Maintenance Therapy with Low-Dose Azacitidine (AZA) after Allogeneic Hematopoietic Stem Cell Transplantation (HSCT) for Relapsed or Refractory AML or MDS: A Dose and Schedule Finding Study.

Relapse is the major cause of treatment failure after HSCT for relapsed AML. Maintenance therapy may provide an “adjuvant’ support for the allogeneic graft-versus-leukemia effect, decreasing the likelihood of recurrence. AZA is a DNA methyltransferase inhibitor that induces DNA hypomethylation, leading to leukemic cell differentiation and potentially increased tumor immunogenicity. FDA approved doses are however unlikely to be tolerated due to myelosuppression. We hypothesized that AZA will reduce relapse rates, and designed a phase I clinical trial to determine the safest dose and schedule combination. Methods . Eligible were patients (pts) with AML/MDS not in 1st complete remission (CR), not candidates for ablative regimens due to age or comorbidities. Conditioning regimen was gemtuzumab ozogamicin 2 mg/m 2 , fludarabine 120mg/m 2 , and melphalan 140mg/ m 2 . GVHD prophylaxis was tacrolimus/mini-methotrexate (ATG in unrelated donor HSCT). We investigated 4 AZA doses: 8, 16, 24 and 32 mg/m 2 SQ daily x 5 starting on day +42, given for 1–4 28-day cycles (schedule). An outcome-adaptive method was used to determine dose and schedule (number of cycles): pts were assigned to a dose/ schedule combination on the basis of the data (organ/hematologic toxicity) from all pts treated previously in the trial. Pts in CR on transplant day+30, with donor chimerism, no grade III/IV GVHD, platelet >10,000/mm 3 and ANC >500/mm 3 were eligible to receive AZA. Methylation status of long interspersed nuclear elements (LINE) was analyzed by pyrosequencing as a surrogate marker of global DNA methylation ( figure ). Results. 80 pts were transplanted; 44 (56%) were eligible to receive AZA: 2 refused; 42 pts. (4 too early) received it. Table shows pts characteristics. 88 cycles were delivered at 8 (n=7), 16 (n=5), 24 (n=21) and 32 mg/m 2 (n=9). AZA-associated/possibly associated toxicities were grade (gd) I/II or III thrombocytopenia (n=5; n=2), gd I nausea (n=5), gd II fatigue (n=2), gd III transaminase elevation (n=1, AZA + posaconazole), conjunctival erythema (n=1), pruritus (n=1), gd I confusion (n=2), retina hemorrhage (possibly pre-existing), gd II creatinine elevation (n=1), oral ulcers (n=1), papilledema (n=1), and pulmonary hemorrhage (n=1; second HSCT, fungal pneumonia/hepatic VOD during the 1st AZA cycle, evolving with thrombocytopenia/bleeding). Reversible thrombocytopenia was more frequent at 32 mg/m 2 . Infections: bacteremia, n=5; respiratory, n=6; parainfluenza, n=2, CMV reactivation, n=6. Most pts were 100% donor chimeras at the start of AZA. AZA did not affect GVHD incidence ( table ). 10 pts have relapsed, 3 while on AZA 16 or 24 mg/m 2 . Median follow-up is 13 months (3–31; n=26). 12 pts have died, 8 of recurrence, 2 of GVHD, 1 of pneumonia and 1 of unknown causes. Day +100 non-relapse mortality was 6 %; 4 pts died within the first 100 days post HSCT due to relapse (n=2), pneumonia (n=1), and GVHD (n=1). Actuarial 1-year event-free and overall survival is 58% and 72%. No dose was found to significantly affect global methylation (gene specific methylation studies are ongoing). The trial has reached the dose 32 mg/m 2 and the projected maximum number of cycles. Thrombocytopenia prevented escalation to the next level (40 mg/m 2 ). Conclusion . AZA 32 mg/m 2 is safe and can be administered for at least 4 cycles to heavily pre-treated pts. with comorbidities. The safety profile indicates that longer periods of administration should be investigated. We will initiate a randomized, controlled study of AZA for one year versus best standard of care (no maintenance therapy) for similarly highrisk patients with AML/MDS.