Abstract
Although anemia of LR MDS initially responds to erythroid stimulating agents (ESA) in 40-50% of patients, response is generally transient. AZA may lead to RBC transfusion independence (RBC-TI) in 30-40% of LR-MDS patients but it has not been studied prospectively in those resistant to ESA. While prognostic factors of response and survival to AZA have been largely analysed in high risk MDS, including by our group (Itzykson, Blood, 2011), they remain uncertain in LR-MDS, especially regarding new genetic factors. We previously reported (J Clin Oncol 30, 2012 # 6523) in a randomized phase II trial comparing AZA and AZA+EPO in 93 LR-MDS that were RBC-TD and resistant to ESA, a similar erythroid response rate in the 2 treatment arms (NCT01015352). Here, we report prognostic factors of response and overall survival in this trial, including genetic markers (SNP array-based karyotype and somatic mutations).
Patients included in the trial received AZA 75mg/m2/d for 5 days every 28 days for 6 cycles (AZA arm) or with added EPO 60000U/week (AZA+EPO arm). Erythroid response was evaluated after 4 and 6 cycles of AZA, according to the IWG 2000 criteria. Responders were then eligible for 12 AZA+/-EPO maintenance cycles. Affymetrix SNP array 6.0 (SNPa) analysis was performed at baseline on marrow mononuclear cells, and mutations of SF3B1, TET2, DNMT3A, ASXL1, JAK2 and 19 other genes were screened by next generation sequencing and validated by Sanger sequencing. The following factors were analysed for their prognostic value on erythroid response (IWG 2000) and survival: age, gender, IPSS, IPSS cytogenetics, WHO diagnosis, time since MDS diagnosis, SNPa karyotype, SF3B1, TET2, DNMT3A, ASXL1 and JAK2 mutations.
In the 93 patients, M/F was 65/28; median age 72 y (IQR: 65-78); WHO diagnosis: RA 5.3%, RARS 40.9%, RCMD 15.1%, RCMD-RS 17.2%, RAEB-1 12.9%, CMML 7.5%, MDS-U 1%. Median MDS duration prior to inclusion was 37.2 months. IPSS was low in 35 and int-1 in 57 patients (NA in 1). IPSS cytogenetics was fav in 65, int in 9, defav in 2 patients, and failed or not done at study entry in 17 (but cytogenetics at MDS diagnosis was available in 16 of them). An abnormal SNPa karyotype was present in 33/79 patients analysed, including 9 with normal conventional cytogenetics. SF3B1, TET2, DNMT3A, ASXL1 and JAK2 mutations were detected in 59/86, 29/87, 12/87, 5/89 and 3/87 patients. Of the 19 other genes analysed, 12 showed no mutation and 7 (UA2F1, CBL, EZH2, RUNX1, ETV6, KIT, IDH2) were mutated in ≤3 patients and were not further considered in this analysis. In ITT analysis (N=93), erythroid response was 34% after 4 courses (37.5 % vs. 31% in the AZA and AZA+EPO arms, respectively, p=0.82) and 30% after 6 courses (35% vs. 24%, p=1.00). In univariate analysis, none of the factors listed above, including genetic markers, significantly predicted treatment response after 4 or 6 courses, while a trend was observed for a higher response rate after 6 cycles for SF3B1 mutated versus wt. patients (35.6 vs. 14.8 % respectively, P=0.07). Median follow-up of the 93 patients was 30 months (IQR: 23-34). Five patients developed AML (1 in AZA arm and 4 in AZA+EPO arm, P=0.19), all 5 in non-responders after 6 cycles. Median OS was 42.2 months and 2.5 years-OS 79.2% (95%CI: 71.2-88.1), with no difference between randomization arms (P=0.69). In univariate analysis, a shorter time since MDS diagnosis (HR=0.97, 95%CI: 0.95-0.99,P=0.011), IPSS int-1 (HR=2.83, 95%CI: 1.05-7.63, P=0.04), an abnormal SNPa karyotype (HR=3.01, 95%CI: 1.26-7.22, P=0.013) and presence of ASXL1 mutation (HR=5.08, 95%CI: 1.48-17.48, P=0.010) were associated with worse survival. In multivariate analysis, however, only time since MDS diagnosis (HR=0.97, 95%CI: 0.95-0.99) and abnormal SNPa karyotype (HR=2.92, 95%CI: 1.07-8.01) remained of prognostic value.
In this prospective AZA trial in LR-MDS resistant to ESA, no significant prognostic factor for response to AZA+/- EPO was identified, but a trend for better response (p=0.07) was seen in SF3B1 mutated patients, i.e 2/3 of patients. Among mutations analysed, only ASXL1 mutation predicted survival in univariate analysis, while abnormal SNPa karyotype predicted survival both in univariate and multivariate analysis (together with MDS duration). Thus, SNP array-based karyotyping and ASXL1 gene mutation analysis may become important tools to predict the long-term outcome of ESA resistant LR-MDS treated by AZA
Guerci-Bresler:Novartis: Honoraria; BMS: Honoraria; Celgene: Honoraria; Amgen: Honoraria. Fenaux:Celgene, Inc: Consultancy, Membership on an entity’s Board of Directors or advisory committees, Research Funding, Speakers Bureau. Gardin:Celgene, inc: Research Funding.
Author notes
Asterisk with author names denotes non-ASH members.
