Input From Clinical Experts Consulted by CADTH

• According to the clinical experts consulted by CADTH, newer treatments are needed that exhibit better control of MM and less toxicity. In particular, needs are not being met for patients who are refractory to certain drug classes, such as immunomodulators (lenalidomide) or proteasome inhibitors (bortezomib), and outcomes tend to be poor in these patients.
• Isatuximab should be combined with other drugs that have unrelated mechanisms/toxicity profiles. Otherwise, it could be used in any line of therapy. For patients with 1 prior line of therapy, an isatuximab regimen could be particularly useful if they had not received a prior anti-CD38 drug. Whether there would be benefit for those previously treated with an CD38 antibody drug is unknown.
• There is no established method for determining which patients would most or least benefit from treatment. A clinically significant response would be improved PFS with acceptable toxicity and quality of life. Response should be assessed before each treatment cycle and disease progression; unacceptable toxicity would warrant discontinuation of treatment. A copy of the input is presented in Appendix 2.

Clinician Group Input

• Input was submitted by the Canadian Myeloma Research Group (CMRG) and the Ontario Health Cancer Care Ontario Hematology Cancer Drug Advisory Committee (OH-CCO DAC).
• There were no notable differences between the input provided by the clinical experts consulted by CADTH on this review and that provided by the clinician groups.
• The clinician groups did not specifically comment on their own experiences with IsaKd; however, they did note that they believed IsaKd would be useful in patients with relapse whose MM had progressed while they were receiving lenalidomide and/or bortezomib.

Description of Studies

The CADTH systematic review included 4 reports of 1 pivotal trial (IKEMA).^5-8 No additional studies were identified from the literature. IKEMA is an ongoing, sponsor-funded, multinational (with Canadian sites) open-label randomized controlled trial (RCT) that randomized 302 adults (> 18 years) patients with relapsed and/or refractory MM and 1 to 3 prior lines of therapy, in a 3:2 manner, to either IsaKd or carfilzomib plus dexamethasone (Kd). Patients in the IsaKd group received isatuximab 10 mg/kg by IV infusion in 28-day cycles (weekly for first cycle, then biweekly thereafter) with carfilzomib 20 mg/m², escalated to 56 mg/m² IV days 1 to 2, 8 to 9, and 15 to 16, and dexamethasone 20 mg twice weekly, while patients in the Kd group received carfilzomib and dexamethasone at those same dosage regimens. Patients were treated until they experienced disease progression or unacceptable toxicity, or they decided to discontinue study treatment. Randomization was stratified by the number of prior lines of therapy (1 versus > 1) and the Revised International Staging System (R-ISS) score (I or II versus III versus not classified).

The primary outcome of the IKEMA trial was PFS, and the key secondary outcomes included overall response rate (ORR), very good partial response (VGPR) or better rate, duration of response (DOR), time to first response (TTR), minimal residual disease (MRD) negativity in patients with VGPR or better, as well as complete response (CR) rate, and OS. PFS, ORR, VGPR or better, and MRD negativity in patients with VGPR or better were included in the statistical testing hierarchy. HRQoL was assessed as an exploratory outcome. The findings in this report are from an interim analysis, which was planned to take place once 103 progression events had occurred (information fraction of 65%). Results for the final analysis, including OS data, are not expected until 2023. Harms including adverse events (AEs), serious adverse events (SAEs), and AEs of special interest, were also measured and reported.

Patients were an average of 63.1 years of age (standard deviation [SD] 9.9); 56% were men and 70.9% were White. The majority of patients had MM of the immunoglobulin G (IgG) subtype (67.9%) at diagnosis, followed by immunoglobulin A (IgA) (22.8%), and these percentages were similar to those observed at study entry (69.9% and 22.5%, respectively). The most common International Staging System (ISS) stage at study entry was stage I (53.0%), followed by stage II (31.1%) and stage III (15.2%). The majority of patients had relapsed and refractory MM (71.5%), while the remainder had relapsed MM (28.5%). The average number of prior regimens was 3.2 (SD 1.7), and the number of prior lines was 1.8 (SD 0.8). Patients were most commonly refractory to an immunomodulatory imide drug (45.0% of patients), followed by a proteasome inhibitor (33.1%) or both (20.5%).

Efficacy Results

The efficacy results for IKEMA are summarized in Table 2. PFS was the primary outcome of IKEMA, and, at the interim analysis (median follow-up of 20.73 months), median PFS was not reached in the IsaKd group and was 19.15 months (95% confidence interval [CI], 15.77 to not calculable) in the Kd group, for a stratified hazard ratio (HR) of 0.531 (99 CI, 0.318 to 0.889), and a P value for a stratified log-rank test of 0.0007. In the IsaKd group, 26.8% of patients had a PFS event, while, in the Kd group, 44.7% of patients had a PFS event. The results for sensitivity analyses performed for the primary outcome were consistent with the primary analysis, and planned subgroup analyses revealed consistent results across various subgroups of patients of interest for this review.

OS will be assessed at the end of study; therefore, no median OS data were available at the time of the interim analysis.

HRQoL was assessed using the European Organisation for Research and Treatment of Cancer Quality of Life of Cancer Patients questionnaire with 30 items (EORTC QLQ-C30), the EORTC MM module Quality of Life questionnaire with 20 items (QLQ-MY20), and the EuroQol 5-Dimensions 5-Levels (EQ-5D-5L) questionnaire. Interpretation of HRQoL data was limited by the large number of withdrawal from the study over time; however, generally, for the EORTC questionnaires, there was little change from baseline in HRQoL scores in the IsaKd group, and numerical increases from baseline over time were observed in the Kd group. An increase in score on these instruments indicates an improvement in HRQoL.

The ORR was assessed in all responders (patients achieving either a stringent complete response [sCR], CR, VGPR or partial response [PR]) and in patients achieving a VGPR or better. An sCR is defined as patients having a CR who also have normalized free light chain (FLC) ratio in the absence of bone marrow plasma cells when assessed by immunohistochemistry or immunofluorescence. The percentage of patients responding was 86.6% in the IsaKd group and 82.9% in the Kd group, and the between-group difference was not statistically significant (P = 0.1930). As this was the second outcome in the statistical hierarchy, testing was to have halted for subsequent outcomes, although the sponsor continued to conduct testing and report P values for descriptive purposes. The percentage of patients achieving a VGPR or better was 72.6% in the IsaKd group and 56.1% in the Kd group. No patients achieved an sCR, while 39.7% of patients in the IsaKd group and 27.6% of patients in the Kd group achieved a CR, and 33.0% and 28.5% of patients, respectively, achieved a VGPR. MRD negativity was achieved by 29.6% of patients in the IsaKd group and 13.0% of patients in the Kd group.

The median DOR was calculated based on 155 patients in the IsaKd group and 102 patients in the Kd group. The median DOR was not yet reached in either treatment group, and the HR was 0.425 (95% CI, 0.269 to 0.672). The median TTR was 1.08 months (95% CI, 1.05 to 1.12) in the IsaKd group and 1.12 months (95% CI, 1.05 to 1.18) in the Kd group, for a stratified HR of 1.143 (95% CI, 0.888 to 1.471).

Harms Results

A summary of the harms reported in the IKEMA trial is available in Table 2. There were 97.2% of patients in the IsaKd group and 95.9% of patients in the Kd group who had at least 1 AE, and 76.8% versus 67.2%, respectively, who had a grade 3 or greater AE, and 3.4% versus 3.3% who had a grade 5 AE. The most common AE in the IsaKd group was an infusion-related reaction, which occurred in 44.6% of patients in the IsaKd group and 3.3% of patients in the Kd group. Other common AEs (IsaKd versus Kd) included hypertension (36.7% versus 31.1%), diarrhea (36.2% versus 28.7%), upper respiratory tract infection (36.2% versus 23.8%), fatigue (28.2% versus 18.9%), and dyspnea (27.7% versus 21.3%). The most common grade 3 or greater AEs (IsaKd versus Kd) were hypertension (20.3% versus 19.7%) and pneumonia (16.4% versus 12.3%).

SAEs occurred in 59.3% of patients in the IsaKd group and 57.4% of patients in the Kd group. The most common SAE was pneumonia (IsaKd versus Kd, 18.1% versus 11.5%).

There were 8.5% of patients in the IsaKd group and 13.9% of patients in the Kd group who had an AE leading to definitive treatment discontinuation. One patient discontinued treatment of isatuximab due to an AE.

Among notable harms, respiratory tract infections occurred in 83.1% of patients in the IsaKd group and 73.8% of patients in the Kd group, and these were grade 3 or greater events in 32.2% versus 23.8% of patients, respectively. Cardiac disorders occurred in 7.3% of patients treated with IsaKd versus 5.7% of patients treated with Kd. Second primary malignancies (solid, non-skin) occurred in 2.8% versus 3.3% of patients in the IsaKd and Kd groups, respectively, and second primary malignancies (solid, skin) in 5.1% versus 2.5% of patients, respectively. There were no hematologic malignancies reported. Events of decreased neutrophil counts occurred in 54.8% of patients in the IsaKd group versus 43.4% of patients in the Kd group, and grade 3 or greater events occurred in 19.2% versus 7.4% of patients, respectively. Events of decreased platelet counts occurred in 94.4% of patients treated with IsaKd and 87.7% of patients treated with Kd, and these were grade 3 or greater events in 29.9% versus 23.8% of patients, respectively.

Table 2

Summary of Key Results From the IKEMA Trial.

