea, hyponatremia, diarrhea, thrombocytopenia, neutropenia, anemia, tumour lysis syndrome, and other selinexor-related AEs. For some AEs, dose interruptions were recommended rather than reductions.¹⁸ Procedures for dose modifications of selinexor are specified in Table 9.

Table 9: Pre-Specified Dose Modifications for Adverse Events Related to Selinexor

b.i.w. = twice weekly; q.w. = once weekly; SVd = selinexor plus bortezomib plus dexamethasone.

^aDose level of + 1 may be considered for patients who meet the following 3 criteria: 1) they do not achieve at least a partial response after the first 2 cycles of SVd; and 2) they are tolerating SVd well at dose level 0; and 3) they do not have any AEs related to study treatment of grade 2 or worse as defined by the National Cancer Institute Common Terminology Criteria for Adverse Events version 4.03 at the time of dose escalation.

Source: BOSTON Protocol.¹⁸

Dose modifications for bortezomib related to PN resulted in a change from a dosage of 1.3 mg/m² twice weekly to once weekly. Other toxicities requiring dose modifications of bortezomib were conducted according to guidelines in the prescribing information for bortezomib. Doses of dexamethasone were to remain constant throughout the study. However, patients who had a partial intolerance to dexamethasone could have had a dosage reduction to a minimum dose of 10 to 12 mg twice weekly.¹⁸

Dose Modifications for Overlapping Toxicities

It was possible for toxicities related to both selinexor and bortezomib to overlap; in these cases, it was strongly recommended that the investigator reduce the dose or delay 1 drug at a time. Thrombocytopenia and neutropenia were considered to have potentially overlapping toxicities for selinexor plus bortezomib. For patients who experienced drug-induced thrombocytopenia and/or neutropenia while they were receiving any study regimen, attempts were made to determine which drug may be responsible and to treat appropriately, with dose modifications if necessary. For cases where the cause of the AE could not be attributed to a single drug, it was strongly recommended that the investigator reduce the dose or delay 1 of the drugs at a time.¹⁸

Missed or Vomited Doses

Pre-specified instructions were provided for patients who missed doses. For patients in the Vd group who missed a dose of bortezomib, their schedules were altered to accommodate 2 doses in that week, with at least 72 hours between 2 consecutive doses of bortezomib. For patients who missed a dose of any study treatment, they were to simply take their next dose as scheduled.¹⁸

Patients who missed doses of treatment for reasons unrelated to the study protocol (e.g., a required medical procedure or an unanticipated personal emergency) were to replace the missed doses on the following cycle. If selinexor doses are vomited within 1 hour of ingestion, the dose was to be replaced. If vomiting occurred more than 1 hour after dosing, it was considered a complete dose.¹⁸

Concomitant Medications

Concomitant medications to treat symptoms, AEs, and intercurrent illnesses were permitted such that they were a necessary part of standard of care. Medications to treat concomitant illnesses such as diabetes and hypertension were permitted. If clinically indicated, patients could receive transfusions of red blood cell or platelets.¹⁸

Restrictions

On the days when selinexor was administered, the use of acetaminophen or products containing acetaminophen was limited to a total daily dose of 1 g; no restrictions were placed on the use of acetaminophen-containing products for other days. Patients were not permitted to take products containing glutathione, S-adenosylmethionine, or N-acetylcysteine during their participation in the BOSTON trial as these products may enhance the metabolism of selinexor; however, patients could use these products if the patient had elevated liver function tests.¹⁸

Outcomes

A list of efficacy and safety end points identified in the CADTH review protocol that were assessed in the clinical trials included in this review is provided in Table 10, and these end points are further summarized in the following section. A detailed discussion and critical appraisal of the outcome measures is provided in Appendix 4.

Table 10: Summary of Outcomes of Interest Identified in the CADTH Review Protocol

CR = complete response; DOR = duration of response; EORTC QLQ-CIPN20 = European Organisation for Research and Treatment of Cancer 20-item Quality of Life Questionnaire for Chemotherapy-Induced Peripheral Neuropathy; IMWG = International Myeloma Working Group; IRC = independent review committee; MRD = minimal residual disease; NA = not applicable; ORR = overall response rate; OS = overall survival; PFS = progression-free survival; PR = partial response; sCR = stringent complete response SdX = selinexor plus dexamethasone after crossover; SVd = selinexor plus bortezomib and dexamethasone; SVdX = selinexor plus bortezomib plus dexamethasone after crossover; VGPR = very good partial response.

Source: BOSTON Protocol.

Safety

Adverse events were graded by severity according to the National Cancer Institute (NCI) Common Terminology Criteria for Adverse Events (CTCAE) version 4.03. For AEs that did not have a specific CTCAE grade, severity was characterized as mild (transient and not interfering with the patient’s daily activities), moderate, severe, or life-threatening, indicating grade 1 (usually transient and do not interfere with the patient’s daily activities), 2 (introduces a low level of inconvenience or concern to the patient and may interfere with daily activities), 3 (interrupts the patient’s usual daily activities), and 4 (life-threatening) AEs, respectively. Adverse events of special interest were reported by the sponsor for selinexor. Those of special interest for selinexor included cataracts and acute cerebellar syndrome. Adverse events were recorded through to the end of the study for at least 30 days following the last dose of study treatment, or until resolution.

The incidence of any PN events of grade 2 or higher was a safety end point, and 1 of the key secondary end points of the BOSTON trial. Non-key secondary safety end points included safety and tolerability of study treatments based on AE reports, physical examination results (including vital signs), ECOG PS scores, 12-lead electrocardiogram results, ophthalmic results, and clinical laboratory tests.

Statistical Analysis

Results from the BOSTON trial were reported for 2 data cut-offs: February 28, 2020 (primary analysis) and February 15, 2021 (updated analysis). The primary analysis was based on a pre-specified interim analysis, which is described in the following section. An additional analysis after the second interim analysis was conducted on February 15, 2021, at the request of the Committee for Medicinal Products for Human Use (CHMP). At this updated-item point, analyses of the primary end point, key secondary end points, and safety data (e.g., PFS, DOR, OS, ORR, AEs, SAEs, AEs of clinical interest, and deaths) were also analyzed. The updated analysis was considered noninferential and P values from this analysis were considered nominal.¹⁹

Sample Size

The sample size was determined to have an 80% power to detect a median time to PFS of 13.5 months for patients treated with SVd versus 9.4 months for patients treated with Vd using a 1-sided alpha of 0.025; this allowed for 15 months of patient accrual and 18 months of follow-up. The median time to PFS of 9.4 months in the Vd group was based on recent ENDEAVOR and CASTOR clinical studies; the eligibility criteria of these studies were considered to be similar to those of the BOSTON trial, with a PFS of 9.4 months in the ENDEAVOR trial and 7.2 months in the CASTOR trial. The median time to PFS in the SVd group was based on preliminary results from the ongoing STOMP trial. Treatment difference was assessed via a log-rank test, which found that 267 PFS events were required for the final analysis. A total of 364 patients was required for enrolment, or approximately 182 patients in each treatment group. An exponential dropout rate of 0.65% per month was assumed, which is equivalent to a dropout rate of approximately 10% after 18 months.

