Overall survival with oral selinexor plus low‐dose dexamethasone versus real‐world therapy in triple‐class‐refractory multiple myeloma

Abstract Triple‐class‐refractory multiple myeloma (MM) describes MM refractory to proteasome inhibitors, immunomodulatory agents, and anti‐CD38 monoclonal antibodies. In the Phase IIb STORM study (NCT02336815), oral selinexor plus low‐dose dexamethasone (Sel‐dex) demonstrated a 26.2% overall response rate in triple‐class‐refractory MM. Here, we compare overall survival (OS) of 122 patients with triple‐class‐refractory MM who received Sel‐dex in STORM Part 2 with that of 64 similar patients treated with other available therapies in a Flatiron Health Analytic Database (FHAD) cohort. OS from the date that the patients’ MM became triple‐class‐refractory was longer in STORM versus FHAD, with an unadjusted hazard ratio (HR) of 0.43 (P = .0002; adjusted HR 0.35 [P = .011]). In a subset analysis of highly resistant patients receiving further therapies after their MM first became at least triple‐class‐refractory (i.e., who received Sel‐dex in STORM, n = 64, and non‐Sel‐dex in FHAD, n = 36), the OS was significantly longer in STORM with an unadjusted HR of 0.52 (P = .0331; adjusted HR 0.33 [P = .041]). Within the limits of this analysis, the OS of patients with at least triple‐class‐refractory MM was significantly better with Sel‐dex versus available therapies, suggesting that Sel‐dex may be associated with a meaningful OS benefit in these patients.

n = 36), the OS was significantly longer in STORM with an unadjusted HR of 0.52 (P = .0331; adjusted HR 0.33 [P = .041]). Within the limits of this analysis, the OS of patients with at least triple-class-refractory MM was significantly better with Sel-dex versus available therapies, suggesting that Sel-dex may be associated with a meaningful OS benefit in these patients.

K E Y W O R D S
multiple myeloma, selective inhibitor of nuclear export, selinexor, triple-class-refractory multiple myeloma INTRODUCTION Multiple myeloma (MM) is the second most common hematologic malignancy after lymphoma, accounting for an estimated 32 110 new cases and 12 960 deaths in the United States alone in 2019 [1]. Current backbone treatment options include proteasome inhibitors (PIs; such as bortezomib, carfilzomib, and ixazomib), immunomodulatory agents (IMiDs; including thalidomide, lenalidomide, or pomalidomide), and anti-CD38 monoclonal antibodies (mAbs; e.g., daratumumab), as well as histone deacetylase inhibitors (e.g., panobinostat), and most of which have been shown to extend survival in patients with MM [2][3][4][5][6][7]. However, the increasing use of multiple lines of PI/IMiD/mAb combination therapies has led to a growing number of patients with penta-treated MM (defined as prior therapy with bortezomib, carfilzomib, lenalidomide, pomalidomide, and daratumumab). The majority of patients with penta-exposed MM have triple-classrefractory MM, defined as MM that is refractory to at least one PI, one IMiD, and one anti-CD38 mAb. These patients have a poor prognosis with a median overall survival (OS) of ≤6 months [8,9]. Therefore, novel therapies are required to address this growing unmet need.
Selinexor is a novel, oral selective inhibitor of nuclear export compound that blocks exportin 1 (also known as chromosome maintenance region 1), a nuclear export protein overexpressed in cancers including MM. Selinexor plus dexamethasone (Sel-dex) was approved by the U.S. Food and Drug Administration (FDA) in July, 2019 for treating patients with MM refractory to at least two PIs, at least two IMiDs, and an anti-CD38 mAb (penta-refractory), based on data from the Phase IIb STORM study [10]. The pivotal Part 2 of STORM enrolled 122 patients with penta-exposed triple-class-refractory MM who achieved an overall response rate of 26.2% [11]. Here, we compare the OS of patients who received Sel-dex in STORM Part 2 with a similar cohort of real-world patients with penta-exposed tripleclass-refractory MM from the observational Flatiron Health Analytic Database (FHAD). An additional exploratory analysis was performed on a subset of patients who received at least one therapy after becoming penta-refractory and triple-class-refractory to assess the impact on OS.