Critical Appraisal

• IKEMA was an open-label trial, and lack of blinding may have biased results, particularly for patient-reported outcomes such as HRQoL and for reporting of harms. Assessment of pathology was conducted by a blinded independent review committee (IRC) and therefore is unlikely to have been influenced by lack of blinding.
• The results of the IKEMA trial were based on a planned interim analysis, with an information fraction of 65%; therefore, there is a risk of overestimation of the primary effect for PFS. However, given the statistically and clinically significant difference observed between the groups for PFS, the potential for overestimation is unlikely to have altered the conclusions.
• Multiplicity was controlled for with the use of a hierarchical testing procedure; however, early failure of the hierarchy meant that statistical testing was conducted only on the primary and first secondary outcomes. This meant that there were several outcomes for which no inferences could be drawn about differences between groups. HRQoL was not included in the hierarchy, and differences between groups were not tested statistically; therefore, no conclusions could be drawn about this outcome.
• The clinical experts consulted by CADTH noted that the patients included in the IKEMA trial were approximately 10 years younger and had a better Eastern Cooperative Oncology Group (ECOG) performance status than patients they would see diagnosed with MM, although this is a common occurrence in clinical trials, which tend to recruit younger, healthier patients. Otherwise, the baseline characteristics and the treatment regimens used in the trial were consistent with what 1 would expect to see in Canadian clinical practice.

Description of Studies

The sponsor conducted several indirect treatment comparisons (ITCs) that included fixed-effects network meta-analyses (NMAs) and matching-adjusted indirect comparisons (MAICs). A systematic review and feasibility assessment were done to identify studies to include in the ITCs. On that basis, it was determined that it was feasible to conduct an NMA including 8 studies (||||||||||) in a connected network that included IsaKd, and 4 separate MAICs based on individual-level data from the IKEMA trial and summary data from 2 studies. ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||

Efficacy Results

|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||

Harms Results

||||||||||||||||||||||||||||||||||||||||||||||||||||

Critical Appraisal

The trial populations included in the NMAs were relatively homogenous in age, ECOG performance status, race, and gender; however, there were some concerns from clinical experts regarding heterogeneity in the prior treatments received. Specifically, prior lenalidomide use is likely a large effect modifier that differs between trials and greatly increases the uncertainty in these findings. In addition, studies included in the network were conducted over a wide span of time, during which the treatment approach for MM has rapidly evolved. Thus, the time span of these trials may further introduce bias to the comparisons in the NMA. Sparsity of the network meant that only a fixed-effects model could be estimated, which limits the ability to detect and/or account for heterogeneity. ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||

In the MAICs, the assumption that all prognostic factors and effect modifiers were adequately adjusted for is unlikely to be the case. In general, the baseline characteristics differed across studies. Specifically, the variation in the prior treatments received may be a serious effect modifier, reflecting differences in care over the wide span of time during which the trials were conducted. Previous lenalidomide use was specifically noted as a likely effect modifier by 1 of the clinical experts, and prior treatment in general is an effect modifier. The choice of the matching factors was based on internal expert opinion (rather than a survey of clinical experts) and availability/completeness of data in the trials (which is inconsistent with the National Institute for Health and Care Excellence [NICE] Decision Support Unit [DSU] guidelines, which recommend the identification of key factors in the data). |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||||| The reported effective sample sizes, and the skewness or outliers apparent in the visualizations of the weight distributions, suggest that the results may be heavily influenced by a small subset of patients from the IKEMA trial. Generalizability may be an issue due to the small sample size remaining after the exclusions and matching — the remaining patients and weighted sample are unlikely to be representative of the entire patient population.

Unmet Needs

The clinical experts consulted by CADTH on this review noted that MM remains an incurable disease, with significant impairment in HRQoL due both to the disease and treatments used to manage it. Newer treatments are needed that exhibit better control over the disease and less toxicity. Additionally, goals are not being met for patients who are refractory to certain classes of drugs (immunomodulators [specifically lenalidomide] or proteasome inhibitors [bortezomib]), and the clinical experts noted that outcomes tend to be particularly poor in these patients.

Place in Therapy

Isatuximab is an anti-CD38 monoclonal antibody, like daratumumab, and is likely best used in combination with other therapies with different mechanisms of action and toxicity profiles. Otherwise, it could be used in any line of therapy. For patients who have received 1 prior line of therapy, an isatuximab-containing regimen could be beneficial, particularly if they had not previously been treated with an anti-CD38 drug like daratumumab. Whether isatuximab is beneficial in patients previously treated with another anti-CD38 is unknown, as trials have not included patients previously treated with a monoclonal antibody.

Patient Population

Any myeloma patient who is suitable for treatment could benefit from isatuximab, and there are no established methods for determining patients who would not benefit. Patients whose disease is actively progressing or not responding to current treatment, or who are intolerant of their current therapy and are not yet adequately treated, could benefit from isatuximab.

The most likely approach to identifying patients who might best respond is clinical judgment. Pre-symptomatic patients should be treated, particularly if there is evidence of disease progression or lack of response to current or prior therapy. Patients who are dying would not be suitable for treatment with isatuximab. It is not possible to identify patients who are more likely to respond to isatuximab.

Assessing Response to Treatment

IMWG response criteria are used to determine whether a patient is responding to treatment. A clinically significant response to treatment is indicated by improved PFS, with acceptable drug toxicity and quality of life. Treatment response should be assessed before each treatment cycle.

Discontinuing Treatment

Disease progression or intolerable toxicity would warrant discontinuation of treatment.

Prescribing Conditions

The most appropriate setting for treatment with isatuximab, according to the clinical experts, is an outpatient systemic cancer therapy unit.

Additional Considerations

According to the clinical experts consulted by CADTH on this review, in Canada, first-line treatment of MM for transplant-eligible patients does not currently incorporate a CD38 antibody drug or carfilzomib except in clinical trials. In transplant-ineligible patients, first-line daratumumab plus lenalidomide plus dexamethasone (DRd) is not currently reimbursed, and, therefore, most patients in Canada have not received a monoclonal antibody in first-line treatment. In the second-line treatment setting, most patients are refractory to lenalidomide that has been given until disease progression, and many of the remaining patients are intolerant of lenalidomide. Therefore, a lenalidomide-free second-line regimen is needed. The most commonly used second-line options are daratumumab plus bortezomib plus dexamethasone (DVd) and Kd regimens, and DVd is often chosen as a means of providing a CD38 antibody drug to the patient, in which case Kd is deferred to third- or fourth-line treatment. DVd, although commonly used, is known to have poor outcomes for patients, as seen in the CASTOR trial (median PFS for DVd of 7.8 months to first relapse in patients refractory to lenalidomide).¹⁵ Real-world data from Canada also show poor outcomes for patients whose disease has progressed while they were receiving lenalidomide and who are then treated with DVd (median PFS of 11.47 months among patients progressing on lenalidomide maintenance following 1 line of treatment).¹⁶ Both DVd and Kd appear more effective than bortezomib plus dexamethasone in randomized trials, likely because of the addition of the CD38 antibody drug in DVd and because of a better-performing proteasome inhibitor in Kd. However, neither of these regimens typically leads to durable remission, and more effective treatments are needed. If the IsaKd regimen were available in Canada, it would be preferred over DVd or Kd due to superior efficacy and good tolerability. Regimens containing pomalidomide are often deferred for use following regimens based on carfilzomib. There are 2 reasons for this: patients refractory to lenalidomide who are not yet refractory to a proteasome inhibitor are thought to be more likely to benefit from a regimen based on a proteasome inhibitor regimen than 1 based on pomalidomide, and access to pomalidomide generally requires prior exposure to lenalidomide and a proteasome inhibitor exposure, although there is variation in this across jurisdictions arising from differences in funding.

Unmet Needs

According to the CMRG, MM can cause significant mortality and morbidity, including fractures, decreased blood counts, renal failure, and infections, and these significantly affect patient quality of life. The most important goals of therapy are to ensure a high-quality and durable response to treatment (measured by response rates and PFS), which will delay disease-related complications and ultimately improve quality of life. The unmet needs include treatment for patients who become refractory to major classes of drug, specifically the immunomodulatory drugs (lenalidomide) and/or proteasome inhibitors (bortezomib). Patients refractory to these therapies tend to have poor outcomes overall. Many patients die before receiving subsequent lines of therapy. Therefore, there is a need to provide patients with optimal treatment to control the disease as early as possible in the disease course. According to OH-CCO DAC, patients who are not eligible for DRd have the greatest unmet need for an intervention such as IsaKd.

Place in Therapy

The CMRG believed that IsaKd is ideal for patients who have progressed after either a bortezomib- and/or lenalidomide-containing regimen. Among these patients, regimens containing CD38 antibody drugs are most commonly used at relapse as (1) they have a different mechanism of action from the drug previously received and (2) exposure to these drugs is required for future enrolment in clinical trials. Currently funded CD38 antibody regimens include DVd and DRd, and DVd is most commonly used second-line, as most patients progress on lenalidomide maintenance (ASCT eligible) or Rd (ASCT ineligible) in the first-line treatment setting. DVd is not optimal, however, as PFS is no more than 8 months in patients refractory to lenalidomide. Therefore, there remains an unmet need in optimizing regimens containing daratumumab in the post-lenalidomide setting, and further options are required. IsaKd would address the need of patients who have relapsed on lenalidomide and/or bortezomib. IsaKd would provide access to an CD38 antibody drug in this setting, along with a different proteasome inhibitor, which would be expected to replace and greatly improve upon current standard of care. Based on the ENDEAVOUR trial, carfilzomib is clearly superior to bortezomib in patients who have relapsed.¹⁷ OH-CCO DAC agreed that IsaKd would be an alternative treatment option for second-line treatment and beyond. This provides a monoclonal antibody treatment option for patients refractory to lenalidomide.

With respect to whether other therapies should be attempted before initiating IsaKd, CMRG noted that, in MM, it is important to use the best therapies up front, given the attrition seen with the disease. Additionally, the current standard of care, DVd, is suboptimal, with data showing poor efficacy in both clinical trials and real-world Canadian settings.^16,18 OH-CCO DAC agreed it was not appropriate to recommend patients try other treatments before initiating treatment with IsaKd.

