Azacitidine and Lenalidomide in Higher-Risk Myelodysplastic Syndromes. Long-Term Results of a Randomized Phase II Multicenter Study and Impact of Cytogenetic Scores and Mutational Status on Long-Lasting Responses

Introduction. The association of Azacitidine (AZA) and Lenalidomide (LEN), either administered concurrently or sequentially, has proven effective in Myelodysplastic Syndromes (MDS), however the optimum dose and schedule remains unknown. The aim of this study was to evaluate the efficacy and safety of the combination vs the sequential use of AZA and LEN in higher-risk MDS pts. Primary endpoint: ORR, defined as the Rate of Complete Remission (CR), Partial Remission (PR), Marrow Complete Remission (mCR), and Hematological Improvement (HI), following the IWG criteria (Cheson, 2006). Moreover, the aim of this analysisis is to enucleate the clinical and biological features of pts who showed long-lasting (≥ 20 cycles) responses. Methods. This is a randomized, phase II, multicenter, open label study, including pts with MDS with IPSS risk High or Intermediate-2, without previous treatment with AZA or LEN. ARM 1 (combined treatment): AZA: 75 mg/m2/day (days 1-5) I.C. + LEN: 10 mg/day (days 1-21), orally, every 4 weeks. ARM 2 (sequential treatment): AZA: 75 mg/m2/day (days 1-5) I.C. + LEN: 10 mg/day (days 6-21), orally, every 4 weeks. The induction treatment was planned for 8 cycles. For responder pts the same treatment was continued until disease progression or unacceptable toxicity. Results. From March 2013, 44 pts (27 males), median age: 72 (48-83 yrs) were enrolled, from 13 hematologic Centers. 21 pts were randomly assigned to ARM 1, and 23 pts to ARM 2. Treatment was given for a median of 8.5 (1-68) cycles; in ARM 1: 9 (1-68) cycles; in ARM 2: 8 (1-63) cycles, respectively. Median follow-up: 15 (2-77) months. 10/44 pts (22.7%) did not complete at least 6 cycles of treatment for causes other than disease progression, and were not considered evaluable for response. Among the 34/44 pts (77.3%) evaluable for response, 26/34 pts (ORR: 76.5 %) showed a favourable response to treatment. Intention-to-treat (ITT) analysis: ORR: 59.1%. First response was observed after a median of 2 (1-8) cycles. The Best Response achieved was: CR: 8 pts (23.5%) (ITT: 18.1%), PR: 1 pt (2.9%) (ITT: 2.2%), mCR: 3 pts (8.8%) (ITT: 6.8%), HI: 8 pts (23.5%) (ITT: 18.1%), mCR+HI: 6 pts (17.6%) (ITT: 13.6%). Median duration of hematologic response: 10.5 months. 37 pts (84.1%) died , and 20 pts (45.4%) showed progression to AML. Grade >2 non haematological toxicity: 54.5%. Median OS: 15 months. OS was significantly longer in responder pts as compared to the other pts (28 vs 7 months, p<0.0001). No significant differences between the 2 arms were observed, in terms of ORR (ARM 1: 76.5%, ITT: 61.9%; ARM 2: 76.5%, ITT: 56.5%), CR rate (ARM 1: 17.6%, ITT: 14.3%; ARM 2: 29.4%, ITT: 21.7%), grade >2 non haematological toxicity (ARM 1: 66.7%; ARM 2: 43.5%), AML incidence (ARM 1: 33.3%; ARM 2: 56.5%; p=0.2150) and OS (ARM 1: 14 months; ARM 2: 16 months). However, among responder pts, sequential treatment showed a longer clinical benefit, as compared to combined treatment. Responder pts of ARM 2 showed a significantly longer median duration of response (18 vs 6 months, p=0.0481), a longer median duration of therapy (28 vs 10 months, p=0.0870; 20 vs 10 cycles, p=0.1181), more long-lasting (≥ 20 cycles) responses (34.8% vs 9.5%, p=0.1017) and a longer OS (35 vs 26 months, p=0.3868), as compared to responder pts of ARM 1. Overall, 10/44 long-responder pts (22.7%) received ≥ 20 cycles; 5/10 pts (50%) achieved CR. IPSS risk: Intermediate-2 (8 pts); High (2 pts); IPSS-R risk: Intermediate (2 pts); High (6 pts); Very High (2 pts); IPSS cytogenetic risk: Good (5 pts); Intermediate (3 pts); Poor (2 pts); IPSS-R cytogenetic risk: Good (5 pts); Intermediate (4 pts); Very Poor (1 pt); 4/6 patients with altered karyotype achieved cytogenetic remission; it is noteworthy that the only 3 pts of the entire series who showed no gene mutations at baseline are included in this subset of long-responders pts, while 5/10 pts showed at baseline ≥ 1 prognostically unfavorable gene mutations (none with TP53 mutations), with variable VAFs during treatment. Moreover all long-responder pts showed a common gene mutation on SOD2 gene, and mutations on PLCG2 gene. Conclusions. Our results confirm the efficacy of both AZA+LEN treatment regimens in higher-risk MDS pts, in terms of ORR and OS, although sequential treatment was associated with a longer clinical benefit among responder pts. A subset of pts (22,7 %) with less unfavorable cytogenetic and molecular characteristics showed a long-lasting response to treatment.