Interim Analyses

Two interim analyses for PFS for futility or superiority were pre-specified (Figure 2). The first interim analysis was planned to be performed for sample size re-estimation. The DSMB reviewed overall survivorship for PFS at this first interim analysis to determine if the sample size, recruitment period, or duration of the follow-up needed to be adjusted to ensure timely achievement of the 267 PFS events required for the final analysis. This first interim analysis was planned to occur after approximately 30% (i.e., 81) of the PFS events had occurred. The unblinded sample size re-estimation was based on the condition power. The sample size re-estimation analyses were conducted by an independent statistician external to the sponsor of the BOSTON trial, and the results were provided to the DSMB, which reviewed the results and then made a recommendation to the sponsor.

The second interim analysis for PFS was planned to occur after approximately 75% of PFS events (i.e., approximately 201 PFS events) and 22 months after the start of the study to allow for patient accrual to be completed before this interim analysis. The results of the interim analysis would allow for a conclusion of significant efficacy at a P value less than or equal to 0.0072, and stopping for futility (non-binding) at a P value greater than or equal to 0.0598. The timing of this second interim analysis was to allow for sufficient follow-up time following the end of patient accrual so that any delayed treatment effect may become sufficiently attenuated. There was a plan to adapt the study at this interim analysis to estimate the treatment effect. The type I error for this second interim analysis was maintained using the Cui, Hung, and Wang (CHW) method if the unblinded sample size re-estimation was conducted. At the time of this second interim analysis, the 3 key secondary end points were also tested: ORR, incidence of PN events of grade 2 or higher, and response rates for responses of a VGPR or better based on an IRC assessment. To maintain the overall type I error of these 3 tests at a 1-sided significance level of 0.025, a hierarchical testing procedure was conducted; testing of these end points was not conducted until the primary end point of PFS reached statistical significance.

The alpha level for the final analysis of PFS was adjusted for the planned interim analyses using the Lan-DeMets alpha spending function with the O’Brien-Fleming type of boundary. If, at the first interim analysis, the unblinded sample size re-estimation was conducted, the CHW method was also used for the final analysis of PFS.

Figure 2: Overview of the BOSTON Trial Interim Analyses

IA = interim analysis; IRC = independent review committee; PD = progressive disease; PFS = progression-free survival; SdX = selinexor plus dexamethasone after crossover; SVdX = selinexor plus bortezomib plus dexamethasone after crossover; Vd = bortezomib plus dexamethasone.

Note: The second interim analysis was based on the data cut-off date of February 18, 2020.

Source: BOSTON Clinical Study Report.⁴

Handling of Missing Data

In general, for all efficacy end points no substitutions for missing data points were made. For HRQoL measurements, missing data were handled as described in the scoring manuals for each of the EORTC QLQ-C30, EORTC QLQ-CIPN20, and EQ-5D-5L. No other imputations for missing efficacy data were planned. Patients without efficacy evaluations were considered censored at time 0 for time-to-event analyses. For AEs, missing dates were outlined according to rules outlined in the statistical analysis plan.¹⁹

Analysis of End Points

Unless otherwise specified, all analyses of efficacy end points were conducted with the intention-to-treat (ITT) population. Summary statistics (i.e., n, median, mean, SD, minimum, and maximum) were computed for continuous variables. Frequencies and percentages were presented for categorical variables. The Kaplan-Meier method was used for time-to-event variables. Graphical displays were provided when appropriate.

The primary end point of the BOSTON trial was PFS. Key secondary end points included ORR, response rates at any time before disease progression or death due to any cause, and incidence of any grade 2 or higher PN events. To maintain the overall type I error at a 1-sided 0.025 level of significance, the ORR was tested under a hierarchical testing procedure. Key secondary end points were tested in the following sequence: ORR, incidence of any grade 2 or higher PN events, and response rates of a VGPR or better based on an IRC assessment.

Non-key secondary end points included OS; DOR; ORR for SVdX patients only (ORR1); PFS for SVdX patients only (PFS1); TTNT, TTR, and PFS for patients who received treatment after SVd, Vd, or SVdX 2 (PFS2), and HRQoL measured using the EORTC QLQ-CIPN20.

Exploratory end points included TTD of SVd and Vd treatment, HRQoL measured using the EORTC QLQ-C30, and the EQ-5D-5L, disease response to SdX treatment according to the IMWG response criteria.

Progression-Free Survival

As clinical deterioration without an increase in M protein levels was not considered disease progression, patients who ended study treatment due to clinical deterioration without an increase in M protein levels were censored for analysis of PFS. Unless otherwise specified, relapse from CR by positive immunofixation or trace amount (defined as < 0.5 g/dL of M protein) was not considered to be disease progression. A summary of PFS used Kaplan-Meier methodology and 25th, 50th, and 75th percentiles with associated 2-sided 95% CIs. Percentages of censored observations and proportions of events were reported. Patients who did not experience a PFS event were censored at the date of their last disease assessment. A stratified log-rank test was used to compare the PFS distributions between treatment groups for the primary efficacy assessment. Analyses were stratified by prior PI therapy, number of prior anti-MM regimens, and the R-ISS stage at study entry. Hazard ratios were estimated using a stratified Cox proportional hazards model with treatment as the only factor and stratification based on the same stratification variables used for randomization.

A supportive analysis was planned to be conducted for PFS using the per-protocol population.

The following sensitivity analyses were conducted for PFS:

• a sensitivity analysis used a nonstratified log-rank test and a Cox proportional hazards model

• PFS events were defined as documented progression as verified by IRC or death when the patient was closely followed, whichever occurred first; patients were censored at the data of their last disease assessment if no there was no progression, as confirmed by IRC, treatment was discontinued for any reason, new anticancer treatment was started, or death or progression occurred after 2 or more missed visits

• similar to the primary analysis for PFS, except treatment discontinuation for any reason was counted as an event

• similar to the primary analysis for PFS, except the initiation of non-study antineoplastic therapy was counted as an event

• similar to the primary analysis for PFS, except clinical progression, defined as when a patient discontinues the treatment with reason of disease progression but the event is not classified as disease progression by IRC, was counted as an event in addition to IRC-confirmed disease progression

• similar to the primary analysis for PFS described above, except the timing of the IRC-confirmed disease progression at an unscheduled visit is changed to the net scheduled visit

• analysis of PFS based on assessment by the investigator.