CMRG noted that sequencing would likely be affected by the approval of IsaKd, as it would most likely be used second-line in patients whose disease had progressed while they were receiving lenalidomide and/or proteasome inhibitors. Regimens containing daratumumab (DRd or DVd) or carfilzomib (isatuximab plus lenalidomide plus dexamethasone [KRd] or Kd) would likely not be used in further lines of treatment. The CMRG also noted that, if DRd is funded for first-line treatment of transplant-ineligible MM patients as per data from the ongoing phase III MAIA trial, IsaKd would likely no longer be used second-line.¹⁹ It would likely continue to have a role among transplant patients whose disease has progressed while receiving lenalidomide maintenance therapy, as they would not have been previously treated with a regimen containing an CD38 antibody drug. OH-CCO DAC agreed that, with respect to sequencing, the only option after IsaKd would be pomalidomide-based therapy.

Patient Population

In terms of which patients are best suited to IsaKd, CMRG noted that the PFS benefit was seen across most subgroups in the IKEMA trial. Patients least suited for IsaKd would be those with prior disease refractory to carfilzomib or previous CD38 antibody drugs, as they were excluded from IKEMA. They also noted emerging evidence suggesting that isatuximab given following therapy with daratumumab likely produces a poor response. The CMRG noted cardiovascular toxicity with carfilzomib is a concern; however, they also noted that IsaKd appeared to have a relatively favourable toxicity profile in IKEMA, although the reasons for this are unknown. As far as those most likely to benefit from treatment, the CMRG suggested patients with less advanced and less biologically aggressive disease would benefit the most, as is the case with all MM treatments. OH-CCO DAC believed that patients best suited to be treated with IsaKd are those who meet the eligibility criteria for the drug. They did not believe it is possible to identify those patients who are most likely to respond to the drug.

Assessing Response to Treatment

CMRG noted that the best ORR, quality of response (VGPR or complete remission), as well as durability of response (PFS) are outcomes used in clinical practice and clinical trials to assess response to treatment. MRD is becoming increasingly important in MM, given its concordance with OS. However, this end point is not routinely used as an indicator of response in Canadian clinical practice. A clinical meaningful response would be absence of progressive disease (PD) with minimal/tolerable side effects. Although deep (VGPR or better) and more sustained MRD-negative responses are ideal, any absence of PD provides a patient benefit, in that it minimizes myeloma-related end organ damage. This is particularly the case if relapse therapy is started when progression is documented by the usual parameters, but before severe myeloma-related organ damage has developed. In this setting, even stable disease may afford benefit. Almost all MM patients are followed continuously, with response assessed every 1 to 3 months with laboratory parameters (most commonly) and occasionally with advanced imaging and/or bone marrow biopsy (depending on the specific scenario). OH-CCO DAC stated that the usual MM response criteria would be sufficient to determine whether a patient is responding to therapy, and a clinically meaningful response would be indicated by improvement in myeloma parameters and in symptoms. They also noted that treatment response should be assessed every treatment cycle.

Discontinuing Treatment

Both clinician groups identified disease progression or intolerance to therapy despite dose reduction as the factors that should be considered when deciding to discontinue treatment with IsaKd.

Prescribing Conditions

The CMRG noted that MM is treated at many community and academic sites across Canada and that the drug is administered in specialized clinical spaces. Therefore, community-setting clinics, hospital outpatient clinics, and IV oncology drug administration facilities would all be appropriate treatment settings for IsaKd. According to OH-CCO DAC, IsaKd should be administered in outpatient chemotherapy suites.

Additional Considerations

OH-CCO DAC noted that IsaKd administration is associated with multiple visits to the chemotherapy suites, which may not be feasible for some patients.

Inclusion and Exclusion Criteria

To be included, patients had to have MM with measurable disease (by serum or urine M protein) and relapsed and/or refractory MM with at least 1 prior line of therapy and no more than 3 prior lines. Primary refractory patients (those who never achieved at least a minimal response with any treatment during their disease course) were excluded, as were those with serum FLC-measurable disease only and those with disease refractory to a prior CD38 monoclonal antibody drug (progression on or within 60 days after end of CD38 antibody treatment or failure to achieve at least a minimal response to treatment). Patients with an ECOG performance status more than 2 were also excluded.

Baseline Characteristics

Patients were an average of 63.1 years of age (SD 9.9); 56.0% were men and 70.9% were White (Table 7). The majority of patients had MM of the IgG subtype (67.9%) at diagnosis, followed by IgA (22.8%), and these percentages were similar to those observed at study entry (69.9% and 22.5%, respectively). The most common ISS stage at diagnosis was stage II (33.4%), followed by stage III (25.5%) and stage I (24.2%); stage was unknown in the remainder. At study entry, the most common stage was stage I (53.0%), followed by stage II (31.1%) and stage III (15.2%). Most patients had relapsed and refractory disease (71.5%), while the remainder had relapsed disease (28.5%). The average number of prior regimens was 3.2 (SD 1.7), and the number of prior treatment lines was 1.8 (SD 0.8). Patients were most commonly refractory to an immunomodulatory imide drug (45.0% of patients), followed by a proteasome inhibitor (33.1%) and both (20.5%).

There were some numerical differences between groups (IsaKd versus Kd) with respect to race (White 73.2% versus 67.5%), ISS stage at study entry (stage I: 49.7% versus 57.7%, stage II: 35.2% versus 25.2%), refractory status (relapsed and refractory: 68.2% versus 76.4%, relapsed: 31.8% versus 23.6%) and in the types of prior regimens used (alkylating antineoplastic drugs: 94.4% versus 82.1%; proteasome inhibitors: 92.7% versus 85.4%; immunomodulatory drugs: 76.0% versus 81.3%).

Table 7

Summary of Baseline Characteristics (Randomized Population).

Progression-Free Survival

The primary outcome was PFS, defined as the time from date of randomization to date of first documented PD (determined by the IRC) or death from any cause, whichever came first. If PD and death were not observed before the data cut-off date or date of initiation of further antimyeloma treatment, PFS was censored at the date of the last valid disease assessment not showing PD or the analysis cut-off date, whichever came first. Patients without an event and without any valid post-baseline assessment were censored at the date of randomization. Responses (including progression) were evaluated by a blinded IRC according to IMWG criteria using central laboratory results and central radiological review. Response and progression based on serum and/or urine M protein were confirmed by 2 consecutive assessments. Progression based on plasmacytomas or bone lesions did not require confirmation. For patients with measurable serum and/or urine M protein, PD was defined as at least 1 of the following (biologic criteria in 2 consecutive assessments):

• 25% or greater increase in serum M component from nadir (the absolute increase had to be 0.5 g/dL or more); serum M component increases of 1 g/dL or more in 2 consecutive assessments were sufficient to define relapse if starting M component was 5 g/dL or more; and/or
• 25% or greater increase in urine M component from nadir (the absolute increase had to be 200 mg/24 hours or more); and/or
• definite development of new bone lesions or soft tissue extramedullary disease or an increase or 50% more from nadir in the sum of perpendicular diameters of existing soft tissue extramedullary disease lesions if there was more than 1 lesion or 50% or greater increase in the longest diameter of a previous soft tissue extramedullary disease lesion greater than 1 cm in short axis.

The date of the PD was defined as the earliest date that indicated PD (provided that PD was confirmed when required). An M protein assessment performed after the initiation of a further antimyeloma treatment was used to confirm PD. In addition, deaths due to PD within 45 days of the first documentation of PD (regardless of initiation of further therapies) were used to confirm PD. If a post-baseline disease assessment for response was determined to be “not evaluable” based on IMWG criteria, or the date of assessment was missing, the disease assessment was considered invalid. For a given time point not showing disease progression, if several examinations were performed on different dates, the date of the last valid assessment was the date of the latest examination. Clinical deterioration could also be considered progression in the primary analysis of PFS if the IRC considered that clinical data reported in the case report form supported clinical progression. In the case of hypercalcemia, a full disease assessment was performed to identify any measurable parameter of myeloma progression (e.g., serum and urine M protein, lytic lesions assessment, and plasmacytoma assessment), and potential alternative causes of hypercalcemia were ruled out. Progression was not diagnosed based on FLC progression only. If both serum and urine M protein were below the level of eligibility in the efficacy analyses performed on cycle 1, day 1, progression was assessed per IMWG response criteria.

Overall Survival

OS was a key secondary outcome and was defined as the time from date of randomization to death from any cause. OS was not assessed in the interim analysis but will be assessed in the final analysis, according to the statistical analysis plan.

Health-Related Quality of Life

HRQoL was assessed, and results were presented descriptively using 3 different instruments: the EORTC QLQ-C30, the EORTC QLQ-MY20, and the EQ-5D-5L. A more complete description is provided and the validity of these patient-reported outcomes is summarized in Appendix 4. Each of these instruments were completed by the patient on days 1 and 2 of each cycle, as well as at the end of treatment visit (30 days after the last treatment was administered) and 90 days (within 5 days) after administration of the last study treatment. The EORTC QLQ-C30 is a 30-item instrument, with 15 items related to function (physical, role, cognitive, emotional, social), 7 questions related to symptoms (7 questions), single-item symptom scales (dyspnea, insomnia, appetite loss, constipation, diarrhea, financial impact), and 2 items related to global quality of life (QoL). All except global QoL are scored from 1 (“not at all”) to 4 (“very much”). Global QoL, also referred to as global health status, is scored on a scale from 1 (very poor) to 7 (excellent).²⁵ The EORTC QLQ-MY20 is a module that was developed specifically for MM and consists of 20 items, 2 symptom scales (disease symptoms and side effects), and 2 functional scales (future perspective and body image).²⁶ The questions are scored on the same 1 to 4 scale and responses are then converted to a standardized score that ranges from 0 to 100. The EQ-5D-5L consists of 5 dimensions (mobility, self-care, usual activities, pain/discomfort, and anxiety/depression), and each dimension is rated on 5 levels, from level 1 (“no problems”) to level 5 (“extreme problems” or “unable to perform”).²⁷ This rating is then converted, using an algorithm, to a scale from 0 (dead) to 1 (full health).²⁸

Overall Response Rate

ORR, a key secondary outcome, was defined as the proportion of patients with sCR, CR, VGPR, and PR as best overall response, assessed by IRC using IMWG criteria, and was assessed at the time of the primary PFS analysis and/or the final PFS analysis. Bone marrow biopsy could have been performed for sCR assessment, per investigator decision. sCR was not defined in the study protocol; however, it is defined by the IMWG on their website as patients with a CR who also have a normalized FLC ratio in the absence of monoclonal bone marrow plasma cells when assessed by immunohistochemistry or immunofluorescence.²⁹

Rate of VGPR or better was a key secondary outcome and was defined as the proportion of patients with an sCR, CR, or VGPR as their best overall response.