Patients and study design
This study included 122 patients from the modified-intent-to-treat population for STORM Part 2 and 64 similar patients from an observational cohort derived retrospectively from electronic health records aggregated in FHAD, a nationally representative oncology platform. The main objective was to characterize the OS of patients with penta-exposed triple-class-refractory MM treated in the realworld who were not treated with selinexor (FHAD cohort) with the OS of similar patients treated with Sel-dex in STORM Part 2 (STORM cohort).

STORM data selection criteria
The study design and enrollment criteria for the STORM study have been previously published [12].

Analysis populations and index date
The overall analysis included the entire study population of 122 patients in STORM, and 64 patients from FHAD. An index date was used to classify the observation time for each patient into a baseline and a follow-up period. For STORM, the index date was defined as the progression date of the last regimen prior to Sel-dex initiation. As all patients enrolled into STORM were had MM that was both penta-exposed and triple-class-refractory, and their index date could not be earlier than the date on which the patient's MM became triple-class-refractory. For FHAD, the index date was defined as the end date of the regimen for which the patient's MM first became pentaexposed. Thus, for FHAD patients the index date could have been earlier than the date on which the patient's MM became triple-classrefractory.
In order to further align these study populations by adjusting for possible survival bias, we performed an additional analysis of OS data utilizing an index date defined as first day of first therapy given after MM became penta-exposed triple-class-refractory. From the STORM cohort, this included 64 patients who received Sel-dex as their first therapy after developing penta-exposed triple-class-refractory MM.
From the FHAD cohort, this included 36 patients who received ≥1 therapy after their MM first became penta-exposed and triple-classrefractory.

Statistical methods
OS was defined as the duration from the corresponding index date until death from any cause, and an unadjusted comparison of OS between the two cohorts was initially performed. The proportion of patients with death and the two-sided, Wald 95% confidence interval (CI) were calculated for the FHAD and STORM cohorts. Median OS with 95% CI and the proportion surviving for 6 and 12 months were estimated based on the Kaplan-Meier method. A Cox proportional hazards regression model with OS as the outcome and study population as the only independent variable was conducted to compare the OS between the two cohorts.
To adjust for potential imbalance between the two cohorts in terms of baseline variable distribution, a regression analysis was conducted.
Specifically, a survival Cox prediction model was developed using the data from FHAD cohort with a set of potentially clinically relevant baseline covariates (see Table 1).  for HR and its nominal P value were obtained via the bootstrapping method applied to the FHAD and STORM populations.

Patient population
Overall, 122 patients with penta-exposed triple-class-refractory MM were included in the STORM cohort. Of these, 64 patients received Seldex as the first therapy after their MM became penta-exposed triple-class-refractory and were included in the STORM cohort subset analysis. Of the 58 patients excluded from this subset analysis, 33 received one, 11 received two, and 14 received three or more additional lines of therapy before starting Sel-dex. The median time for these 58 patients from first developing penta-exposed triple-class-refractory MM and starting Sel-dex was 5.3 months (range 0.9-27.2 months).
Of 126 patients in FHAD, 64 had disease documented to be triple-class-refractory and Eastern Cooperative Oncology Group (ECOG) performance status ≤2; these were included in the FHAD cohort ( Figure 1). Of these, 36 had received one or more available therapies (including unapproved agents) after their MM first became penta-exposed triple-class-refractory and were included in the subset analysis. From the index date, in the first regimen, 19 of the 36 patients were treated with a PI-containing regimen, 17 with an IMiD-containing regimen, and nine with an anti-CD38 mAb-containing regimen, reflecting the use of retreatment strategies in this patient population.
Baseline characteristics and patient demographics were mostly similar between the two cohorts for both the overall and subset analyses (Table 1 and Table S1). However, patients in the STORM cohort tended to be more heavily pretreated, more likely to have had prior treatment with an alkylating agent, an anthracycline, and/or stem cell trans-

Unadjusted analysis of OS
In the overall analysis, the median unadjusted OS was 10.  Table   S2). The unadjusted 6-month and 12-month OS probabilities for the STORM versus FHAD cohort are presented in Figure S1.