Overall Response Rate

ORR was 1 of the key secondary end points tested hierarchically upon statistical significance of PFS at the time of the second interim analysis. A Cochrane-Mantel-Haenszel (CMH) test stratified by prior PI therapy, number of prior anti-MM regimens, and the R-ISS stage at study entry were used to compare ORRs between the 2 treatment groups. The Breslow-Day test was used to evaluate the homogeneity of ORs across the strata associated with ORR. Patients missing MM disease assessments after cycle 1 day 1 (C1D1) were imputed as nonresponders. Forest plots of estimated ORs were provided for each stratification factor. The analysis of the response rate for responses of a VGPR or better were based on assessments by the IRC and performed in a similar manner to the primary efficacy end point of ORR using the CMH test.

A supportive analysis was planned to be conducted for ORR using the per-protocol population.

The following sensitivity analysis was conducted for ORR:

• Patients who did not have the opportunity to complete at least 2 post-C1D1 MM evaluations were considered nonresponders.

Incidence of Peripheral Neuropathy Events of Grade 2 or Higher

This analysis for PN was 1 of the key secondary end points tested hierarchically upon statistical significance of PFS at the time of the second interim analysis. A CMH test stratified by randomization stratification factors was used to compare the differences in incidence of PN events of grade 2 or higher between the 2 treatment groups. This analysis was conducted using the safety population. The number and percentage of events were summarized by each treatment group along with ORs and associated 95% CIs. The Breslow-Day test was used to assess homogeneity of ORs across strata. Similar analyses were conducted for assessing the incidence of PN events of grades 2, 3, and 4 separately.

The following sensitivity analysis was conducted:

• All PN events of grade 2 or higher that occurred for patients in the Vd group, regardless of crossover status, were included in the sensitivity analysis.

Response Rate for Responses of VGPR or Better Based on IRC Assessment

This end point was 1 of the key secondary end points tested hierarchically upon statistical significance of PFS at the time of the second interim analysis. The analysis for this end point was performed in a manner similar to that of the analysis for ORR using the CMH test. The unadjusted number and percentage of patients were summarized for each treatment group with associated 95% CIs. The Breslow-Day test was used to assess homogeneity of ORs across the strata. Only responses better than a VGPR that occurred before IRC-confirmed progressive disease or initiating a new MM treatment were included in the analysis.¹⁹

Overall Survival

Analyses for OS were performed by treatment group using a stratified log-rank test stratified by prior PI therapy, number of prior anti-MM regimens, and the R-ISS stage at study entry. The Kaplan-Meier method was used to estimate medians and associated 95% CIs. Kaplan-Meier curves were graphically displayed where appropriate.¹⁹

Duration of Response

Analyses for DOR were conducted in the same manner as OS.¹⁹

Time to Next Treatment

Analyses for TTNT were conducted in the same manner as OS.¹⁹

Time to Response

Analyses for TTR were conducted in the same manner as OS.¹⁹

Objective Response Rate for SVdX Patients Only (ORR1)

The percentage of patients who achieved a confirmed PR or better (i.e., PR, VGPR, CR, or sCR) was tested assuming a null hypothesis fixed-threshold value of 10% against a 1-sided alternative hypothesis of greater than 10% using exact methods for a 1-sample binomial without stratification.¹⁹

Time to Discontinuation of SVd and Vd

Analyses were performed using a stratified log-rank test by treatment group, stratified by prior PI therapies, the number of prior anti-MM regimens, and the R-ISS stage at study entry. The Kaplan-Meier method was used to determine median TTD with 95% CIs; graphical representation of the Kaplan-Meier curve was provided for each treatment group. Patients were censored if they did not discontinue treatment; these patents were censored at the date of the last dosing on or before the data cut-off date.¹⁹

Disease Response to SdX treatment According to IMWG Response Criteria

The IRC-confirmed response according to IMWG response criteria was summarized as the number and percentage of patients achieving the following responses: sCR, CR, VGPR, PR, minimal residual disease, stable disease, progressive disease, and NE. The corresponding 95% exact CI for each response category were also provided.¹⁹

Health-Related Quality of Life

Health-related quality of life was measured using 3 questionnaires: the EORTC QLQ-CIPN20, EORTC QLQ-C30, and EQ-5D-5L. A brief explanation of each instrument is provided here, and further detail is provided in Appendix 4.

EORTC QLQ-CIPN20: The EORTC QLQ-CIPN20 is a 20-item tool to elicit a patient’s experience of symptoms and functional limitations related to chemotherapy-induced PN. The questionnaire contains 3 subscales, including sensory (9-item), motor (8-item), and autonomic (3-item) subscales. Responses indicate the degree of sensory, motor, or autonomic symptoms during the past week as measured on a 4-point Likert scale, with 1 indicating “not at all” and 4 indicating “very much.” Responses are converted to a 0-to-100 scale, with higher scores indicating greater symptom burden. The scale was only calculated if a minimum of 50% of items (i.e., 5 of 9 items or 4 of 8 items) from the subscale were answered. Descriptive statistics were used to measure the value and change from baseline before initiating a new MM treatment for each of the 3 subscales. A linear-effects model was also used to analyze the change from baseline with treatment group as the fixed effect; analysis was stratified by prior PI therapies, the number of prior anti-MM regimens, and the R-ISS stage at study entry. The baseline value of the corresponding subscale score was included as a covariate, as well as the random effect of patients and repeated measures over time points. Adjusted mean changes of each treatment group and treatment differences were presented with 95% CIs and P values. The P values were considered nominal.¹⁹

EORTC QLQ-C30: The instrument consists of 5 functional scales, 3 symptom scales, a GHS/QoL scale, and 6 single items. The actual value and change from baseline before initiating a new MM treatment was summarized using descriptive statistics over time for each scale: functional scale, symptom scale, GHS/QoL scale, and single items.¹⁹

EQ-5D-5L: The actual value and change from baseline before initiating a new MM treatment were summarized using descriptive statistics over time for EQ-5D-5L health states and VAS scores.¹⁹

Analysis Populations

A summary of the analysis populations used in the BOSTON trial is provided in Table 11.

Table 11: Analysis Populations in the BOSTON Trial

IRC = independent review committee; ITT = intention-to-treat; ORR1 = objective response rate for patients receiving selinexor plus bortezomib plus dexamethasone after crossover; PFS = progression-free survival for patients receiving selinexor plus bortezomib plus dexamethasone after crossover; PP = per-protocol; SdX = selinexor plus dexamethasone after crossover; SVdX = selinexor plus bortezomib plus dexamethasone after crossover; Vd = bortezomib plus dexamethasone.