The CR rate was defined as the proportion of patients with sCR or CR as best overall response. Patients with demonstrated isatuximab interference were considered in the best overall response category, corresponding to the M protein assessment obtained without interference, when the antibody-capture interference assay was available. Antibodies that are used therapeutically can interfere with M protein assessment, because they can be detected by immunoelectrophoresis or immunofixation, making it difficult to determine whether the therapeutic antibody or the disease marker is being detected. Stable disease, non-PD, and PD were also reported. A patient was considered to have a non-PD response when they had no more measurable serum or urine M protein on cycle 1, day 1.

Depth of Response

The MRD negativity rate in patients with VGPR or better was a key secondary outcome. This was defined as the proportion of patients for whom MRD was negative at any time point after the first dose of study treatment. MRD status was assessed centrally by ClonoSEQ assay in bone marrow samples from patients who achieved VGPR or better, to determine the depth of response at the molecular level. The threshold for negativity was 10⁻⁵. MRD status was considered negative if at least 1 result of the assessment was negative in the patient; otherwise, the MRD was considered positive. Those in the intention-to-treat (ITT) population without MRD assessment were considered positive.

Bone marrow aspirates were collected at screening and at the time of VGPR or better for confirmation. Because bone marrow aspiration is an invasive procedure, guidance was provided to limit the number of bone marrow assessments. A maximum of 6 bone marrow assessments could be performed.

For patients with CR without previous documentation of VGPR:

• The first bone marrow for MRD assessment was collected at time of confirmation of CR (i.e., at the second time point showing CR). If the patient was determined to be MRD-positive, another bone marrow sample was collected 3 months (3 cycles) later, to identify late negativity. A third sample may have been collected after another 3 months if the patient remained MRD-positive and was still being treated.

For patients with VGPR:

• The first bone marrow assessment was performed when VGPR was confirmed at the second time point or at a later time point, as per investigator judgment, based on kinetics of M protein decrease and/or if a plateau phase was reached (defined as variation less than 20% over 12 weeks).
• If MRD was positive at the first bone marrow assessment, a second assessment was performed 3 months later (3 cycles) to identify late negativity.
• If MRD remained positive at the second bone marrow assessment, a third assessment was performed while the patient had VGPR; the timing to perform the third assessment could be postponed until CR was achieved.
• If the patient had CR and was MRD-positive at the last bone marrow assessment performed during VGPR, a bone marrow assessment was performed to confirm CR.
• If the patient had CR and the first bone marrow assessment was MRD-positive, the additional protocol-specified bone marrow assessment could be discussed with the patient.

Duration of Response

DOR was a secondary outcome, defined as the time from date of first IRC-determined response that was subsequently confirmed for patients achieving PR or better to the date of first documented PD determined by IRC, or death, whichever occurred first. DOR was censored at the date of last valid disease assessment not showing PD performed before initiation of a new antimyeloma treatment (if any) or the analysis cut-off date, whichever occurred first.

The observation period for safety variables was divided into the pre-treatment period (from signed informed consent up to first dose of study medication), the treatment period (time from first dose of study treatment until last dose of study treatment plus 30 days) and the post-treatment period (starting the day after the end of the treatment period up to the end of the study).

Adverse Events

AE data were collected from time of signed informed consent to 30 days following last administration of study treatment. All AEs were graded according to NCI CTCAE v4.03. Treatment-emergent AEs were defined as AEs that developed, worsened (according to investigator opinion), or became serious during this period.

Adverse Events of Special Interest

AEs of special interest included infusion reactions of grade 3 or greater, and reports of pregnancy (including female partners of male patients), and symptomatic overdose with the study treatment or noninvestigational medical products.

Primary Outcome

Secondary Outcomes

For the key binary secondary end points, the significance levels at the interim and final analyses were determined using alpha-spending function specific to each end point, except if the information fraction was 100% at the interim analysis of PFS (i.e., information on binary secondary end points were available for every patient). For ORR and rate of VGPR or better, a Pocock-type boundary was used. For rates of CR and MRD negativity, the O’Brien-Fleming alpha-spending function was used. If the null hypothesis for any key efficacy end point was rejected at the interim PFS analysis, then any subsequent key efficacy end point was not tested until the final PFS analysis. CIs at (1 to 2 alpha) percent level (alpha being the adjusted 1-sided nominal significance level using the alpha-spending function specific to each outcome at final and interim analyses) were determined for ORR, rate of VGPR or better, CR rate (including sCR) and MRD negativity rate using the Clopper-Pearson method. Since the information fraction was 100% at the interim analysis, significance levels were 0.025 for all binary secondary outcomes. These outcomes were compared between treatment groups using the Cochran-Mantel-Haenszel test stratified by stratification factors, as entered in the IRT. A sensitivity analysis of key secondary outcomes (except MRD negativity rate) was performed using investigator assessment.

Analyses of “other” secondary outcomes was for descriptive purposes only, and any testing that was carried out was considered exploratory. Time-to-event outcomes were assessed using Kaplan–Meier methods. For HRQoL outcomes, a descriptive summary of results was provided at each visit, including end of treatment and an assessment performed 90 days after the last study treatment. Mean change from baseline was also provided, and data were reported for the safety population.

Table 9

Statistical Analysis of Efficacy End Points in the IKEMA Trial.

Protocol Amendments

Notable protocol amendments included Amendment 3 (February 8, 2018), adding an objective to evaluate the rate of VGPR or better in patients who were MRD-negative (original outcome evaluated rate of CR in MRD-negative patients). The HR estimate and corresponding 95% CI was changed to HR and corresponding (1 to 2 alpha) percent level (alpha being the 1-sided nominal significance level: alpha = 0.023 at final analysis and 0.005 at PFS interim analysis). Amendment 3 also clarified that further antimyeloma therapies were collected until OS analysis and added delay for carfilzomib infusion if grade 3 to 4 infusion reactions occurred during the isatuximab infusion. Amendment 4 (July 2, 2018) allowed that, in the absence of radiological and M protein progression, if clinical and biologic data provided clear evidence of clinical progression, the IRC could consider clinical progression as a PFS event. Amendment 7 (November 13, 2019) changed censoring rules for the primary PFS analysis; the PFS definition was modified such that the date of initiation of further antimyeloma treatment was considered when determining the cut-off date for PFS.

Analysis Populations

The randomized population included all patients randomized into the study, regardless of whether they received study treatment. The ITT population is the randomized population, and patients were analyzed based on the treatment they were randomized to, regardless of whether they received any treatment or the treatment to which they were randomized. This was the primary population for all efficacy analyses. The safety population included all patients in the ITT population who received at least 1 dose of study drug, or a partial dose of study drug. This population was the primary population for analysis of all safety parameters, and analyses were based on the treatment actually received.

Protocol Deviations

The most common reason for protocol deviations was wrong stratum of randomization, which occurred in 7.3% of IsaKd patients and 16.3% of Kd patients. This was due to violations in number of prior lines (IsaKd: 2.8% versus Kd: 5.7%) and R-ISS stage (4.5% versus 10.6%, respectively).

Progression-Free Survival

PFS was the primary outcome of the IKEMA trial. At the interim analysis (median follow-up of 20.73 months), median PFS was not reached in the IsaKd group and was 19.15 months (95% CI, 15.77 to not calculable) in the Kd group, for a stratified HR of 0.531 (99% CI, 0.318 to 0.889) and a P value by log-rank test of P = 0.0007, well below the planned threshold for statistical significance of P = 0.005 (Table 11). At this time, 48 (26.8%) patients in the IsaKd group and 55 (44.7%) of patients in the Kd group had a PFS event (Figure 3).

Figure 3

Progression-Free Survival From IKEMA Trial.

A number of sensitivity analyses of the primary outcome were performed, including per IRC without censoring for further antimyeloma treatment; per investigator, including symptomatic deterioration as an event; per investigator, ignoring symptomatic deterioration; per IRC, including initiation of further antimyeloma treatment as an event; per IRC stratified by stratification factors entered in the eCRF; and per IRC with censoring of progression or death occurring more than 8 weeks after last valid disease assessment. All results were consistent with that of the primary analysis.

Subgroup analyses for the primary outcome are summarized in Table 19, Appendix 3. Based on the interaction P values, there were no clear differences in response between subgroups of interest for this review.

Overall Survival

OS will be assessed at the end of the study; therefore, no data for median OS were reported at the time of the interim analysis. At the time of the interim analysis, after a median follow-up of 20.73 months at the data cut-off date, 31 (17.3%) patients had died in the IsaKd group and 25 had died (20.3%) in the Kd group.

Health-Related Quality of Life

HRQoL was assessed using the EORTC QLQ-C30, the EORTC QLQ-MY20, and the EQ-5D-5L instruments. Results were reported for change from baseline in each treatment group, but no statistical analyses were planned. Results from the EORTC QLQ-C30 global QoL scale are provided in Figure 4, and the results for the remainder of the EORTC QLQ-C30 scales and other questionnaires are available in Appendix 3, presented in graphs.

