The Effect of Duration of Lenalidomide Maintenance and Outcomes of Different Salvage Regimens in Patients with Multiple Myeloma (MM)

The optimal duration of lenalidomide maintenance post-autologous stem cell transplant (ASCT) in Multiple Myeloma (MM), and choice of therapy at relapse post-maintenance, need further evaluation. This retrospective study assessed outcomes of patients with MM (n = 213) seen at Mayo Clinic, Rochester between 1/1/2005–12/31/2016 who received lenalidomide maintenance post-ASCT. The median PFS was 4 (95% CI: 3.4, 4.5) years from diagnosis of MM; median OS was not reached (5-year OS: 77%). Excluding patients who stopped lenalidomide maintenance within 3 years due to progression on maintenance, ≥3 years of maintenance had a superior 5-year OS of 100% vs. 85% in <3 years (p = 0.011). Median PFS was 7.2 (95% CI: 6, 8.5) years in ≥3 years vs. 4.4 (95% CI: 4.3, 4.5) years in <3 years (p < 0.0001). Lenalidomide refractoriness at first relapse was associated with inferior PFS2 [8.1 (95% CI: 6.4, 9.9) months vs. 19.9 (95% CI: 9.7, 30.2; p = 0.002) months in nonrefractory patients]. At first relapse post-maintenance, median PFS2 was superior with daratumumab-based regimens [18.4 (95% CI: 10.9, 25.9) months] versus regimens without daratumumab [8.9 (95% CI: 5.5, 12.3) months; p = 0.006]. Daratumumab + immunomodulatory drugs had superior median PFS2 compared to daratumumab + bortezomib [NR vs 1 yr (95% CI: 0.5, 1.5); p = 0.004].


INTRODUCTION
The current most widely agreed upon frontline treatment strategy for multiple myeloma (MM) involves induction with triplet therapies followed by high-dose chemotherapy and autologous stem-cell transplantation (ASCT) for eligible patients, and then maintenance therapy with or without preceding consolidation. Lenalidomide is the preferred maintenance strategy following ASCT, especially in non-high-risk patients with MM, and multiple randomized studies (e.g., CALGB100104, IFM2005-02, GEMIMA, Myeloma IX, German multicenter study) have demonstrated an improvement in progression-free survival (PFS) with lenalidomide maintenance [1][2][3][4][5]. Only the CALGB100104 study showed a significant overall survival (OS) improvement with lenalidomide maintenance while an OS benefit was not demonstrated in the other studies; possibly because they were not powered for OS as the primary endpoint [3,4,6]. A meta-analysis of CALGB, GEMIMA, and IFM studies confirmed a statistically significant improvement in OS in patients receiving lenalidomide maintenance [7]. While there is no prospective data to guide the optimal duration of lenalidomide maintenance, the current practice consensus is to continue lenalidomide indefinitely until progression or unacceptable toxicity. This is supported by retrospective data from two studies showing improvement in PFS (and OS in one) with a longer duration of lenalidomide maintenance up to 32 months [8,9]. Importantly, while maintenance therapy has improved outcomes for patients with MM, progression on maintenance therapy is common and there are limited data to guide the selection of optimal therapy at relapse. In this study, we follow-up on patients treated with lenalidomide-based maintenance therapy to assess the impact of duration of maintenance and report the outcomes with various treatment regimens used at first relapse postmaintenance.

PATIENTS AND METHODS
After Mayo Clinic institutional review board approval, 213 patients with MM diagnosed consecutively between 1/1/2005 and 12/31/2016 receiving early transplant (i.e., frontline ASCT within 1 year of diagnosis) and treated with lenalidomide (with or without dexamethasone) as maintenance therapy were included in this retrospective study. The timing of starting maintenance therapy following ASCT and starting dose of lenalidomide was based on routine current practices. Typically, patients were recommended to start lenalidomide 2-4 months after ASCT at an initial daily dose of 10-15 mg. The mSMART 3.0 criteria were used for risk stratification based on cytogenetics features on interphase FISH [5,10]. Median follow-up and median duration of maintenance therapy were calculated using the reverse Kaplan-Meier estimator method. Response to therapy was determined using the 2016 IMWG criteria [11]. For deepening of response with maintenance, the best-recorded response after initiation of maintenance therapy was used for comparison with the pre-maintenance therapy response. Choice of treatment regimens at first relapse after maintenance therapy was based on individual provider preference and prevalent data for salvage therapies. Second progressionfree survival (PFS2) was calculated from the start of post-maintenance therapy until discontinuation of therapy due to progression, therapylimiting toxicity, or death. The Kaplan-Meier method was used to estimate the median OS, PFS, and PFS2. All statistical analyses were performed using SPSS (version 19; SPSS Inc., Chicago, IL, USA).
M. Ho et al.
Patients who achieved or maintained a best response of ≥VGPR within 2 years of maintenance were noted to have a significantly better OS from diagnosis (5-year OS: 82%) compared with patients with ≤PR (67%; p = 0.003) ( Table 3). The median PFS from diagnosis was 4.4 (95% CI: 3.9, 4.9) years in patients with ≥ VGPR as a best response within 2 years of maintenance versus 3.3 (95% CI: 2.2, 4.3) years in patients with ≤ PR (p = 0. 003) ( Table 3). Adjusting for age, ISS stage 3, cytogenetic risk group, and patients who received Rd maintenance, the hazard ratio (HR) for OS and PFS in patients who achieved or maintained a best response of ≥ VGPR within 2 years of maintenance was 0.32 (95% CI: 0.18, 0.58; p < 0.0001) and 0.56 (95% CI: 0.38, 0.83; p = 0.004), respectively.
There was no significant difference in OS and PFS between patients who achieved ≥ VGPR prior to the start of maintenance (n = 128) compared with patients who first achieved ≥ VGPR (i.e., had deepening of response to ≥VGPR) within the first 2 years of lenalidomide maintenance (n = 35). The 5-year OS from diagnosis was 77% in patients who achieved ≥VGPR prior to maintenance compared with 80% in patients who first achieved ≥ VGPR within the first 2 years of maintenance (p = 0.41). Median PFS from diagnosis was 4.3 (95% CI: 3.5, 5.2) years in patients who achieved ≥ VGPR prior to maintenance versus 5 (95% CI: 3.6, 6.5) years (p = 0.17) in patients who first achieved ≥ VGPR within the first 2 years of maintenance ( Table 3).
Outcomes with salvage regimens at first relapse postmaintenance One-hundred-thirty-six (64%) patients relapsed during or after stopping maintenance; 126 (59%) patients were documented to have received therapy at relapse (10 patients were either lost to follow-up or opted for no treatment).