Model parameters for OS adjusted for prognostic factors in the FHAD (steps 1 and 2)
Number of prior regimens squared †  Table 2 and Figure 3. The predicted OS for patients treated with "other therapy" was significantly worse than observed OS in STORM patients treated with Sel-dex (HR = 0.35; 95% CI 0.14-0.85; P = .011).
The same method was implemented for the subset analysis with the same criteria and baseline variables considered. The initial number of predictors was reduced in this analysis due to the smaller subset sample size. The selected variables to be adjusted in this  (Table S2 and Figure S2).

OS by carfilzomib, pomalidomide, and daratumumab-refractory status
In an exploratory analysis of the effect of carfilzomib + pomalidomide + daratumumab-resistant status on OS, patients from the subset analysis were further analyzed by their carfilzomib + pomalidomide + daratumumab-resistant status. The majority of STORM patients (n = 60/64) had documented carfilzomib + pomalidomide + daratumumab-resistant disease and thus were not considered further in this analysis.

Postindex treatment in STORM and FHAD
Where possible, patients in both cohorts continued to be followed after discontinuation of their first postindex therapy. Of the 64 patients in the STORM cohort subset, 34 went on to receive additional therapies after Sel-dex: 17 patients received combination therapy; five received a venetoclax-containing regimen; four were re-challenged with a PI-, IMiD-, and anti-CD38-containing regimen; two received CAR T-cell therapy; two received immunotherapy; and one patient received radiation therapy.

DISCUSSION
Based on the STORM study, Sel-dex is the only therapy currently approved by the U.S. FDA for use in patients with MM whose disease is refractory to daratumumab and other approved agents, including PIs and IMiDs. The current analyses evaluated the OS of patients in STORM and comparable patients with penta-exposed triple-classrefractory MM. These analyses demonstrate that treatment with Sel-dex is associated with a doubling of the median OS of patients receiving their first therapy after their disease becomes penta-exposed and triple-class-refractory, compared with patients from a large contemporary cohort receiving available (mostly approved, and some not yet approved) therapies, but not including Sel-dex.
Importantly, the OS results observed in the FHAD cohort of this study were consistent with median OS values reported by other recent studies for patients with MM refractory to three or four drugs using real-world data from the United States, Europe, and Israel [8,9,13]. While the median number of prior systemic regimens was higher in the STORM versus the FHAD cohort, this difference may be attributable to differences in treatment preference between academic and community-based settings. Patients treated in the community setting, as in FHAD, may be more likely to receive fewer combination therapies (including salvage regimens containing alkylating agents) and less likely to receive experimental therapies than patients in an academic setting, as in STORM. In the current analysis, approximately one-half of patients with penta-exposed triple-classrefractory MM in the FHAD cohort went on to receive at least one further treatment regimen after the index date. Retreatment/recycling with PI and IMiD therapies was chosen most frequently for subsequent treatment, either as monotherapy or as part of a combination regimen. However, only ∼50% of patients in the FHAD cohort had developed carfilzomib and pomalidomide-refractory MM prior to developing penta-exposed triple-class-refractory MM. Therefore, while carfilzomib-and/or pomalidomide-based therapy could reasonably be expected to induce responses for some patients in the FHAD cohort, more than 90% of patients in the STORM cohort had MM refractory to these agents, along with daratumumab. In this context, the key finding in this analysis was that the median OS with Sel-dex in the STORM cohort was nearly double that compared to retreatment approaches used in the FHAD cohort. Adverse events with Sel-dex are generally manageable and reversible (as expected by the ∼7-hour halflife of selinexor), allowing patients to also go onto subsequent therapies with a lower burden of disease (and potentially different MM clones and resistance patterns) after being treated with Sel-dex. This supports the use of novel therapies that do not have cross-resistance with existing mechanisms of action and, in the case of Sel-dex, provide an oral, non-cross resistant regimen. Within the limitations of these analyses, the OS of patients receiving Sel-dex as first therapy for penta-exposed triple-class-refractory MM in the STORM cohort was significantly longer than those receiving available therapy in the FHAD cohort. These results were also found across a number of subgroups. Sel-dex may be favorable for the treatment of relapsed and/or refractory MM when patients first become triple-class-refractory MM, as opposed to current retreatment strategies, with either Sel-dex as in STORM, or in combination with currently available agents [15][16][17][18]. These results highlight the ability of Sel-dex to address the critical unmet medical need for patients with tripleclass-refractory MM.

ACKNOWLEDGMENTS
We thank all the patients who participated in the study and their fami-