Therapy-related myelodysplastic syndrome (tMDS) is a poor-risk subtype of MDS with no standard treatment options, and yet patients (pts) with tMDS are often excluded from trials where they would have the opportunity to benefit from novel treatment approaches. DNA methyltransferase inhibitors are a treatment option for tMDS, and are being evaluated as a bridge to stem cell transplant in these often heavily pre-treated patients to avoid the organ toxicity of intensive chemotherapy. However, the response rate of tMDS to DNA methyltransferase inhibitor (DNMTI) therapy is unknown. In this retrospective study, adult patients tMDS were culled from a fully annotated, IRB-approved database of all MDS patients who received either decitabine (DAC) on both the 3-day and 5-day schedules or 5-azacytidine (5-aza) at our institution from 4/8/2002 to 6/18/2010. Patients who received interrupted therapy were only analyzed for response to their initial course of therapy. Patients who received sequential DNMTI therapy (i.e., DAC followed by 5-aza or 5-aza followed by DAC) were included, but response to only their initial therapy was assessed. Responses were determined using the modified International Working Group criteria (Cheson BD, et al, 2006). Of the 35 patients initially identified with tMDS who received DNMTI therapy, 4 were deemed inevaluable for response due to marrow involvement with the primary malignancy (n= 1), missing records (n=2), or delivery of < 1 full cycle of therapy (n=1). The 31 evaluable pts included 14 males and 17 females with a median age of 65 years (range 25–85). Therapy for the primary malignancy included chemotherapy alone (n=13), chemotherapy plus radiotherapy (n=14), radioactive iodine and radiotherapy (n=2), radioactive iodine and chemotherapy (n=2), and autologous stem cell transplant (n=3). Prior to DNMTI therapy, the MDS FAB subtypes were as follows: RA, n= 6; RARS, n= 3; RAEB, n= 19; RAEBt, n=2; CMMoL, n=1. Pre-DNMTI therapy included lenalidomide (n=4) and alloSCT (n=1). Of the 31 evaluable patients, 20 received DAC, including 7 pts who received tretinoin with DAC in a clinical trial, and 11 received 5-aza. DAC recipients received a median of 2 cycles of therapy (range, 1–12) and 5-aza recipients received a median of 5 cycles (range, 1–9). Best responses were as follows: CR, n=1; Marrow CR plus HI, n=6 (3 trilineage HI, 1 HI-P+ HI-N, 2 HI-P); Stable Disease, n=6; Progressive Disease, n=6; Failures (death during 1st cycle or before response evaluation), n=3. Rate of CR + mCR was 22% (n=7). Additional patients had inevaluable (aparticulate) marrows, or refused follow-up marrow studies, but showed signs of stable (n= 3), improved, (n=2; HI-P, HI-P+HI-N), or progressive cytopenias (n= 3). Median time to best response was 1.5 cycles (range 1–6). Fifty-eight percent (n=18) of 31 pts achieved stable disease or better responses (including 4 pts with stable cytopenias or HI with inevaluable marrow response). Four patients proceeded directly to transplant after DNMTI therapy. Two subsequently died from relapsed disease after transplant, while 1 pt is lost to follow-up and 1 pt is without evidence of MDS 2.5 years after transplant. Nine pts had persistence of their primary malignancy during DNMTI therapy, and 5/9 pts required interruption or cessation of their DNMTI therapy because of progressive primary malignancy. 24/31 pts died from complications of MDS (n=5) or subsequent AML (11), complications of MDS/AML with likely contribution from their primary malignancy (n=4), infection during DNMTI nadir (n=2), GVHD post AlloSCT (n=1), or unknown reasons (n=1). Living pts (n=7; median follow-up from start of DNMTI therapy = 12.5 months, range 4.1 – 35.1 months) include 5 who are not transplant candidates. In conclusion, DNMTI therapy produced modest clinical benefit in our tMDS cohort. In some patients, persistence of the primary malignancy interfered with our ability to deliver optimal DNMTI therapy and to assess response. Although DNMTIs represent a potential therapeutic option for tMDS, treatment of a larger cohort is required to clarify the response rate of these agents in tMDS.
Asterisk with author names denotes non-ASH members.