Selinexor with bortezomib and dexamethasone for previously treated multiple myeloma

For this evaluation, the company positioned selinexor plus bortezomib and dexamethasone (selinexor combination) as a second-line treatment for people whose condition is refractory to previous treatment with both daratumumab and lenalidomide. The clinical experts agreed with the company's positioning of selinexor combination as a second-line treatment option. They highlighted that most people who cannot have a stem cell transplant would be offered first-line treatment with daratumumab plus lenalidomide and dexamethasone. They said that the relevant comparator at second line is carfilzomib plus dexamethasone because this is the only option available to people whose condition is refractory to both daratumumab and lenalidomide. They explained that different factors would be considered when choosing between carfilzomib and selinexor such as comorbidities, individual preferences and ease of administration. Carfilzomib is associated with cardiac side effects and is administered by intravenous infusion in hospital. Selinexor, an oral tablet, is associated with gastrointestinal side effects and bortezomib is a subcutaneous medicine, so takes less time to administer in hospital.

The company also positioned selinexor combination at third line, for people who had 2 previous lines of any treatments. The clinical experts explained that, for people who can still have lenalidomide, ixazomib plus lenalidomide and dexamethasone (ixazomib combination) would be the preferred option. But they considered that, at third line, most people's condition would be refractory to lenalidomide. So, they explained that for these people, there are limited treatment options available. One treatment option is panobinostat plus bortezomib and dexamethasone (panobinostat combination). But they highlighted that, because of the toxicity associated with panobinostat, it is not often used. The clinical experts acknowledged the toxicity associated with selinexor. But they explained that emerging real-world data shows that a dose reduction of selinexor can help reduce gastrointestinal side effects and thrombocytopenia, with potentially the same level of clinical effectiveness. The company explained that Jagannath et al. (2023), a study analysing data from the main clinical trial (BOSTON, see section 3.6) showed that progression-free survival improved with dose reduction. It explained that a reason for this result could be that people continue to have selinexor for longer at the reduced dose. The clinical experts agreed that selinexor provides an option with a different mechanism of action. They also agreed that there are limited treatments available at third line, particularly for people with multiple myeloma that is refractory to lenalidomide.

The committee acknowledged that the company's positioning of selinexor combination is narrower than its marketing authorisation. It agreed with the company's positioning of selinexor combination at second and third line, and concluded that the choice of comparators was appropriate.

The clinical-effectiveness evidence for selinexor combination came from BOSTON, a phase 3, randomised, open-label, multi-national trial. The trial was stratified by previous proteasome inhibitor treatments, number of previous lines of treatment and the Revised International Staging System stage at entry. It included people aged at least 18 years with relapsed or refractory multiple myeloma, or both, who had already had 1 to 3 previous lines of treatment. They were randomised to selinexor combination (n=195) or bortezomib plus dexamethasone (n=207). Seventy-seven people randomised to the bortezomib plus dexamethasone group changed treatment to selinexor combination or selinexor plus dexamethasone after their condition progressed during the trial. The company used data from subgroups that had treatment at second line (49%) and at third line (32%). The remaining subgroup, which had treatment at fourth line (19%), was not included in the analyses for this evaluation (see section 3.3 and section 3.4). The primary outcome was progression-free survival assessed by an independent review committee that was blind to treatment group allocation. The committee noted that the average age of people in the subgroups was 67 years at second line and 65 years at third line. This was younger than people seen in the NHS, whose condition is usually diagnosed around 75 years (see section 3.2). The committee also noted that BOSTON was not statistically powered to find differences in outcomes in the subgroups. Also, the trial comparator was not relevant to the decision problem at second or third line. The committee noted that many people in the BOSTON subgroups had not had previous treatment with lenalidomide (68%). The committee was aware that most people having second- or third-line treatment in NHS clinical practice would have already had lenalidomide. So, it considered that there may be issues about how representative the population in BOSTON is to people likely to have selinexor combination in NHS clinical practice. These issues may lead to uncertainty in the generalisability of the results.

The company did Bayesian network meta-analyses using Markov chain Monte Carlo simulations to estimate the comparative effectiveness of selinexor combination to:

• carfilzomib plus dexamethasone at second line

• ixazomib combination at third line

• panobinostat combination at third line.
  
