A Phase II Study of Nivolumab or Ipilimumab with or without Azacitidine for Patients with Myelodysplastic Syndrome (MDS)

Introduction: Myeloid cells express PD1 and CTL-A. The expression of these molecules is activated by treatment with a hypomethylating agent (HMA) both in cell lines and patient samples. We hypothesized that treatment of patients with MDS with immune checkpoint inhibitors (ICPI) blocking CTLA or PD1 with or without azacitidine could have activity both in front line and relapsed MDS.

Methods: To study this, we designed a basket exploratory phase 2 trial of ICPI in MDS. Patients with MDS age 18 or older with adequate renal and hepatic function without history of autoimmune disorders were eligible. Patients were divided into front-line and HMA-failure cohorts. Front-line patients were treated in two different cohorts: AZA + nivolumab (Nivo) and AZA + ipilimumab (Ipi). Patients in the HMA failure cohort were treated in 2 cohorts: single agent Nivo and single agent Ipi. In the HMA failure cohort, patients were first treated with single agent ICPI, and after 6 cycles (or earlier if evidence of progression), the use of azacitidine was allowed to test the concept of re-sensitization. If there was no response, azacitidine was added back. Cohorts could enroll a max of 20 patients with stopping rules for toxicity and response. Nivolumab was administered at a dose of 3 mg/kg on days 1 and 15 every 4-week cycle and ipilimumab at 3 mg/kg on day 1 every 3-week cycle. Azacitidine was used at the standard dose. When combined with azacitidine, nivolumab was administered on days 6 and 20 every 4-week cycle and ipilimumab on day 6 every 4-week cycle.

Results: From 11/12/2015 to 8/10/2017, 76 patients were treated. 41 pts (54%) on front-line cohort and 35 (46%) on HMA failure. The median age was 71 years (range, 39.5-85.7). IPSS risk was Low 3 patients (4%), INT-1 30 patients (40%), INT-2 28 patients (37%), High 11 (15%), and unknown 4 (5%). 29 patients (38%) had complex karyotype; 17 patients (22%), diploid; 25 (33%), other; 5 (7%), insufficient metaphases. Next generation sequencing on whole bone marrow extracted DNA was performed using a 28 or 81-gene panel. Distribution of mutations was as follows: ABL, 4 (5%); ASXL1, 17 (22%); BRAF, 2 (3%); CBL 2 (3%); DNMT3A, 7 (9%); IDH1, 2 (3%); IDH2, 4 (5%); IKZF2, 1 (1%); JAK2, 1 (1%); KIT, 2 (3%); NPM1, 1 (1%); RAS, 14 (18%); RUNX1, 15 (20%); 1, SF3B1, 1 (3%); SRSF2, 2 (3%); TET2, 12 (16%); TP53, 22 (29%); U2AF1, 1 (1%). Median number of marrow blasts was 7% (range, 0-18). Median WBC and platelets was 2.9 (range, 0.5-48.2) and 46 (range, 2-319), respectively. 20 patients were treated with AZA+Nivo, 21 with AZA+Ipi, 15 with Nivo, and 20 with Ipi. In summary, toxicities were as follows; skin rash, 26 (11%); fatigue, 22 (9%); pain, 16 (7%); infection, 14 (6%); febrile neutropenia, 13 (5%); pruritus, 14 (6%); diarrhea, 11 (5%); constipation, 9 (4%); nausea, 10 (4%), ALT elevations, 8 (3%); anorexia, 7 (3%); cough, 7 (3%). Early mortality was observed in 1 patient (1%). Overall response was observed in 15/20 (75%), 15/21 (71%), 2/15 (13%), and 7/20 (35%) of pts treated with AZA+Nivo, AZA+Ipi, Nivo, and Ipi, respectively; CR/CRp was observed in 10/20 (50%), 8/21 (38%), 0 (0%), and 3 (15%) in pts treated with AZA+Nivo, AZA+Ipi, Nivo, and Ipi, respectively. In addition, clearance of detectable mutations through the course of therapy was observed in 3 (15%) pts treated with Ipi, 4 (20%) with AZA+Nivo and 3 (14%) with AZA+Ipi. The median cycle received was 4 cycles (range, 1-25). Among 37 patients with response, the median cycle to response was 3 cycles (range, 1-10). With a median follow up of 20 months, the median overall survival were 12 months, not reached, 8 months, and 8 months with AZA+Nivo, AZA+Ipi, Nivo, and Ipi, respectively (Figure 1A; Figure 1B). Event-free survival were 10 months, not reached, 7 months, 6 months with AZA+Nivo, AZA+Ipi, Nivo, and Ipi, respectively. One-year survival rates were 50%, 68%, 25%, and 45%, respectively. The median event-free survival were 16 months and 7 months in the frontline and HMA failure cohort, respectively (p=0.096); the median overall survival were 17 months and 8 months in the frontline and HMA failure cohort, respectively (p=0.030).

Conclusion: The incorporation of ICPI is feasible in MDS. These agents have significant activity as single agents and in combination in MDS with an acceptable toxicity profile and significant response and survival outcomes, particularly with ipilimumab. Further randomized studies are needed.

View largeDownload slide

View largeDownload slide

Close modal