Source: BOSTON Clinical Study Report.⁴

Results

Patient Disposition

A summary of patient disposition for all patients in the BOSTON trial is provided in Table 12. A total of 195 patients were randomized to the SVd group and 207 patients were randomized to the Vd group. All patients in the SVd group received their allocated treatment; only 3 patients in the Vd group did not receive the allocated treatment due to withdrawal of consent, death, and AEs (N = 1 for each). At the primary analysis, there were no differences in patients discontinuing from the study between treatment groups (81.0% in the SVd group versus 82.4% in the Vd group). The reasons for discontinuation were mainly due to disease progression (34.4% versus 52.5%, respectively), withdrawal by the patient (19.0% versus 8.8%, respectively) and AEs and/or toxicity (16.9% versus 11.3, respectively). A greater proportion of patients discontinued due to disease progression in the Vd group, but a greater proportion of patients in the SVd group discontinued due to withdrawal from the patient and AEs and/or toxicity. At the time of the primary analysis, more patients in the SVd group than in the Vd group were followed up for survival (33.3% versus 25.0%, respectively). Similar proportions of patients in both treatment groups had discontinued from the BOSTON trial (47.7% versus 50.5% in the SVd and Vd groups, respectively); the primary reasons for withdrawal were death (24.1% versus 29.9%) and withdrawal by the patient (19.0% versus 17.2%).

Table 12: Patient Disposition of All Patients

AE = adverse event; ITT = intention-to-treat; NA = not applicable; PP = per-protocol; SdX = selinexor plus dexamethasone after crossover; SVd = selinexor plus bortezomib plus dexamethasone; SVdX = selinexor plus bortezomib plus dexamethasone after crossover; Vd = bortezomib plus dexamethasone.

^aPatients in follow-up who were willing to continue into survival follow-up, still alive, not lost to follow-up, and had not discontinued from the study.

^bPatients who have been followed for 5 years after the last dose of SVd, Vd, SVdX, or SdX treatment.

^cPatients in the safety population who cross over from the Vd arm to the SVdX treatment and have received at least 1 dose of selinexor.

^dPatients in the safety population who cross over from the Vd arm to the SdX treatment and have received at least 1 dose of selinexor.

Source: BOSTON Clinical Study Report.⁴

At the time of the updated analysis, few patients from both treatment groups were still on treatment as the majority had discontinued from the trial, with 89.2% discontinuing from the SVd group and 92.2% discontinuing from the Vd group. As with the primary analysis, the primary reasons for discontinuation were disease progression (29.0% versus 57.8% for the SVd and Vd groups, respectively), withdrawal by the patient (19.0% versus 10.2%, respectively), and AEs and/or toxicity (16.9% versus 12.7%, respectively). More patients in the Vd group discontinued due to disease progression, while more patients in the SVd group discontinued due to withdrawal by the patient and AEs and/or toxicity. Similar proportions of patients were being followed for survival in the SVd (26.7%) and Vd (28.4%) groups. A total of 73 patients (27.4%) in the SVd group and 78 patients (38.2%) in the Vd group were still in the study by the updated analysis. The primary reasons for discontinuation were withdrawal by the patient (20.0% versus 19.1%, respectively) and death (34.9% versus 38.7%, respectively).

It should be noted that none of the patients enrolled in the BOSTON trial had completed the 5-year survival follow-up at either the primary or updated analysis.

A summary of the disposition of crossover patients (SVdX and SdX groups) is provided in Table 13. The disposition of crossover patients was similar to all randomized patients.

Table 13: Patient Disposition of Crossover Patients

SdX = selinexor plus dexamethasone after crossover; SVd = selinexor plus bortezomib plus dexamethasone; SVdX = selinexor plus bortezomib plus dexamethasone after crossover; Vd = bortezomib plus dexamethasone.

^aPatients in follow-up were willing to continue into survival follow-up, still alive, not lost to follow-up, and had not discontinued from the study.

^bPatients who have been followed for 5 years after the last dose of SVd, Vd, SVdX, or SdX treatment.

Source: BOSTON Clinical Study Report.

Protocol Deviations

At the primary analysis (February 18, 2020), 22 major protocol deviations were reported for 21 patients due to issues involving informed consent (n = 17), eligibility criteria (n = 2), study conduct (n = 1), and investigational medicinal product (n = 1). At the updated analysis (February 15, 2021), an additional 9 major protocol deviations involving informed consent (n = 7), study drug dosing (n = 1), and investigational medicinal product (n = 1) were reported.

Protocol Amendments

Three protocol amendments were made to the original conduct of the BOSTON trial. The first protocol amendment was dated February 22, 2017. At this amendment, crossover for treatment to SdX was permitted as an option for patients in the Vd group after they experienced disease progression confirmed by an IRC if they had significant tolerability issues with bortezomib, such as a PN of greater than grade 2 or PN of 2 or higher with pain. At the third protocol amendment dated August 17, 2018, a series of changes were made to the total number of required PFS events:

• at the final analysis, the number of PFS events required changed from 284 events to 267

• for the interim analysis for sample size re-estimation, the number of PFS events required changed from 85 to 81

• for the interim analysis for futility or superiority, the number of PFS events required changed from 213 to 201.

Exposure to Study Treatments

Dose Intensity

A summary of patients’ exposures to treatments at the primary and updated analyses are reported in Table 14. The mean duration of study treatment exposure was similar between both treatment groups at both the primary and updated analyses. The mean duration of study treatment exposure remained similar between both treatment groups throughout the trial. The mean duration of selinexor exposure was 39.5 weeks (SD = 29.98) at the primary analysis and 47.2 weeks (SD = 44.00) at the updated analysis, with a mean number of 36 doses (SD = 28.06) at the primary analysis and 43 doses (SD = 41.55) at the updated analysis, and an average mean dose of 79 mg per week at both time points. The mean duration of bortezomib and dexamethasone exposure was similar between both treatment groups at both the primary and updated analyses. However, the average total dose and the average number of doses of bortezomib and dexamethasone were slightly lower in the SVd group than in the Vd group at both the primary and updated analyses.

The exposure to study treatment among crossover patients (SVdX and SdX groups) is reported in Table 15. The mean duration of study treatment exposure was 17.2 weeks (SD = 12.53) in the SVdX group and 11.5 weeks (SD = 7.06) in the SdX group.

Table 14: Exposure to Treatment (Safety Population)

NA = not applicable; SD = standard deviation; SVd = selinexor plus bortezomib and dexamethasone; Vd = bortezomib plus dexamethasone.

Note: For patients who cross over from the Vd Arm to the SVdX or SdX treatment, dosage information after the crossover is not included. Study treatment was selinexor with bortezomib and dexamethasone for the SVd arm. Study treatment was bortezomib with dexamethasone for the Vd arm.

^aAverage dose received per week is defined as total dose received divided by duration of exposure.

Source: BOSTON Clinical Study Report.⁴

Table 15: Exposure to Treatment for Crossover Patients (Safety Population; Primary Analysis)

NA = not applicable; SD = standard deviation; SdX = selinexor plus dexamethasone after crossover; SVdX = selinexor plus bortezomib plus dexamethasone after crossover.

^aAverage dose received per week is defined as total dose received divided by duration of exposure.