  The company chose to use random effects models because of the significant heterogeneity in the studies in the network meta-analyses. The EAG considered that the network meta-analyses used for the second-line comparison and the third-line comparison with panobinostat combination were appropriate. But it considered that the third-line network meta-analysis with ixazomib combination was at high risk of bias because of:

• the unanchored matching-adjusted indirect comparison (MAIC) used between pomalidomide plus dexamethasone (ICARIA‑MM study) and bortezomib plus dexamethasone (BOSTON)

• the double use of bortezomib plus dexamethasone BOSTON data to estimate hazard ratios

• using by-arm median progression-free survival data from the MM‑009 and MM‑010 trials

• potential violation of the proportional hazards' assumption for many comparisons in the networks for progression-free and overall survival

• including the MM-003 trial, which may not be representative of NHS clinical practice because people in this trial had an average of 5 previous lines of treatment

• substantial heterogeneity in some trials

• unadjusted crossover in some of the trials.
  
  The EAG preferred to use an unanchored MAIC with the ixazomib combination. The company explained that it had adopted a pragmatic approach to deal with the heterogeneity between the trials. It considered that the unanchored MAIC did not solve all the underlying uncertainty. It emphasised that the network meta-analysis is still its preferred approach because of the very small numbers included in the unanchored MAIC. It confirmed that the unanchored MAIC was not adjusted for subsequent treatments because no data was available. The committee agreed with the EAG that the network meta-analyses for the second-line comparison, and the third-line comparison with panobinostat combination were appropriate. It acknowledged the limitations of the unanchored MAIC for ixazomib combination, in particular, that it had not been adjusted for subsequent treatments, which may have affected overall survival. But it concluded that the unanchored MAIC was preferred for this third-line comparison with ixazomib combination because of the substantial limitations of the network meta-analysis.

The company presented a partitioned survival model with 3 health states: progression-free (including on and off treatment), progressed and death. The probability of being in each health state was calculated using extrapolated progression-free survival, overall survival and time-on-treatment curves. People started the model in the progression-free health state on second-line or third-line treatment. The model included a cycle length of 1 week with a half-cycle correction over a 35‑year time horizon. The committee concluded that the company's model structure was acceptable for decision making.

In its base case, the company modelled differences in overall survival between treatments based on data from BOSTON Kaplan–Meier curves and indirect treatment comparisons. The EAG's base case assumed no overall survival differences between treatments and used bortezomib plus dexamethasone as the baseline overall survival curve. The EAG considered that this was justified because the overall survival data from BOSTON was immature and uncertain. Also, there were no statistically significant overall survival differences for any of the comparisons. The EAG noted that an overall survival benefit likely includes varying effects of subsequent treatments on overall survival after disease progression. The company argued that an overall survival benefit was plausible. This was because selinexor combination provides a new mechanism of action, and a statistically significant improvement in overall survival was seen in the lenalidomide-refractory subgroup of BOSTON (see section 3.8). At the second committee meeting, the clinical experts agreed with the company that it is plausible for overall survival to improve, as shown with other treatments in this disease area. They added that this is because selinexor has a different mechanism of action, But they acknowledged that the clinical evidence did not show an overall survival benefit with selinexor combination compared with its comparators. The company considered that the EAG had taken a pessimistic and inconsistent approach by adopting the indirect treatment comparison results for progression-free survival but assuming no benefit for overall survival. The EAG reiterated that there were no statistically significant differences in overall survival in any of the comparisons (see section 3.8). It also noted that, for the lenalidomide-refractory subgroup, the analysis included people at all treatment lines (second to fourth line) and not just third line. Also, there was no available comparative data from indirect treatment comparisons for the lenalidomide-refractory subgroup. The EAG explained that there was a difference in the degree of uncertainty for progression-free survival and overall survival because progression-free survival was the primary outcome of the key trials. BOSTON's progression-free survival data was more mature than its overall survival data. The EAG noted that overall survival was confounded by using different subsequent treatments in the key trials. It also noted that BOSTON's overall survival data was sensitive to adjustments for treatment cross over. The committee considered that it had not been presented with evidence of overall survival benefit for selinexor combination compared with the relevant treatment options at second and third line. So, it considered the EAG's assumption of no differences in overall survival between treatments to be appropriate for decision making.

