Superior Response Rates for Patients with Myeloid Malignancies Treated with the Combination of Decitabine (DAC) and Arsenic Trioxide (ATO) in a Phase II Adaptive Three Arm Randomization Study: DAC +/- Carboplatin or ATO

Background: Epigenetic therapy is the mainstay of treatment for patients with advanced myelodysplastic syndromes (MDS) and for patients with acute myeloid leukemia (AML) who are not candidates for intense chemotherapy. Decitabine (DAC) is an FDA approved hypomethylating agent for the treatment of MDS and has clinical activity in AML. Despite DAC's activity in MDS and AML, the duration of response is limited, and once patients relapse the prognosis is poor.

In pre-clinical studies, we identified potent epigenetic effects of arsenic trioxide (ATO) alone and in combination with DAC, and also found that carboplatin (Carbo) can enhance gene reactivation by DAC. We therefore initiated a randomized phase 2 clinical trial (NCT02188706) to compare the safety and efficacy of DAC in combination with either Carbo or ATO as compared to single agent DAC in patients with MDS and AML.

Methods: Patients with MDS INT-1 or above or patients with AML that was relapsed or refractory or considered unfit for intense chemotherapy based on age (>60 years), poor performance status (PS2) or comorbidities were randomized to one of three regimens: DAC 20 mg/m2 days 1-5, DAC as above and Carbo AUC 5 on day 8, or DAC as above and ATO 0.15 mg/kg days 1-5. We used adaptive randomization based on response rate, and the adaptive randomization started after 10 patients were randomized to each arm equally. Cycles were scheduled every 28 days for a minimum of 4 cycles. Dose reductions/delays were allowed based on response and tolerability. Patients remained on treatment as long as they continued to benefit without toxicity. The primary endpoint was the composite response rate: the complete and partial response rates using the modified IWG 2006 criteria in MDS and IWG 2003 criteria in AML. Secondary endpoints included survival (median, one year), safety and an evaluation of the epigenetic effects of each arm.

Results: 42 patients have been enrolled on study as of August 2015 (22 AML and 20 MDS). Median age was 65 years (range, 30 to 85 years). There was no statistically significant difference in patients' characteristics between the three arms. Twenty-three (60%) of the patients were previously treated and had relapsed/refractory disease. 38 patients were evaluable for response (13 on DAC alone, 11 on DAC/Carbo and 14 on DAC/ATO). There was not a significant difference in toxicity between the arms, with Grade 3 neutropenia and Grade 3 thrombocytopenia as the most common side effects.

Responses (CR, Cri, PR) were seen in 3/13 patients on DAC alone (23%), 5/11 patients on DAC/Carbo (45%, p=0.18 compared to DAC alone) and 10/14 patients on DAC/ATO (71%, p=0.01 compared to DAC alone). Responses in treatment naïve patients were seen in 3/7 patients with DAC alone, 2/4 patients with DAC/Carbo and 6/6 patients with DAC/ATO. Fitting a logistic regression model comparing treatments accounting for whether the patients had received prior therapy, the DAC/ATO arm showed a statistically higher response rate as compared to DAC alone (p=0.02).

Conclusion: In this randomized phase II study, we find that the combination of decitabine and arsenic trioxide is well tolerated and yields a higher response rate compared to decitabine alone or to decitabine/carboplatin. Our results warrant further studies comparing these regimens for patients with MDS and AML.