Compliance with filling out the EORTC QLQ-C30 was reported to be between 87.5% and 100% at each treatment cycle. Baseline scores were generally consistent between the treatment groups, although there were differences of up to 5 points for some subscales, including the EORTC QLQ-C30 global QoL scale. Mean changes from baseline in global QoL are presented in Figure 4. There was a considerable decline in patients sampled over time, largely due to attrition. Graphs in Appendix 3 provide an indication of the change in scores over time. Generally, there was little change from baseline in global QoL scores over time in the IsaKd group, while there were numerical increases from baseline over time in the Kd group. Increases in the global QoL score indicated an improvement in HRQoL. The threshold estimates for a small deterioration in the global health status is −5 to −10.

Figure 4

Mean Change From Baseline in EORTC QLQ-C30 Global QoL Scores Over Time.

Overall Response Rate

The ORR was assessed as 2 secondary outcomes, as responders (patients achieving either an sCR, CR, VGPR, or PR) and as patients achieving a VGPR or better, both by IRC assessment. The percentage of patients responding was 86.6% (95% CI, 0.8071 to 0.9122) in the IsaKd group and 82.9% (95% CI, 0.7509 to 0.8911) in the Kd group, and the difference between groups was not statistically significant (P = 0.1930) (Table 11). As this was the second outcome in the statistical hierarchy, testing was to have halted for subsequent outcomes; however, the sponsor continued to conduct testing and report P values for descriptive purposes. The percentage of patients achieving VGPR or better was 72.6% (95% CI, 0.6547 to 0.7901) in the IsaKd group and 56.1% (95% CI, 0.4687 to 0.6503) in the Kd group. No patients achieved an sCR, while 39.7% of patients in the IsaKd group and 27.6% of patients in the Kd group achieved a CR, and 33.0% and 28.5% of patients, respectively, achieved a VGPR. The results for ORR by investigator assessment were consistent with those conducted by the IRC.

Depth of Response

Bone marrow samples for the assessment of MRD were collected for patients with VGPR or better and included 78% and 74% of patients with at least 1 evaluable MRD sample in the IsaKd and Kd groups, respectively. MRD negativity (10⁻⁵ sensitivity level by central laboratory) was achieved by 29.6% of patients in the IsaKd group and 13.0% of patients in the Kd group (Table 11). At a sensitivity of 10⁻⁴, MRD negativity occurred in 40.2% and 24.4% of patients, respectively; and at a sensitivity of 10⁻⁶, 8.4% and 1.6% of patients, respectively.

Median Duration of Response

The median DOR was calculated based on 155 patients in the IsaKd group and 102 patients in the Kd group. The median DOR was not yet reached in either treatment group, and the HR was 0.425 (95% CI, 0.269 to 0.672) (Table 11).

Time to Response

The median TTR was 1.08 months (95% CI, 1.05 to 1.12) in the IsaKd group and 1.12 months (95% CI, 1.05 to 1.18) in the Kd group, for a stratified HR of 1.143 (95% CI, 0.888 to 1.471) (Table 11).

Health Resource Utilization

This outcome was not assessed.

Table 11

Summary of Efficacy.

Adverse Events

There were 97.2% of patients in the IsaKd group and 95.9% of patients in the Kd group who had at least 1 treatment-emergent AE; 76.8% versus 67.2%, respectively, who had at least a grade 3 AE; and 3.4% versus 3.3% who had a grade 5 AE (Table 12). The most common AE in the IsaKd group was infusion-related reaction, which occurred in 44.6% of patients in the IsaKd group and 3.3% of patients in the Kd group. Other common AEs, IsaKd versus Kd, were hypertension (36.7% versus 31.1%), diarrhea (36.2% versus 28.7%), upper respiratory tract infection (36.2% versus 23.8%), fatigue (28.2% versus 18.9%), and dyspnea (27.7% versus 21.3%). The most common AEs of grade 3 or higher were hypertension (IsaKd: 20.3% versus Kd: 19.7%) and pneumonia (IsaKd: 16.4% versus Kd: 12.3%).

Serious Adverse Events

SAEs occurred in 59.3% of patients in the IsaKd group and 57.4% of patients in the Kd group (Table 12). The most common SAE was pneumonia (IsaKd: 18.1% versus Kd: 11.5%).

Withdrawal Due to Adverse Events

There were 8.5% of patients in the IsaKd group and 13.9% of patients in the Kd group who had an AE leading to definitive treatment discontinuation, which was defined as discontinuing all treatments (Table 12). One patient discontinued isatuximab due to an AE. There were 14.7% of patients in the IsaKd group and 0.8% of patients in the Kd group who discontinued carfilzomib, and 6.2% of patients in the IsaKd group and 3.3% of patients in the Kd group who discontinued dexamethasone.

Notable Harms

Among notable harms, respiratory tract infections occurred in 83.1% of patients in the IsaKd group and 73.8% of patients in the Kd group, and these were grade 3 or greater events in 32.2% versus 23.8% of patients, respectively (Table 12). Cardiac disorders occurred in 7.3% of patients in the IsaKd group and 5.7% of patients in the Kd group; second primary malignancies (solid non-skin) in 2.8% versus 3.3% of patients, respectively; and second primary malignancies (solid, skin) in 5.1% versus 2.5% of patients, respectively. There were no hematologic malignancies reported in either treatment group. Events of decreased neutrophil count occurred in 54.8% of patients in the IsaKd group versus 43.4% of patients in the Kd group, and grade 3 or greater events occurred in 19.2% versus 7.4% of patients, respectively. Events of decreased platelet counts occurred in 94.4% of IsaKd patients and 87.7% of Kd patients, and these were grade 3 or greater events in 29.9% versus 23.8% of patients, respectively.

Table 12

Summary of Harms.

Internal Validity

The IKEMA trial appears to be a well-conducted study. It was planned with adequate power to meet the goals of the primary outcome (PFS), and enrolment targets were met, with a target enrolment of 300 patients and 302 patients randomized. The assumptions made when performing power calculations were based on a previous study³⁰ and seemed reasonable. The randomization method appeared adequate, and randomization was stratified based on important prognostic factors in an effort to minimize important differences in baseline characteristics between groups. There were, however, some limitations and potential sources of bias, which are outlined in this section.

Results for the primary outcome were based on a pre-specified interim analysis (103 PFS events), which occurred at a 65% information fraction, relative to the planned final analysis (159 PFS events). As a result, there is a risk that the treatment effect for PFS has been overestimated. However, given the statistical and clinical significance of the difference observed between the IsaKd and Kd groups, this is likely not a major concern.

The included trial was open-label, and lack of blinding may bias results, particularly for patient-reported outcomes such as HRQoL. Objective clinical outcomes such as PFS and ORR are less likely to be influenced by lack of blinding. Reporting of harms may also be influenced by blinding, as patients may anticipate known adverse effects of IsaKd and thus may be more likely to report them when they do occur. Physician knowledge of their patient’s assigned treatment may also affect the way they manage their patient, and patient knowledge of their assigned treatment may make them more or less likely to remain in the study. Assessment of radiological and laboratory-related outcomes (such as ORR) was conducted by a blinded IRC, and therefore would not be subject to bias from lack of blinding.

Multiplicity was controlled for in the study by use of a hierarchical testing procedure. The hierarchy failed early (after the primary outcome). Therefore, as described in the protocol, testing should have stopped there. However, the sponsor continued to report P values for other outcomes in the hierarchy. The P values that were reported beyond the failure of the hierarchy should not be used to draw conclusions. Subgroup analyses were not adjusted for multiplicity and therefore should not be used to draw conclusions. Otherwise, subgroup analyses were planned a priori, and these included many of the subgroups included in the protocol for this review. Most subgroup results for PFS were consistent with the primary analysis results in favour of IsaKd, although some of these analyses were limited by small numbers.

There was a relatively large numerical difference between groups in patients discontinuing the trial early, with fewer discontinuations in the IsaKd group than the Kd group, and a large number of patients overall who discontinued. The largest number of these discontinuations were due to PD. This was expected, given that discontinuations due to PD were more frequent in the Kd group. A large difference in attrition between study groups can influence interpretation of harms outcomes, as the IsaKd group had longer exposure to study drug than the Kd group.

There does not appear to have been any accounting for missing data in analysis of efficacy outcomes such as ORR and patient-reported outcomes. For example, missing data for ORR are often counted as nonresponders. However, this procedure was not explicitly stated in the study protocol, although it was presumably followed. For best overall response, 2% of patient data in the IsaKd group and 4% in the Kd group were described as “not available/not assessed;” these patients do not appear to have been counted as responders, and this is appropriate.

There were some notable differences between groups in baseline characteristics. There were a larger percentage of White patients in the IsaKd group than in the Kd group (73.2% versus 67.5%) and a smaller percentage of Asian patients (14.5% versus 19.5%, respectively). It is unclear whether race plays a significant role in response to therapy in MM. A smaller percentage of IsaKd patients had MM at ISS stage I compared to Kd patients (49.7% versus 57.7%), and a larger percentage had MM at stage II (35.2% versus 25.2%). Although the percentage of patients with MM at R-ISS stage I was similar between IsaKd and Kd groups at baseline (25.1% versus 26.8%, respectively), there were more patients with MM at R-ISS stage II in the IsaKd group (61.5% versus 56.9%). Although R-ISS score was a stratification factor, the cut point was stage I or II versus III or not classified. Given that, in both cases, there were more patients with MM at stage II in the IsaKd group than in the Kd group, this could suggest that patients in the IsaKd had more advanced disease than those in the Kd group. However, the importance of this is unclear, although it could bias efficacy results against IsaKd. There were fewer patients in the IsaKd group than in the Kd group who had relapsed and refractory MM (68.2% versus 76.4%) and more patients who had relapsed MM (31.8% versus 23.6%). There were more patients in the IsaKd than in the Kd group who had 2 prior lines of therapy (35.8% versus 29.3%) and fewer who had 3 prior lines of therapy (18.4% versus 24.4%). Although the number of prior lines of therapy was a stratification factor, the cut point was 1 prior line of therapy versus more than 1 prior lines. A larger number of IsaKd than Kd patients had prior treatment with alkylating neoplastic drugs (94.4% versus 82.1%) and proteasome inhibitors (92.7% versus 85.4%), while fewer had been treated with immunomodulatory drugs (76.0% versus 81.3%).