DISCUSSION
Lenalidomide maintenance is an effective and well-tolerated strategy to prolong PFS even in the era of novel agents alongside ASCT, and several trials have demonstrated a remarkable risk reduction in progression and death especially in the post-ASCT setting [1,3,4,6]. However, several important questions on the ideal duration of maintenance, as well as the optimal choice of therapy at first relapse remain unanswered.
The PFS and OS data in our uniform cohort was largely comparable to the previously reported outcomes in phase 3 trials studying lenalidomide maintenance. The median PFS in our study was 3 years (i.e., 36 months) from the start of maintenance and 4 years (i.e., 48 months) from diagnosis, which is comparable to the median PFS reported in the Myeloma IX (39 months from randomization), IFM (41 months from randomization), CALGB (46 months from randomization), and GEMIMA (42 months from randomization) studies [1,3,4,6]. The 3 year OS from the start of maintenance was 84% (89% from diagnosis) which is comparable to the 3 year OS reported in Myeloma IX (79%), IFM (80%), CALGB (88%), and GEMIMA (88%) [1,3,4,6]. In our cohort, achievement or maintenance of a deeper response (≥VGPR) within 2 years of maintenance therapy was associated with an improvement in PFS and OS. This is consistent with multiple prior studies demonstrating improvement in outcomes with deeper responses [12][13][14].
Our study also suggests that a longer duration of lenalidomide maintenance (i.e., ≥3 years) was associated with an improved OS and PFS, and this association was significant even after adjusting for high-risk cytogenetics, age, and ISS stage 3. We controlled for guarantee-time bias by excluding patients that had stopped lenalidomide maintenance within 3 years due to disease progression. Consistent with our results, a pooled analysis of the GIMEMA MM-03-05, RV-MM-PI-209, and CC-5013-MM-015 trials showed that continuous therapy (defined as upfront therapy followed by maintenance lasting ≥ 2 years) was superior to fixed duration therapy (defined as upfront treatment for ≤1 year) [8].
Patients who started maintenance therapy on or after 1/1/2014 (n = 124) had significantly better OS compared with patients who started maintenance on or before 12/31/2013 (n = 89; p < 0.0001), which is most probably explained by the availability and use of novel therapies. We next compared the efficacy of various groups of therapies in the post-maintenance relapse setting and found that, consistent with the above, the use of daratumumab (first FDAapproved in 2015) at initial relapse was associated with improved PFS2. Daratumumab in combination with an IMiD was significantly superior compared to daratumumab with bortezomib, even after adjusting for age, high-risk cytogenetics, ISS stage 3, and lenalidomide refractoriness at time of salvage. One possible explanation for the synergism between daratumumab and IMiDs is the latter's ability to enhance NK cell proliferation, cytotoxic activity, and therefore daratumumab-mediated antibody-dependent cellular cytotoxicity (ADCC); even in the setting of IMiD-refractory MM cells [15,16].  5 Daratumumab + IMiD vs daratumumab + PI at first relapse post-maintenance. Daratumumab + IMiD compared with daratumumab + PI at first relapse after maintenance (a) The median PFS2 in patients who received daratumumab + IMiD (n = 16) was not reached compared with 1 (95% CI: 0.5, 1.5) year in patients who received daratumumab + PI (n = 15) (p = 0.004). There were 84% of patients event-free at 5 years in the dara + IMiD group and 0% in the dara + PI group. b The median PFS2 in patients who received dara + Rd or dara + Pd were not reached. 5-year event-free survival was 69% in the dara + Rd group versus 100% in dara + Pd (p = 0.138). Dara + Pd was associated with a significantly better 5-year EFS (100%) compared with dara + Vd (0%; p = 0.006).