Source: BOSTON Clinical Study Report.⁴

Dose Modifications

In general, dose modifications, including reductions, delays, and interruptions, were required more frequently in the SVd group than in the Vd group (91.8% versus 82.4%, respectively). Dose modifications related to selinexor occurred in 89.2% of patients in the SVd group with a mean of 4.4 (SD = 3.38) dose modifications, and most patients required 3 or more (60.5%) dose modifications compared to only 1 (13.3%) or 2 (15.4%) dose modifications in the Vd group. There were no differences in dose modifications related to bortezomib in the SVd (84.1%) and Vd (80.4%) groups, with a mean of 3.5 and 3.7 dose modifications in each group, respectively. There were no differences in dose modifications related to dexamethasone in the SVd (80.5%) and Vd (76.0%) groups, with a mean of 3.4 and 3.5 dose modifications, respectively. In general, few patients missed treatment doses to any study treatment in both treatment groups.

There were more dose reductions for any study treatment in the SVd group (73.3%) compared to the Vd group (53.9%), although the mean number of dose reductions was similar between the 2 groups (2.7 versus 1.8, respectively). Patients in the SVd group also had a greater proportion of patients needing 2 or more dose reductions compared with the Vd group; 21.0% of patients in the SVd group required 2 dose reductions versus 14.2% of patients in the Vd group, and 34.9% versus 10.3% of patients required 3 or more dose reductions, respectively. A dose reduction of selinexor was reported for 64.6% of patients in the SVd group, with a mean of 1.8 (SD = 0.96) dose reductions; most patients required only 1 (27.7%) or 2 (24.1%) dose reductions, with 12.8% of patients reporting 3 or more reductions. There were no differences in reductions of bortezomib doses in the SVd and Vd groups (43.1% versus 44.6%, respectively), with similar proportions of patients reporting the number of reductions of bortezomib doses. Fewer dose reductions for dexamethasone were reported in the SVd group (27.2%) than in the Vd group (36.3%), but the mean number of dose reductions for dexamethasone was similar between the 2 groups (1.3 versus 1.4, respectively).

Dose delays or interruptions for any study treatment were more frequent in the SVd group (88.7%) than in the Vd group (76.0%). A dose delay or interruption for selinexor was reported for 86.7% of patients, with a mean of 3.2 dose reductions or interruptions (SD = 2.81) in the SVd group; most patients reported 2 (26.2%) or 3 or more (42.1%) dose delays or interruptions, with 18.5% reporting only 1 dose delay or interruption. A greater proportion of patients required a dose delay or interruption for bortezomib in the SVd group (81.5%) than in the Vd group (72.5%), although the mean number of dose delays or interruptions was similar between the 2 treatment groups (2.9 in the SVd group versus 3.2 in the Vd group). Dose delays or interruptions for dexamethasone were similar between the SVd and Vd groups (79.5% versus 72.5%, respectively) with a mean number of 3.0 dose delays or interruptions in both treatment groups.

More patients in the SVd group required dose escalation for any study treatment than in the Vd group (24.6% versus 7.4%, respectively); this was expected, as a total of 45 patients (23.1%) in the SVd group underwent a dose escalation with selinexor.¹⁹ There were no differences between the SVd and Vd groups in the proportion of patients requiring an escalation in doses of bortezomib (5.1% versus 2.9%, respectively) or dexamethasone (3.1% versus 4.4%).

Table 16: Dose Modifications (Safety Population)

NA = not applicable; SdX = selinexor plus dexamethasone after crossover; SD = standard deviation; SVd = selinexor plus bortezomib and dexamethasone; SVdX = selinexor plus bortezomib plus dexamethasone after crossover; Vd = bortezomib plus dexamethasone.

Note: For patients who cross over from the Vd Arm to the SVdX or SdX treatment, dosage information after the crossover is not included. Study treatment was selinexor with bortezomib and dexamethasone for SVd arm. Study treatment was bortezomib with dexamethasone for Vd arm. Dose modification includes dose reduction, dose delay and dose interruption.

Source: BOSTON Clinical Study Report.

Subsequent Therapy

A summary of subsequent therapies received by patients in the BOSTON trial is reported in Table 17. |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||.

Table 17: Subsequent Therapies (Primary Analysis)

NA = not applicable; SdX = selinexor plus dexamethasone after crossover; SVd = selinexor plus bortezomib plus dexamethasone; SVdX = SVd after crossover; Vd = bortezomib plus dexamethasone.

Source: Additional information from sponsor.⁵

Treatment Adherence

Adherence to treatment, which was measured as a percentage of the number of doses taken by patients compared to the number of doses scheduled, was high in both treatment groups; nearly all patients (99% in both groups) received scheduled doses of treatment. The high level of adherence was maintained through to the updated analysis. Adherence to treatment was similarly high among the subgroup of patients who crossed over (the SVdX and SdX groups).

Efficacy

Only those efficacy outcomes and analyses of subgroups identified in the review protocol are reported here. Refer to Appendix 3 for detailed efficacy data.

Overall Survival

At the time of the primary analysis, results for OS were based on a median follow-up time of 17.28 months (95% CI, 16.56 to 19.27) in the SVd group and 17.51 (95% CI, 17.08 to 18.23) in the Vd group (Table 18). Similar proportions of patients experienced death in the SVd (24.1%) and Vd (30.0%) treatment groups. The median OS was NE (95% CI, NE to NE) in the SVd group and 24.97 months (95% CI, 23.49 to NE) in the Vd group. The HR of death was 0.84 (95% CI, 0.57 to 1.23; 1-sided P = 0.1852, stratified log-rank test). At this time, 75 patients (36%) from the Vd group had crossed over to the SVdX or SdX groups (Figure 3).

At the updated analysis, results were based on a median follow-up of 28.71 months (95% CI, 27.24 to 29.90) in the SVd group and 28.65 months (95% CI, 27.63 to 29.67) in the Vd group, and were consistent with the primary analysis. In the SVd group, 34.9% of patients experienced an OS event compared to 38.6% of patients in the Vd group. Median OS was 36.67 months (95% CI, 30.19 to NE) in the SVd group and 32.76 (95% CI, 27.83 to NE) in the Vd group (HR = 0.88; 95% CI, 0.63 to 1.22) (Figure 4).

Table 18: Overall Survival (Intention-to-Treat Population)

CI = confidence interval; NE = not evaluable; SVd = selinexor plus bortezomib and dexamethasone; Vd = bortezomib plus dexamethasone.

Note: Overall survival is calculated from date of randomization to date of death. Patients without events were censored at the date of study discontinuation or date of last participating visit, whichever occurred first.

^aStratified for prior proteosome inhibitor therapies, number of prior anti–multiple myeloma regimens and Revised International Staging System stage at screening.

^bBased on a stratified Cox proportional hazard models with the Efron method of handling ties.

^cP value is considered nominal as results for the updated analysis were not pre-specified or controlled for multiplicity.

