To the Editor,

Therapeutic options have significantly advanced for patients with multiple myeloma (MM) including combination therapies employing complementary mechanisms or targeting mechanisms distinct from previous regimens [1, 2]. Selinexor is a first-in-class, orally-available, selective inhibitor of nuclear export (SINE) compound that has shown definitive activity with low dose dexamethasone in patients with triple class refractory MM in the STORM study [3] and synergistic activity with bortezomib and dexamethasone (XVd) in patients with 1–3 prior therapies in the BOSTON study [4]. Here we analyzed pre-specified subpopulations from the BOSTON study to determine the impact of prior lines of therapy and identify those who might optimally benefit from the XVd regimen.

Baseline characteristics were well balanced between treatment arms across subgroups (Additional file 1: Table S1). Median progression-free survival (PFS) was longer on XVd versus Vd in patients with 1 prior line (P = 0.0148) or 2–3 prior lines (P = 0.0295), lenalidomide-naïve (P = 0.0150) or lenalidomide-treated (P = 0.0177) patients, and PI-naïve patients (P = 0.0003), with a strong trend in PI-treated patients. Patients with IMiD-refractory MM had a significantly longer median PFS (P = 0.0051), as did patients with or without prior ASCT (P = 0.0074 and P = 0.0341). A post-hoc analysis showed a trend towards longer PFS with XVd in patients who received limited bortezomib induction prior to ASCT treatment (Table 1).

Table 1 Progression-free survival by subgroup

Treatment with XVd was associated with a significantly higher overall response rate including patients with 1 prior line, 2–3 prior lines, lenalidomide-naïve or treated, PI-naïve or treated, and prior ASCT (Fig. 1). Subgroups with 1 prior therapy, lenalidomide-naïve, and prior PI treatment had significantly higher rates of  ≥ VGPR (Additional file 1: Table S2). Median time-to-next-treatment was significantly improved with XVd versus Vd: 1 prior line, 2–3 prior lines, lenalidomide-naïve or treated, PI-naïve or treated, and prior ASCT. Across the entire study, overall survival (OS) trended in favor of XVd over Vd (HR, 0.84 [95% CI 0.57–1.23]; P = 0.19). The median OS for lenalidomide-naïve and PI-naïve patients was not reached, but favored XVd over Vd (HR, 0.76 [95% CI 0.45–1.29] P = 0.16 and HR 0.63 [95% CI 0.25–1.61], P = 0.16, respectively) (Additional file 1: Table S3).

Fig. 1
figure 1

Depth of response by subgroup and treatment arm. The distribution of response pattern in subgroups based on number of prior lines, lenalidomide (LEN) or proteasome inhibitor (PI) treatment, IMiD refractoriness, and autologous stem cell transplant (ASCT). Bort bortezomib, CR complete response, IMiD immunomodulatory drug, NS not significant, PR partial response, sCR stringent complete response, VGPR very good partial response. Odds ratio and P value shown. *P < 0.05; **P < 0.01, ***P < 0.001

Overall grade ≥ 3 adverse events (AEs) occurred more frequently with XVd and were generally well managed. Importantly, grade ≥ 2 peripheral neuropathy occurred significantly less frequently across all XVd subgroups. The incidence of serious AEs and drug discontinuation due to AEs trended higher with XVd (Additional file 1: Table S4). There was no clear trend regarding AEs leading to a fatal outcome, although the slight excess number of deaths with XVd in the PI-treated subgroup were restricted to India prior to the institution of increased monitoring, after which there were no additional deaths.

Our observations are particularly noteworthy as the once weekly XVd regimen utilizes ~ 40% less bortezomib and 25% less dexamethasone and requires ~ 37% fewer clinic visits for bortezomib injections than the standard Vd regimen. Despite the number of additional, subsequent therapies available to patients in this study, allowing patients on Vd with objective progressive disease to cross-over to a selinexor regimen, and the relatively short follow up, the results were accompanied by favorable trends on OS. Given its unique role in reactivating multiple tumor suppressor proteins and demonstrated synergy with PIs as well as other anti-MM drugs [5,6,7,8,9], these findings are consistent with the use of oral selinexor earlier in the MM treatment course. It is possible that some of the benefits of selinexor in those PI-treated patients may reflect the documented synergy between selinexor and PIs, even cells with marked PI refractoriness [5]. Moreover, benefits in duration and depth of response of XVd over Vd were most pronounced in patients who were PI-naïve, suggesting that selinexor could be an optimal partner for combining with weekly bortezomib as the first PI-containing MM regimen. Moreover, as daratumumab + lenalidomide/dexamethasone (DRd) is increasingly utilized in frontline MM treatment, the once weekly XVd regimen in second line could lead to a marked reduction in the development of prolonged or permanent bortezomib-associated neuropathy [10, 11]. Furthermore, the use of XVd following DRd allows for optimal mechanistic switching, thus preserving second generation agents (PIs, IMiDs and anti-CD38 mAbs) for subsequent lines of therapy where they may be more effective [1, 2, 12].

In conclusion, the earlier use of selinexor in treating MM may provide better, more durable outcomes with lower rates of peripheral neuropathy, using one of the simplest triplet regimens currently available for the treatment of patients with MM [4].