External Validity

The IKEMA trial included all of the outcomes of most importance to patients. However, because the study is ongoing, OS data were not yet available (anticipated in 2023) and the difference in HRQoL outcomes between the groups was not formally assessed. The lack of formal assessment of HRQoL precludes any conclusions from being drawn about this important outcome, and this should therefore be considered an important limitation of this study. The lack of formal analysis of OS at this time should also be considered a limitation, given the importance of this outcome. PFS is considered a surrogate for OS, and the clinical experts consulted by CADTH believed it was an important outcome for patients and an appropriate primary outcome. The definition of PFS was changed after a late protocol amendment, adding that the date of initiation of subsequent myeloma treatment was censored. This change was requested by regulatory bodies and is therefore unlikely to have been influenced by investigator awareness of study data.

The clinical experts consulted by CADTH believed that the patients included in IKEMA were younger, on average, than the real-world patients they would be expecting to treat with IsaKd in routine clinical settings. The clinical experts noted that clinical trials in MM generally enrol patients who are younger and healthier. They also pointed out that the ECOG performance status in IKEMA is also better than 1 would expect to see in practice. Results from a planned subgroup analysis performed by the sponsor do not suggest that age affects response to IsaKd. Otherwise, the demographics and baseline disease characteristics seemed consistent with the patients the clinical experts would expect to receive the drug in Canadian clinical practice. Of the 302 patients included in the study, 8 patients were enrolled at Canadian sites.

The treatment regimens used in IKEMA appear consistent with what 1 would expect to see used in Canada, according to the clinical experts consulted by CADTH, and were aligned with the dosage in the Health Canada Product Monograph.

Objectives

The sponsor’s ITC aimed to compare the efficacy, as measured by PFS, OS, and ORR, among the following treatments for patients with RRMM:

• Bortezomib + dexamethasone (Vd)
• Bortezomib + pomalidomide + dexamethasone (VPd)
• Cyclophosphamide + Kd (CKd)
• Cyclophosphamide + Rd (CRd)
• Cyclophosphamide + Vd (CVd)
• Daratumumab + Kd (DKd)
• Daratumumab + Vd (DVd)
• Elotuzumab + Vd (EloVd)
• IsaKd
• Kd
• Panobinostat + Vd (PanVd)

Study Selection Methods

The criteria used for inclusion and exclusion of studies in the systematic review are described in Table 14. The search was done in June 2020 using Ovid MEDLINE, Embase, CENTRAL, CDSR, DARE, Northern Light Life Sciences Conference Abstracts, Clinicaltrials.gov, NICE, CADTH, EMA, and FDA websites. In addition, the bibliographies of review articles and meta-analyses were reviewed to identify additional studies. An initial screening step was conducted by 2 reviewers independently, based on the titles and abstracts of the identified studies. If it could be determined from the abstract that the study did not match the eligibility criteria, then it was excluded. Otherwise, the studies proceeded to the full-text screening step. The full-text screen was done by 2 reviewers independently, and discrepancies were solved by a third reviewer.

Data extraction was done using a custom data extraction sheet in Excel by a primary reviewer, and an audit was performed by a second reviewer. Any disagreements between the 2 reviewers were adjudicated by a third reviewer. The risk of bias was assessed following the NICE criteria for assessment of risk of bias for RCTs. The study quality was used for descriptive purposes and not used to justify exclusion of studies.

The efficacy outcomes were PFS, OS, and ORR. For each study, the following were extracted: the outcome definition and how it was assessed, the unit of measurement, the number of individuals included in the analysis, the size of the estimated effect, and a measure of precision for each estimated effect.

For included studies, the risk of bias was assessed using the NICE criteria for assessment of risk of bias for RCTs. The assessment criteria were as follows:

• Was the randomization method adequate?
• Was the allocation adequately concealed?
• Were the groups similar at the outset of the study in terms of prognostic factors; for example, severity of disease?
• Were the care providers, participants, and outcome assessors blind to treatment allocation? If any of these people were not blind to treatment allocation, what might be the likely impact on the risk of bias (for each outcome)?
• Were there any unexpected imbalances in dropouts between groups? If so, were they explained or adjusted for?
• Was there any evidence that the authors measured more outcomes than they reported?
• Did the analysis include an intention-to-treat analysis? If so, was this appropriate and were appropriate methods used to account for missing data?

Summary of Included Studies

A total of 11,635 records were assessed for eligibility; following this review, a total of 18 RCTs were included as eligible. The heterogeneity of the included trials was assessed according to the trial design, patient characteristics, and outcome data. All studies were determined to have low to medium risk of bias, and all but 2 studies were open-label design. Quality assessment was done systematically using a standardized set of criteria. Heterogeneity assessments and comparisons are summarized in Table 15. Inclusion and exclusion criteria were, in general, consistent across trials. All trials included patients with relapsed or refractory MM and at least 1 prior therapy. In general, patients could not have had prior treatment with the investigational drug or be refractory to the other drugs under study. Some variation was observed between trials, however, in the requirements for previous treatment, specifically lenalidomide and bortezomib. Many of the baseline characteristics were not reported among the trials, which makes it difficult to assess heterogeneity.

Four trials were excluded from the evidence network due to heterogeneity in the number of prior lines of treatment between them and the other trials. Of the remaining trials, 12 were in a connected network with sufficient data available for the NMA, while 4 were in a disconnected network. Of the 4 in the disconnected network, 2 were excluded from the MAICs, because of the inclusion of primary refractory patients in 1, and a sample size that was too small in the other.

Table 14

Assessment of Homogeneity for Indirect Treatment Comparisons.

Network Meta-Analysis Methods

Separate NMA models were fit for PFS, OS, and ORR. The models were Bayesian fixed-effects mixed treatment comparisons, as recommended by the NICE DSU.³¹ These were used instead of random-effects models because the random-effects models did not converge due to sparsity in the network. The models were estimated using MCMC as implemented in JAGS and R software, using noninformative priors, 10,000 burn-in iterations, and 50,000 posterior iterations for 3 independent chains. Methods to assess convergence were not specified. For the PFS and OS outcomes, the analysis was based on HR estimates under the proportional hazards assumption and used normal likelihood models with a linear link to model the log HRs. For the ORR results, fixed-effects models with the binomial likelihood and logit link were used. In addition to the HRs, modelling of the time-to-event outcomes using fractional polynomials embedded within the normal likelihood and linear link functions on the log hazard scale was performed. This approach relaxed the proportional hazards assumption and allowed estimates of HRs to vary over follow-up time. Model fit was assessed using the deviance information criterion. One sensitivity analysis was performed in the ORR analysis by including trials that evaluated treatments in the second-line-only population.

Table 15

Network Meta-Analysis Methods.

Network Meta-Analysis Results

|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||

|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| 　|　 |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| ||||||||||||||||||||||||||||||||||||||||||||||

|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| ||||||||||

|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||.

Figure 5

Evidence Network for PFS — Redacted.

Figure 6

Effect of IsaKd on PFS Relative to Other Treatments — Redacted.

Figure 7

Evidence Network for OS and ORR — Redacted.

Figure 8

Effect of IsaKd on OS Relative to Other Treatments — Redacted.

Figure 9

Effect of IsaKd on ORR Relative to Other Treatments — Redacted.

Matching-Adjusted Indirect Comparison Analysis Methods

|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| || |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| 　|　 |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||

Matching-Adjusted Indirect Comparison Results

|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| || |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| ||||||||||||||||||||||||||||||||||||||||||||||||||||||

|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| 　|　 |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| 　|　 |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| ||||

|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| 　|　 |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||

|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||

Table 16

Results of Matching-Adjusted Indirect Comparisons — Redacted.

Network Meta-Analysis Critical Appraisal

The ITC methods and results, in general, were reported thoroughly and adequately for the purposes of this review, with some exceptions, as noted in this section. The study selection was pre-specified with clear criteria and conducted in a rigorous manner. Heterogeneity and risk of bias were assessed. The trial populations were relatively homogenous in age, ECOG status, race, and gender. However, there were some concerns from the clinical experts regarding heterogeneity in the prior treatments received. Specifically, prior lenalidomide use is likely a key effect modifier that differs between trials and greatly increases the uncertainty in the findings. In addition, studies included in the network were conducted over a wide span of time during which the treatment approach for MM has rapidly evolved. Thus, the time span of these trials may further introduce bias to the comparisons in the ITCs. ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||

The study authors noted that results from fixed-effects models were reported, since results from random-effects models were considered unreliable or unstable. This was likely due to the sparsity of the network. So the fixed-effects approach was considered to be the only reasonable choice. However, the key assumption of fixed-effects models — that the true treatment effect is common in all studies comparing the same treatment — is unlikely to be plausible. Since the ITCs were based on sparse networks, in most cases, each direct comparison was supported by data from a single trial. For the most part, the network estimate is based solely on the indirect estimate (||||||||||||||||||||||). Typically, 1 benefit of a NMA is that it yields more precise estimates because it combines information from the direct and indirect estimates. This is not the case in the current NMAs, as many of the estimates are quite imprecise. Assessment of convergence for the NMA models was not described in the ITC report; however, the noted issues with the random-effects model would imply that some convergence assessment was done.

|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| The statistical methods for the NMA were appropriate given the limitations of the network. |||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| ||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||

Matching-Adjusted Indirect Comparison Critical Appraisal

In the MAICs, the unanchored nature of the comparison imposes the assumption that all prognostic factors and effect modifiers are adequately adjusted for, which is unlikely to be the case.³⁶ |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||, and previous lenalidomide use, which was specifically noted as a likely effect modifier by 1 of the clinical experts, and prior treatment in general. As for the choice of the matching factors, it was based on internal expert opinion (rather than a survey of clinical experts) and availability and completeness of data in the trials (which is inconsistent with the NICE DSU guidelines, which recommend that all necessary factors be identified, regardless of availability in the data). Many of the matching factors were categorized or dichotomized, which discards information and may result in residual confounding and inadequate adjustment. |||||||||||||||||||||||||||||||||||||||||||||||||| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||

The reported effective sample sizes, and the skewness and outliers apparent in the visualizations of the weight distributions, suggest that the results may be heavily influenced by a small subset of patients from the IKEMA trial. Generalizability may be an issue due to the small sample size remaining after the exclusions and matching — the remaining patients and weighted sample are unlikely to be representative of the entire patient population.