Source: BOSTON Clinical Study Report.⁴

Figure 3: Overall Survival (Intention-to-Treat Population; Data Cut-Off: February 18, 2020)

SVd = selinexor plus bortezomib and dexamethasone; Vd = bortezomib plus dexamethasone.

Source: BOSTON Clinical Study Report.⁴

Figure 4: Overall Survival (Intention-to-Treat Population; Data Cut-Off: February 15, 2021)

SVd = selinexor plus bortezomib and dexamethasone; Vd = bortezomib plus dexamethasone.

Source: BOSTON Clinical Study Report.⁴

Subgroup Analysis

Forest plots of all subgroup analyses are depicted in Figure 37. Results of subgroup analyses highlighted in the CADTH protocol are listed in Table 19. Subgroup analysis did not indicate any differences between the SVd and Vd in terms of OS.

Table 19: Subgroup Analysis of Overall Survival

CI = confidence interval; MM = multiple myeloma; NE = not estimable; PI = proteasome inhibitor.

^aBased on an unstratified model.

^bBased on a Cox proportional hazards model with the Efron method of handling ties.

Source: BOSTON Clinical Study Report.⁴

Progression-Free Survival per IRC Assessment

The median follow-up times were 13.17 months (95% CI, 10.64 to 15.34) for the SVd group and 16.53 months (95% CI, 14.39 to 17.71) for the Vd group (Table 20). At the primary analysis, a higher proportion of patients in the Vd group experienced a PFS event than did patients in the SVd group (59.9% versus 41.0%, respectively). The median PFS was longer in the SVd group (13.93 months; 95% CI, 11.73 to NE) compared to 9.46 months (95% CI, 8.11 to 10.78) in the Vd group. An HR of 0.70 (95% CI, 0.53 to 0.93) was reported for PFS, indicating an increase in PFS of 4.47 months and a 30% reduction in risk of disease progression or death in the SVd group compared to the Vd group (1-sided P = 0.0075, stratified log-rank test; Figure 5).

At the time of the updated analysis, the median follow-up time was 13.47 months (95% CI, 10.64 to 24.87) in the SVd group and 24.48 months (95% CI, 21.16 to 29.17) in the Vd group. The results at the updated analysis were consistent with those of the primary analysis. There was a higher proportion of PFS events in the Vd group compared to the SVd group (66.2% versus 47.2%, respectively) with a longer median PFS in the SVd group at 13.24 months (95% CI, 11.73 to 23.43) compared to the Vd group at 9.46 months (95% CI, 8.11 to 10.78); results indicated an increase of 3.78 months and a 29% reduction in risk of disease progression or death in the SVd group (HR = 0.71; 95% CI, 0.54 to 0.93; Figure 6).

A high proportion of patients were censored for these analyses of PFS. More patients in the SVd group were censored compared with patients in the Vd group at both the primary (59.0% versus 40.1%, respectively) and updated (52.8% versus 33.8%) analyses.

Sensitivity Analyses

Results of the sensitivity analyses are reported Table 46. In general, the results continued to support the primary analysis, which suggested that treatment with SVd was superior to Vd for PFS.

Table 20: Progression-Free Survival Based on IRC Assessment (ITT Population)

CI = confidence interval; IRC = independent review committee; ITT = intention-to-treat; MM = multiple myeloma; NE = not evaluable; SVd = selinexor plus bortezomib plus dexamethasone; Vd = bortezomib plus dexamethasone.

^aStratified for prior proteosome inhibitor therapies, number of prior of anti-MM regimens and Revised International Staging System stage at screening.

^bBased on a stratified Cox proportional hazards model with the Efron method of handling ties.

Source: BOSTON Clinical Study Report.⁴

Figure 5: Progression-Free Survival Based on IRC Assessment (ITT Population; Data Cut-Off: February 18, 2020)

IRC = independent review committee; ITT = intention-to-treat; SVd = selinexor plus bortezomib plus dexamethasone; Vd = bortezomib plus dexamethasone.

Source: BOSTON Clinical Study Report.⁴

Figure 6: Progression-Free Survival Based on IRC Assessment (ITT Population; Data Cut-Off: February 15, 2021)

IRC = independent review committee; ITT = intention-to-treat; SVd = selinexor plus bortezomib plus dexamethasone; Vd = bortezomib plus dexamethasone.

Source: BOSTON Clinical Study Report.⁴

Crossover Patients

Results for PFS of patients in the Vd group who crossed over to receive SVdX are provided in Table 21. A total of 36 patients (57.1%) experienced a PFS event. Twenty-seven patients (42.9%) were censored at the time of the analysis (Figure 7).

Table 21: Progression-Free Survival Based on IRC Assessment (SVdX Population)

IRC = independent review committee; SVdX = selinexor plus bortezomib plus dexamethasone after crossover.

Source: BOSTON Clinical Study Report.⁴

Figure 7: Progression-Free Survival Based on IRC Assessment of SVdX Patients

IRC = independent review committee; SVdx = selinexor plus bortezomib plus dexamethasone after crossover; Vd = bortezomib plus dexamethasone.

Note: Progression-free survival is calculated from date of first SVdX treatment after crossover from the Vd group to the date of first IRC-confirmed progressive disease, according to International Myeloma Working Group response criteria, or death due to any cause, whichever occurred first.

Source: BOSTON Clinical Study Report.⁴

PFS After Treatment with SVd, Vd, or SVdX

A summary of PFS results for patients who received SVd, Vd, or SVdX after receiving subsequent therapy post-progression on BOSTON study therapy is provided in Table 22. PFS events were similar between the SVd (35.4%) and Vd (32.8%) groups, but higher in the SVdX group (49.2%); this is expected, as patients in the SVdX group would have progressed on an additional line of therapy compared to patients in the SVd and Vd groups (Figure 8).

Table 22: Progression-Free Survival After Treatment With SVd, Vd, or SVdX

MM = multiple myeloma; SVd = selinexor plus bortezomib plus dexamethasone; SVdX = selinexor plus bortezomib plus dexamethasone after crossover; Vd = bortezomib plus dexamethasone.

Note: Progression-free survival is calculated from the date of first dose of post-SVd, Vd, or SVdX treatment to the date of first PD on post-SVd, Vd, or SVdX treatment, or death due to any cause. Patients who do not have an event are censored at the date of last disease assessment, or the database cut-off date, whichever occurs first.

Source: BOSTON Clinical Study Report.⁴

Figure 8: Progression-Free Survival After Treatment With SVd, Vd, or SVdX

SVd = selinexor plus bortezomib plus dexamethasone; SVdX = selinexor plus bortezomib plus dexamethasone after crossover; Vd = bortezomib plus dexamethasone.