The company included the cost of subsequent treatments by using a weighted average of these treatments in BOSTON, and adjusted for treatments available in the NHS. The EAG considered that the subsequent treatments modelled by the company do not reflect NHS practice. The EAG instead used market share data provided by the company, and assumptions based on the NHS treatment pathway and adjusted for the proportion of people from BOSTON having subsequent treatments (80%). At the first committee meeting, the clinical experts explained that there is significant attrition with each line of treatment. They noted that the average age of diagnosis is 75 to 80 years. They also noted that some study results suggested that, at fourth and fifth line, only about 20% of people remain on treatment. The committee recalled the younger cohort included in BOSTON. It considered that, by using data from this study, the proportion of people continuing on subsequent treatments may have been overestimated. The clinical experts explained that, after third line, multiple myeloma is likely to be refractory to lenalidomide and people are more likely to have pomalidomide plus dexamethasone. They considered that there would be no significant differences in subsequent treatments after third line based on whether people had selinexor combination or panobinostat combination. The committee was aware that subsequent treatment assumptions were a key driver of costs and cost effectiveness in the model, particularly when assuming no overall survival benefits between treatments (see section 3.12). The committee acknowledged that the EAG's distributions of subsequent treatments excluded treatments that were not relevant to NHS clinical practice. But it considered that these distributions, particularly after third line, did not reflect the opinion of the clinical experts.

For the second committee meeting, the EAG provided 2 scenarios for modelling subsequent treatment costs. Scenario 1 was based on feedback from the clinical experts at the first committee meeting. It assumed that:

• 20% of people at third line would have subsequent treatments, and that there were no differences in subsequent treatments regardless of the third-line treatment

• 80% of people at fourth line would have pomalidomide plus dexamethasone and 20% would have chemotherapy

• 20% of people at fifth line would have pomalidomide plus dexamethasone and 80% would have chemotherapy.
  
  For scenario 2, the EAG estimated a one-off cost for subsequent treatments for selinexor combination and its comparators. This approach balanced the costs of subsequent treatments across all treatment arms so that the incremental cost was zero.
  
  At the second committee meeting, the clinical experts clarified that they considered 20% of people continuing to have subsequent treatment at third line was too low. They considered 50% more plausible. They also explained that the composition of subsequent treatments at fourth line did not reflect clinical practice. The EAG explained that, because selinexor combination has a shorter progression-free survival than its comparators at third line, people having selinexor combination would accrue greater subsequent treatment costs. The committee recalled that it accepted the EAG's assumption that there was no overall survival benefit between treatment arms (see section 3.12). It was concerned that this assumption, combined with the shorter duration of progression-free survival for selinexor combination, was creating a difference in subsequent treatment costs. This was not supported by any evidence, and may not exist in clinical practice. The committee considered that scenario 2, which assumed that there were no differences in subsequent treatment costs between arms, would be a more consistent and appropriate assumption.

In the company's base case, grade 3+ treatment-emergent adverse events that occurred in at least 5% of people in BOSTON were modelled as weekly rates for the duration on therapy. The impacts of the adverse events on quality of life were also modelled as a weekly disutility. The company assumed that adverse events are managed in primary and secondary care. In its base case, the EAG modelled adverse events as a one-off event in cycle 1 in line with previous multiple myeloma technology appraisals. Also, it assumed that adverse events are managed in secondary care. The clinical experts explained that adverse events are generally managed in secondary care. If side effects are not manageable, treatment is stopped. They considered that adverse events are likely to become less frequent over time with improved management and dose reduction, if appropriate. The committee considered that modelling a one-off event does not consider the distribution of adverse events over the duration on treatment. It also thought that it does not capture the long-term impact on quality of life. It acknowledged that the company's approach, using cumulative adverse event rates divided to provide weekly event rates ,assumed a constant incidence of adverse events over time. It did not think this was appropriate and led to benefits for treatments with shorter estimated progression-free survival such as selinexor. The committee concluded that the approach of modelling adverse events as a one-off event in cycle 1 was the best option.