ClinicalTrials.gov

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

[Search terms – isatuximab, Sarclisa, SAR-650984, or SAR650984]

WHO ICTRP

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

[Search terms – isatuximab, Sarclisa, SAR-650984, or SAR650984]

Health Canada’s Clinical Trials Database

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

[Search terms – isatuximab, Sarclisa, SAR-650984, or SAR650984]

EU Clinical Trials Register

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

[Search terms – isatuximab, Sarclisa, SAR-650984, or SAR650984]

Description

The European Organisation for Research and Treatment of Cancer (EORTC) Quality of Life Questionnaire-Core 30, or EORTC QLQ-C30, is one of the most commonly used patient-reported outcomes measures in oncology clinical trials.⁴⁸ It is a multi-dimensional, cancer-specific, evaluative measure of HRQoL. The core questionnaire of the EORTC QLQ-C30 consists of 30 questions that are scored to create 6 multi-item functional scales, 2 multi-item symptom scales, 6 single-item symptom scales, and a 2-item quality of life (QoL) scale, as outlined in Table 21.^25,49 Version 3.0 of the questionnaire, used in the included trial in this report, is the most current version. The questionnaire is available in more than 100 different languages and has been used in more than 3,000 studies.⁵⁰

Table 21

Scales of EORTC QLQ-C30.

Scoring

The EORTC QLQ-C30 uses a 1-week recall period in assessing function and symptoms. Most questions have 4 response options (“not at all,” “a little,” “quite a bit,” “very much”), with scores on these items ranging from 1 to 4.²⁵ For the 2 items that form the global QoL scale, however, the response format is a 7-point Likert-type scale, with anchors between 1 (very poor) and 7 (excellent).

Raw scores for each scale are computed as the average of the items that contribute to a particular scale.²⁵ This scaling approach is based upon the assumption that it is appropriate to provide equal weighting to each item that comprises a scale. There is also an assumption that, for each item, the interval between response options is equal (for example, the difference in score between “not at all” and “a little” is the same as “a little” and “quite a bit,” at a value of one unit). Each raw scale score is converted to a standardized score that ranges from 0 to 100 using a linear transformation, with a higher score reflecting better function on the function scales, higher symptoms on the symptom scales, and better quality of life (i.e., higher scores simply reflect higher levels of response on that scale). Thus, a decline in score on the symptom scale would reflect an improvement, whereas an increase in score on the function and QoL scale would reflect an improvement. According to the EORTC QLQ-C30’s scoring algorithm, if there are missing items for a scale (i.e., the participant did not provide a response), the score for the scale can still be computed if there are responses for at least one-half of the items. In calculating the scale score, the missing items are simply ignored — an approach that assumes that the missing items have values equal to the average of those items for what the respondent completed.

Validity

Osborne et al. (2012)³⁹ assessed the reported construct and criterion validity of the EORTC QLQ-C30. A systematic review was conducted to identify HRQoL tools validated for use in MM; identify issues important to HRQoL from the point of view of patients with myeloma; describe the measurement properties of each HRQoL tool; evaluate the content validity of HRQoL tools in terms of their ability to capture all issues important to patients and to explore the suitability of each HRQoL tool for use in different settings. Results of the systematic review showed that all subscales of patients with MM were shown to be impaired compared to the general population.^51-55 General QoL scales significantly improved with increasing time following hematopoietic SCT.⁵⁵ Sixty-seven percent and 43% of patients scored below the 10th percentile for the physical functioning and global QoL subscales, respectively.⁵⁴ Functional subscales and global QoL were found to be lower in MM than in general hematology populations.⁵³ The subscales for pain, fatigue, physical and global QoL were able to discriminate between those who improved versus those who were stable/deteriorated.⁵⁶ All subscales except the single-item diarrhea scale discriminated between MM patients with different performance status and response status.⁵⁷ There were significant differences in global QoL between the different treatment arms in 2 examined trials of patients with MM.^58,59 Additionally, there was no correlation of any subscale with the SEIQoL‐Index (an instrument which allows patients to select the 5 most important domains for their present QoL and measures their satisfaction in these domains) suggesting independence.⁴⁰

Reliability

A sample of MM patients (n = 89) from 2 tertiary hospitals in Greece were surveyed with the EORTC QLQ-C30 and various demographic and disease-related questions.⁴¹ Internal consistency of the QLQ-C30 was assessed in this population. The 5 functional scales reported an internal consistency of greater than 0.7 (range, 0.77 to 0.90).7 except for cognitive function (α = 0.57). The global health status/HRQoL scale reported an internal consistency (α) of 0.92. Of the symptom scales, fatigue (α = 0.89), nausea and vomiting (α = 0.74) and pain (α = 0.80) were assessed for internal consistency, and all were considered acceptable. The 5 symptom scales/items of the core QLQ-C30, that is, nausea/vomiting, appetite loss, constipation, diarrhea, and financial difficulties suffered from high (> 50%) floor scores, implying a lack of these symptoms in this sample, but also suggesting an underlying reduced discriminative ability. Conversely, no ceiling effects were observed on the core instrument despite 3 scales being close to the threshold value (role, cognitive, and social functioning).

Responsiveness to Change

One study by Kvam et al. (2011)⁴² assessed HRQoL in patients with MM (n = 239) in Norway using the global health ⁄ QoL domains of the EORTC QLQ-C30. To assess responsiveness, the study used the global rating of change (GRC) to identify whether MM patients have changed over time. A Wilcoxon signed-rank test for pair differences was used to calculate the significance of differences in the mean score changes between baseline (T1) and after 3 months (T2). Due to the small sample sizes in some of the GRC categories, data were pooled into the categories improved, unchanged, and deteriorated to yield sufficient numbers of cases in each category. “Improved” represented patients ‘who reported themselves as improved’ and similarly for deteriorated and unchanged patients. To assess the magnitude of the difference in scores between patients who improved ⁄ unchanged ⁄ deteriorated, standardized response means (SRMs) were calculated by dividing the mean score changes by the SD of the change. This was compared against Cohen’s theory for interpreting the magnitude of mean differences in HRQoL scores, which suggests that a change of 0.20 represents a small change, 0.50 a moderate change, and > 0.80 a large change.

In patients rating themselves as unchanged, mean score changes clustered around 0, and the SRMs were negligible.⁴² MM patients who deteriorated reported lower global QoL scores at T2 compared with T1. The global QoL scale of the EORTC QLQ-C30 was the most responsive in deteriorating patients (SRM 0.57).

MID

A study⁴³ examined 118 published studies on various types of cancer such as breast, lung, or head and neck as well as clinician expert input to evaluate meaningful differences and magnitude of change in the EORTC QLQ-C30 scores. A meta-analysis was conducted to estimate a weighted average change within each size class for large, medium, small, and trivial changes. Small changes indicated a subtle, clinically relevant change. The calculations or symptom subscales were reversed to achieve consistency in improvement or deteriorations over time across all scales. MIDs for improvement and deterioration for small changes in QoL are shown in Table 19.

One study⁵⁶ assessed the MID of the EORTC QLQ-C30 by recruiting 239 patients with MM to complete the EORTC QLQ-C30 at baseline (T1) and after 3 months (T2). At T2, patients were asked if they had noticed any change in the domains pain, fatigue, physical function, and global QoL. The MID was determined using the mean score changes as observed by the patients stating improvement or deterioration for each domain. A combination anchor and distribution approach were used. The MIDs (SD) for patients rating themselves as improved was 6.2 (15.3) for physical function, -14.7 (35.9) for pain, -13.5 (24.7) for fatigue and 7.6 (23.7) for QoL. Patients reporting deterioration had MIDs (SD) of 8.6 (23.4) for fatigue, 17.3 (23.1) for pain, -12.8 (19.2) for physical function, and -12.1 (21.2) for QoL. However, there was considerable variation in the observed scores.

Description

The EQ-5D-5L is a generic preference-based HRQoL instrument that has been applied to a wide range of health conditions and treatments including MM.²⁷ The EQ-5D-5L was developed by the EuroQol Group as an improvement to the EQ-5D 3 level (EQ-5D-3L), to measure small and medium health changes and reduce ceiling effects. The instrument is comprised of 5 dimensions: mobility, self-care, usual activities, pain/discomfort, and anxiety/depression. Each dimension is rated on 5 levels: level 1 “no problems,” level 2 “slight problems,” level 3 “moderate problems,” level 4 “severe problems,” and level 5 “extreme problems” or “unable to perform.”²⁸ A total of 3,125 unique health states are possible, with 55555 representing the worst health state and 11111 representing the best state.²⁸ The corresponding scoring of EQ-5D-5L health states is based on a scoring algorithm that is derived from preference data obtained from interviews using choice-based techniques (e.g., time trade-off) and discrete choice experiment tasks. The lowest score varies depending on the scoring algorithm used. The anchors are 0 (dead) and 1 (full health), however negative values are also allowed to represent health states that a society considers worse than death. As an example, a Canadian scoring algorithm results in a score of -0.148 for health state 55555 (worst health state). Another component of the EQ-5D-5L is a visual analogue scale (VAS), which asks respondents to rate their health on a visual scale from 0 (worst health imaginable) to 100 (best health imaginable).²⁷

The psychometric properties of the EQ-5D-5L have not been assessed in patients with MM; therefore, its validity and reliability have not been evaluated in this patient population of interest.