Note: Only patients who do not cross over are included in the Vd arm. Progression-free survival is calculated from the date of first dose after treatment with SVd, Vd, or SVdX to the date of first progressive disease after treatment with SVd, Vd, or SVdX, or death due to any cause. Patients who do not have an event is censored at the date of last disease assessment, or database cut-off date, whichever occurred first.

Source: BOSTON Clinical Study Report.⁴

Subgroup Analysis

Forest plots depicting all results for subgroup analyses are presented in Figure 35 and Figure 36. Results for subgroup analysis of subgroups highlighted in the CADTH protocol are reported in Table 23. In general, point estimates at the primary of subgroup analyses at the primary analysis appeared to favour treatment with SVd compared to Vd, although results were not statistically significant for most subgroups. Subgroup analysis demonstrated improved PFS with SVd compared with Vd among patients without previous PI therapy (HR = 0.23; 95% CI, 0.11 to 0.60), with only 1 prior line of anti-MM therapy (HR = 0.63; 95% CI, 0.41 to 0.96), with stem cell transplant (HR = 0.55; 95% CI, 0.34 to 0.90), with del[17p] (HR = 0.38; 95% CI, 0.16 to 0.86), with 1q21 (HR = 0.62; 95% CI, 0.40 to 0.96), with del[17p] or t[4;14] or t[14;16] or 1q21 (HR = 0.67; 95% CI, 0.45, 0.98), and with previous exposure to lenalidomide (HR = 0.63; 95% CI, 0.41 to 0.97). Results of subgroup analyses at the updated analysis were consistent with the primary analysis, except that differences between SVd and Vd groups in the subgroup of patients with previous exposure to lenalidomide was no longer statistically significant. Subgroup analyses were not powered to detect differences between treatment groups in the BOSTON trial at either the primary or updated analyses.

Duration of Response

Results for DOR are presented in Table 24. At the primary analysis, more patients in the SVd group had achieved a PR or better (76.4%) compared with the Vd group (62.3%). The median DOR was 20.27 months (95% CI, 12.55 to NE) in the SVd group compared to 12.88 months (95% CI, 9.26 to 15.77) in the Vd group (Figure 9). Results at the updated analysis were consistent with results of the primary analysis. A greater proportion of patients in the SVd group achieved a PR or better compared to the Vd group (76.9% versus 63.3%, respectively). The median DOR was 17.28 months (95% CI, 12.55 to 26.25) in the SVd group and 12.88 months (95% CI, 9.26 to 15.77; Figure 10) in the Vd group.

Table 24: Duration of Response Based on IRC Assessment

CI = confidence interval; DOR = duration of response; IRC = independent review committee; MM = multiple myeloma; NE = not evaluable; PR = partial response; SVd = selinexor plus bortezomib plus dexamethasone; Vd = bortezomib plus dexamethasone.

Note: DOR is defined for patients with a confirmed PR or better as the duration from the date of first IRC-confirmed PR or better to the date of first IRC-confirmed progressive disease, or death due to any cause, whichever occurs first. All percentages are calculated based on patients with a PR or better.

^aStratified for prior proteosome inhibitor therapies, number of prior of anti-MM regimens and Revised International Staging System stage at screening.

^bBased on a stratified Cox proportional hazards model with the Efron method of handling ties.

Source: BOSTON Clinical Study Report.⁴

Figure 9: Duration of Response Based on IRC Assessment (ITT Population; Data Cut-Off: February 18, 2020)

IRC = independent review committee; ITT = intention-to-treat; SVd = selinexor plus bortezomib plus dexamethasone; Vd = bortezomib plus dexamethasone.

Source: BOSTON Clinical Study Report.⁴

Figure 10: Duration of Response Based on IRC Assessment (ITT Population; Data Cut-Off: February 15, 2021)

IRC = independent review committee; ITT = intention-to-treat; SVd = selinexor plus bortezomib plus dexamethasone; Vd = bortezomib plus dexamethasone.

Source: BOSTON Clinical Study Report.⁴

Time to Next Treatment

Results for TTNT are reported in Table 25. There were fewer patients in the SVd groups with TTNT events (45.1%) versus the Vd group (65.2%). The median TTNT was longer in the SVd group at 16.13 months (95% CI, 13.92 to NE) than in the Vd group at 10.84 months (95% CI, 9.82 to 13.40; Figure 11). There was a longer median treatment-free interval for patients with new MM treatment in the SVd group at 28.0 days (range, 1 to 447) than the Vd group at 14.0 days (range, 1 to 419).

Table 25: Time to Next Treatment (Intention-to-Treat Population)

CI = confidence interval; MM = multiple myeloma; SVd = selinexor plus bortezomib plus dexamethasone; Vd = bortezomib plus dexamethasone.

^aBased on a stratified Cox proportional hazard models with the Efron method of handling ties.

^bStratified for prior proteosome inhibitor therapies, number of prior of anti-MM regimens and Revised International Staging System stage at screening.

Source: BOSTON Clinical Study Report.⁴

Figure 11: Time to Next Treatment (Primary Analysis; Intention-to-Treat Population)

SVd = selinexor plus bortezomib plus dexamethasone; Vd = bortezomib plus dexamethasone.

Source: BOSTON Clinical Study Report.⁴

Time to Treatment Discontinuation

A summary of TTD is reported in Table 26. There were no differences between the SVd and Vd treatment groups in patients discontinuing treatment (81.0% versus 82.6%, respectively). The median TTDs were 7.10 months (95% CI, 6.44 to 8.54) in the SVd group and 7.95 months (95% CI, 6.80 to 9.23; Figure 12) in the Vd group.

Table 26: Time to Treatment Discontinuation (Intention-to-Treat Population)

CI = confidence interval; MM = multiple myeloma; SVd = selinexor plus bortezomib plus dexamethasone; Vd = bortezomib plus dexamethasone.

^aStratified for prior proteosome inhibitor therapies, number of prior anti-MM regimens and Revised International Staging System stage at screening.

^bBased on a stratified Cox proportional hazards model with the Efron method of handling ties.

Source: BOSTON Clinical Study Report.⁴

Figure 12: Time to Treatment Discontinuation

SVd = selinexor plus bortezomib plus dexamethasone; Vd = bortezomib plus dexamethasone.

Note: Time to treatment discontinuation is calculated from date of randomization to date of treatment discontinuation. Patients without events were censored at the last non-zero dose date or database cut-off, whichever occurred first. For crossover patients, only the end of treatment of the randomization phase was considered.

Source: BOSTON Clinical Study Report.⁴

Time to Response

Results for TTR are reported in Table 27. A greater proportion of patients in the SVd group had an IRC-confirmed response of PR or greater (76.4%) compared with the Vd group (62.3%). The median TTR was numerically shorter in the SVd group at 1.41 months (95% CI, 1.35 to 1.51) than in the Vd group at 1.61 months (95% CI, 1.51 to 2.14) (Figure 13).