In its base case, the company used EQ‑5D‑5L data from BOSTON for health-state utilities mapped to the EQ‑5D‑3L using the algorithm published in Hernandez Alava et al. (2020). It applied pooled utilities from trial arms and assumed that health-related quality of life did not depend on treatment, lines of treatment or differences in treatment-emergent adverse event profiles. The company also provided a scenario analysis using utility values from Hatswell et al. (2019), a source used in other multiple myeloma technology appraisals. In its base case, the EAG used EQ‑5D‑5L data from BOSTON and line of treatment as a covariate. The clinical experts noted the small difference in utilities between the progression-free survival and progressed health states in BOSTON, whereas a larger difference was shown using the data from Hatswell et al. (2019). The company explained that, often in trials, utilities for the progressed health state are based on an assessment at 1 timepoint shortly after disease progression. The clinical experts explained that disease-related and non-disease-related comorbidities increase over time. So, it is clinically plausible that over time and with later lines of treatment, the impact on health-related quality of life will be greater. The committee acknowledged the likely overestimation of the utility for the progressed health state but considered that it preferred to use the trial utility value data by line of treatment. It considered that the EAG's base-case approach for health-state utilities was more appropriate for decision making.

In response to the draft guidance consultation, the company accepted all the assumptions in the EAG's base case except for the assumption of no overall survival benefit. It still preferred to use indirect treatment comparison results for overall survival in its model (see section 3.12). The committee's preferred assumptions were largely in line with that of the EAG's base case, which were:

• using the results of the network meta-analyses for the second-line comparison with carfilzomib plus dexamethasone, and the third-line comparison with panobinostat combination (see section 3.7)

• using the results of the unanchored MAIC for the third-line comparison with ixazomib combination (see section 3.7)

• using independently fitted models for extrapolations of progression-free survival, overall survival and time on treatment (see section 3.10)

• using baseline curves of bortezomib plus dexamethasone for comparator extrapolations (see section 3.11)

• assuming no difference in overall survival between treatments and using bortezomib plus dexamethasone as baseline for overall survival for all treatments (see section 3.12)

• modelling of adverse events and associated disutility as a one-off event in cycle 1 (see section 3.14)

• modelling health-state utilities by line of treatment (see section 3.15).
  
  Also, for subsequent treatment costs, the committee preferred to instead assume that there was no difference in costs between arms (see section 3.13). The committee considered that the EAG's base case with scenario 2 for subsequent treatment costs reflected its preferred assumptions.

NICE's health technology evaluations manual notes that judgements about the acceptability of a technology as an effective use of NHS resources will take into account the degree of certainty around the incremental cost-effectiveness ratio (ICER). The committee will be more cautious about recommending a technology if it is less certain about the ICERs presented. NICE's health technology evaluations manual also states that decisions about the acceptability of the technology will consider aspects that relate to uncaptured benefits and non-health factors. The committee recalled the statements from the clinical and patient experts on the significant unmet need for effective and safe treatments at later lines in the treatment pathway. It also noted that selinexor has a novel mechanism of action and, as an oral treatment, would be easily administered and fit into the existing care pathway. The committee acknowledged the high unmet need for novel treatments, especially at later lines in the treatment pathway. But it also noted the high levels of uncertainty, including:

• the representativeness of the population from BOSTON and the generalisability of the results to people with multiple myeloma likely to have selinexor combination in NHS clinical practice (see section 3.6)

• the wide credible intervals for outcomes in the indirect treatment comparisons (see section 3.8)

• the uncertainty in longer-term extrapolations of progression-free survival, overall survival and time on treatment (see sections 3.10 to 3.12).
  
  The committee concluded that an acceptable ICER would be below £20,000 per quality-adjusted life year (QALY) gained.

The recommendations apply equally to all people with relapsed or refractory multiple myeloma. The clinical experts noted that multiple myeloma is common in men, elderly people, and people from African or Caribbean ethnic groups. The committee considered that its recommendations apply equally, regardless of sex, age and ethnicity. It concluded that the difference in prevalence did not represent an equality issue in this evaluation.

The clinical experts considered that selinexor combination provides an alternative treatment option with a novel mechanism of action. The company highlighted some uncaptured benefits, including the impact on carer health-related quality of life. The committee considered that there may be benefits uncaptured in the economic modelling. For example, the value of an additional oral treatment option, particularly at second line, for which the ease of administration and adverse effect profile may make it more acceptable than existing options. The committee also acknowledged that there would be an increasing number of people whose condition is refractory to lenalidomide and daratumumab. It concluded that selinexor combination provides an alternative treatment option with a novel mechanism of action.

NICE's advice about conditions with a high degree of severity did not apply.