Responsiveness

Kvam et al. (2011)⁴² also assessed the responsiveness of the EQ-5D-3L using the aforementioned methods. The results found that the EQ-5D-3L was the most responsive among improved patients (SRM 0.43). The global QoL scale of the EQ-5D-3L for deteriorating patients had a SRM 0.45. The study also assessed the presence of floor and ceiling effects for EQ-5D-3L. A small floor or ceiling effect was defined as < 15% of patients attaining the worst and best health state and a serious effect was defined as > 15% of patients attaining these states. The results found small floor and ceiling effects for the EQ-5D-3L and noted that 10% of the patients achieved the maximum score (ceiling effect).

MID

Kvam et al. (2011)⁴² used both distribution and anchor-based approaches for the whole sample (n = 239) to determine MIDs for the EQ-5D-3L. The distribution-based approach was determined by multiplying the SDs at baseline and expected differences in scores associated with small (0.2), moderate (0.5), or large (0.8) changes as per Cohen’s criteria for interpreting the absolute magnitude of a change. From this analysis, using the small effect size as a value of MIDs, the expected MID score was 0.04 for the EQ-5D-3L. The anchor-based approach used the GRC as previously described as the anchor. From this analysis, an MID of 0.08 (95% CI: 0.04 to 0.12) in MM patients who thought their HRQoL improved, and -0.10 (95% CI: -0.16 to -0.04) in those who thought their HRQoL deteriorated.

To estimate the MID values of the EQ-5D-3L for each country-specific scoring algorithm, a simulation-based approach based on instrument-defined single-level transitions has been used. The simulation-based instrument-defined generally accepted MID estimate (mean ± SD) for Canada is 0.056 ± 0.011.⁶⁰

Description

The questions in the QLQ-MY20 also have 4 response options (“not at all,” “a little,” “quite a bit,” “very much”), with scores on these items ranging from 1 to 4.^26,46 Raw scores for each scale are computed as the average of the items that contribute to a particular scale. This scaling approach is based upon the assumption that it is appropriate to provide equal weighting to each item that comprises a scale.²⁵ There is also an assumption that, for each item, the interval between response options is equal (for example, the difference in score between “not at all” and “a little” is the same as “a little” and “quite a bit,” at a value of one unit). Each raw scale score is converted to a standardized score that ranges from 0 to 100 using a linear transformation, with a higher score reflecting better function on the function scales, higher symptoms on the symptom scales, and better quality of life (i.e., higher scores simply reflect higher levels of response on that scale). Thus, a decline in score on the symptom scale would reflect an improvement, whereas an increase in score on the function and quality of life scale would reflect an improvement.⁴¹

Validity & Responsiveness to Change

A cross-sectional study in Brazil assessed the validity of the EORTC QLQ-MY20 questionnaire among 225 MM patients recruited from 3 clinics, more than 70% of which were more than the age of 60.⁴⁶ Both convergent and divergent validity were assessed between the EORTC QLQ-C30 and QLQ-MY20 instruments using Spearman’s correlation coefficient. Correlation coefficients were defined as <0.4 for a weak correlation and ≥ 0.4 for a moderate to strong correlation. For convergent validity, a strong correlation was hypothesized between the 2 instruments’ symptom scales. For divergent validity, a weak correlation was hypothesized between the EORTC QLQ-C30 symptom scale and the QLQ-MY20 function scale. Results demonstrated a strong correlation for convergent validity and a weak correlation for divergent validity as hypothesized with correlation coefficients of 0.771 and -0.386 for each assessment, respectively.

Cocks et al. (2007)⁴⁴ assessed the validity and sensitivity of the EORTC QLQ-MY20 questionnaire. MM patients were recruited prospectively from ongoing or new clinical trials at the time of the study rather than setting up a separate questionnaire validation study. Trials had to be for newly diagnosed or relapsed/refractory MM patients with a minimum prognosis of 3 months. Maintenance therapy trials were not eligible. Questionnaires were issued to patients at the baseline for each trial (randomization or registration) according to the procedures for that trial. A panel reviewed each protocol and identified an appropriate follow-up time point for the field study QoL assessment, this had to be during treatment or within 4 weeks of completion of treatment and was 3 or 6 months from randomization for all protocols. Responsiveness analysis compared the baseline and follow-up scale scores for the 137 (57% of total) patients who achieved at least a PR. The mean disease symptom score significantly decreased (a higher score indicates worse symptoms) from 31.9 (23.2) at baseline to 21.1 (18.7) at follow-up (p < 0.0001). The mean body image score significantly decreased over time (a higher score indicates better support/functioning) from 80.0 (29.8) at baseline to 63.2 (32.5) at follow-up (p < 0.0001). The EORTC QLQ-MY20 scale scores for patients with performance status 0, 1 or 2 (n = 200) were compared to those with performance status 3 or 4 (n = 25) at baseline. The performance status score is a measure of how disease impact’s a patient’s daily living abilities with a score ranging from 0 (“fully active”) to 5 (“dead”).⁶¹ Three of the scales from the module: disease symptoms, side effects of treatment and body image showed evidence of a significant difference (P = 0.0013) between the patient groups. The future perspective scale showed a trend toward a difference (p = 0.065).⁴⁴

Reliability

In total, 215 patients with MM were recruited from Imam Khomeini Hospital, Tehran.⁴⁵ Patients were eligible for the study if they were 18 years of age or older, had been diagnosed with MM at least one month prior to study entry, and were able to read and speak Persian/Farsi. Patients who had a life expectancy of less than 3 months or had serious cognitive problems (as assessed by the Mini-Mental State Examination) were excluded from the study. Participating patients were asked to complete the EORTC QLQ-MY20 3 times, at study entry, after 2 weeks, and again after 3 months. Internal consistency reliability reported was greater than 0.80 for all 3 multi-item scales (ranging from 0.82 to 0.93).⁴⁵ All 4 scales (including the single-item body image measure) had test-retest reliability of 0.85 or greater. The disease symptoms scale had the highest test-retest reliability (ICC=0.95).

The aforementioned study⁴⁶ in Brazil had a small sample of the studied population (n = 25) take the EORTC QLQ-MY20 at 2 time points (14 days apart) to measure test-retest reliability using intraclass correlation coefficient. ICC values between 0.6 and 0.8 were defined as satisfactory temporal stability and values > 0.8 suggested excellent temporal stability. Results found ICC values ranged from 0.41 to 0.79 for body image and diseases symptoms, respectively, indicating a moderate to high ability to create reproducible results. A sample of MM patients (n = 89) from 2 tertiary hospitals in Greece were surveyed with the EORTC QLQ-MY20 and various demographic and disease-related questions.⁴¹ Patients were approached during a scheduled visit to one of the hospitals. None were suffering from cancer metastases or severe comorbid conditions which could further compromise HRQoL. Internal consistency of the EORTC QLQ-MY20 was assessed in this population. Of the symptom scales, disease symptoms (α = 0.77) and side effects of treatment (α = 0.72) demonstrated adequate (adequate measured at α = 0.70) internal consistency. Future perspective was reported as an internal consistency of 0.80, side effects of treatment at an internal consistency of 0.72, and disease symptoms at 0.77. Body image was not assessed.

MID

A mixed‐methods study by Sully et al. (2019)⁴⁷ assessed MID scores of the EORTC QLQ-MY20 using patient interviews that were conducted alongside an analysis of existing trial data. Data were pooled across 3 clinical trials: ASPIRE,⁶² ENDEAVOR,³⁰ and CLARION⁶³ (n = 2,147). Each was a phase III, randomized, open-label study comparing carfilzomib‐based regimens, with a primary end point of progression‐free survival (PFS) in patients with relapsed or refractory MM. Patients were eligible for entry into the pooled sample if they had completed the EORTC QLQ‐C30 or QLQ‐MY20 at baseline plus at least one other of the following time points: mid‐treatment (MT), or end of treatment (EOT). Twenty patient interviews were conducted.

Two anchor‐based methods were applied, mirroring an approach in another pooled study of cancer patients.⁴⁷ Potential anchors were identified by the authors reviewing the clinical trial protocols to identify measures available across the trials at MT and EOT. Anchors were selected via review of the protocols and case report forms. Potential anchors were chosen if they were deemed to have conceptual overlap with any of the EORTC QLQ‐MY20 scales. Anchor‐based MIDs were estimated by calculating the mean change score of patients classified as improved and deteriorated according to anchor definitions. Patient‐reported anchors were the EORTC QLQ‐C30 Global Health Status/QoL scale and Functional Assessment of Cancer Therapy‐Gynecologic Oncology Group Neurotoxicity (FACT‐GOG‐Ntx). Clinical anchor measures were the Eastern Cooperative Oncology Group performance status (ECOG PS), matched adverse events (AEs) and peripheral neuropathy‐related AEs. Linear regression models were also fitted with EORTC QLQ‐MY20 score change as the outcome and a binary indicator of stable versus improved/worsened according to the anchor as a predictor, where the coefficient of this indicator was the MID estimate (incorporating the change score of stable patients). Diagnosis (newly diagnosed/relapsed) was accounted for in the model.⁴⁷

Based on this mixed-method study, the recommended MIDs for the EORTC QLQ-MY20 were as follows: Disease Symptoms, -10 and +10 points for improvement and deterioration, respectively; Side Effects of Treatment, -10 and +10 points for improvement and deterioration, respectively; Body Image, +13 points and -13 for improvement and worsening, respectively; and Future Perspective (body image) +9 and -9 points for improvement and deterioration respectively.⁴⁷