Table 27: Time to Response (Intention-to-Treat Population)

CI = confidence interval; IRC = independent review committee; MM = multiple myeloma; PR = partial response; SVd = selinexor plus bortezomib and dexamethasone; Vd = bortezomib plus dexamethasone.

^aBased on a stratified Cox proportional hazards model with the Efron method of handling ties.

^bStratified for prior proteosome inhibitor therapies, number of prior of anti-MM regimens and Revised International Staging System stage at screening.

Source: BOSTON Clinical Study Report.⁴

Figure 13: Time to Response

SVd = selinexor plus bortezomib and dexamethasone; Vd = bortezomib plus dexamethasone.

Source: BOSTON Clinical Study Report.⁴

Overall Response Rate

At the primary analysis, a total of 149 patients (76.4%; 95% CI, 69.8 to 82.2) in the SVd group had an ORR compared to 129 patients (62.3%; 95% CI, 55.3 to 68.9) in the Vd group. There were no differences in the best overall response of patients between the 2 treatment groups. Most patients achieved a PR (31.8% in the SVd group versus 30.0% in the Vd group), VGPR (27.7% versus 21.7%, respectively), or stable disease (12.8% versus 19.3%, respectively).

At the updated analysis, the ORR of the SVd group increased to 150 patients (76.9%; 95% CI, 70.4 to 92.6) versus 131 patients (63.3%; 95% CI, 56.3 to 69.9) in the Vd group. Little difference remained between treatment groups in best overall response. The results were consistent with those of the primary analysis.

A summary of results for ORR in the crossover (SVdX) group is provided in Table 28. The ORR was reported in 12 patients (19.0%; 95% CI, 10.2 to 20.9) and results for best overall response were similar to results of the total population.

Table 28: Overall Response Rate Based on IRC Assessment (Intention-to-Treat Population)

CI = confidence interval; IRC = independent review committee; MM = multiple myeloma; SVd = selinexor plus bortezomib plus dexamethasone; Vd = bortezomib plus dexamethasone.

^aOverall response rate is the proportion of patients who achieve a partial response or better, before IRC-confirmed progressive disease or initiating a new MM treatment or crossover.

Source: BOSTON Clinical Study Report.⁴

Table 29: Overall Response Rate Based on IRC Assessment (SVdX Population)

CI = confidence interval; IRC = independent review committee; MM = multiple myeloma; SVdX = selinexor plus bortezomib plus dexamethasone after crossover.

^aOverall response rate is the proportion of patients who achieve a partial response or better, before IRC-confirmed progressive disease or initiating a new MM treatment or crossover.

Source: BOSTON Clinical Study Report.⁴

Subgroup Analysis

A depiction of all subgroup analyses is provided in Figure 38. Results of subgroup analyses highlighted in the CADTH protocol are reported in Table 30. Subgroup analysis suggested an improvement with treatment with SVd over Vd in patients who had received prior PI therapies (OR = 2.26; 95% CI, 1.37 to 3.7), with only 1 prior line of anti-MM therapy (OR = 2.20; 95% CI, 1.15 to 4.21), with stem cell transplant (OR = 2.91; 95% CI, 1.35 to 6.29), with del(17p) (OR = 5.33; 95% CI, 1.28 to 22.2), with del(17p) or t(4;14) or t(14;16) (OR = 3.21; 95% CI, 1.23 to 8.37), with 1q21 (OR = 2.06; 95% CI, 1.0359 to 4.0798), with del(17p) or t(4;14) or t(14;16) or 1q21 2.7015 (OR = 2.70; 95% CI, 1.45 to 5.04), and previous bortezomib exposure (OR = 2.38; 95% CI, 1.41 to 4.02).

Table 30: Subgroup Analysis of Overall Response Rate Based on IRC Assessment

CI = confidence interval; IRC = independent review committee; MM = multiple myeloma; OR = odds ratio; PI = proteasome inhibitor; SVd = selinexor plus bortezomib plus dexamethasone; Vd = bortezomib plus dexamethasone.

Source: BOSTON Clinical Study Report.⁴

Rate of Very Good Partial Responses or Better Responses

At the primary analysis, a VGPR, CR, or sCR was observed in 87 (44.6%) of 195 patients from the SVd group and 67 (32.4%) of 207 patients from the Vd group (OR = 1.6594; 95% CI, 1.0993, 2.5049; P = 0.0082). At the updated analysis, a VGPR, CR, or sCR was observed in 87 (44.6%) of 195 patients from the SVd group and 67 (32.4%) of 207 patients from the Vd group (OR = 1.6594; 95% CI, 1.0993 to 2.5049; P = 0.0082).

Health-Related Quality of Life

Patient-Reported Peripheral Neuropathy Measured by EORTC QLQ-CIPN20

Baseline scores for the sensory, motor, and autonomic neuropathy symptoms subscales were similar between the 2 treatment groups. A summary of post-baseline score increases for the 3 subscales is presented in Table 31. Regarding the sensory and motor subscales, a greater proportion of patients in the Vd group had higher post-baseline scores that showed increases from baseline equal to or greater than 10, 20, 30, 40, and 50 compared with the SVd group, indicating worse symptoms for patients in the Vd group. Regarding the autonomic subscale, a greater proportion of patients in the SVd group had higher post-baseline scores that showed increases from baseline equal to or greater than 10, 20, 30, 40, and 50 compared with the Vd group, indicating worse symptoms for patients in the SVd group. Linear mixed-effect models were also conducted for the EORTC QLQ-CIPN20 scores; a lower mean change from baseline was observed in the SVd group compared to the Vd group for the sensory, motor, and autonomic subscale, indicating reduced symptom burden in the SVd treatment group.

Figure 14 presents the absolute values of sensory and motor subscale scores over time. Symptom scores were numerically higher in the Vd group than in the SVd group, suggesting greater symptom burden; however, there were no statistically significant differences between the 2 treatment groups. Absolute values of PN autonomic symptom scores are presented in Figure 15. Patients in the SVd group showed higher symptom scores compared to the Vd group, indicating greater symptom burden; however, the results were not statistically significantly different. The results of the autonomic symptom score were broken down to its 3 components of blurred vision, difficulty with erection, and dizzy when standing up (Figure 16). The SVd and Vd groups showed similar scores for the dizziness and erectile-function components. The SVd group showed higher scores for blurred vision than the Vd group, indicating greater symptom burden.

Absolute values of PN autonomic symptom scores are presented in Figure 15. There were no statistically significant results between the SVd and Vd groups for any of the subscales. The results of the autonomic symptom score were broken down to its 3 components of blurred vision, difficulty with erection, and dizzy when standing up (Figure 16). The SVd and Vd groups showed similar scores for the dizziness and erectile-function components. The SVd group showed higher scores for blurred vision than the Vd group, indicating greater symptom burden.

Table 31: Change From Baseline of EORTC QLQ-CIPN20 (Primary Analysis; ITT Population)