Decitabine plus ipilimumab in r/r MDS/AML in the post-BMT or transplant naïve settings

Dr Jacqueline Garcia - Dana-Farber Cancer Institute, Boston, USA

At ASH 2020 I presented abstract 170 which is a clinical trial that I’m presenting on behalf of my co-investigators. It’s a phase I study evaluating the safety and efficacy of decitabine plus the CTLA-4 blockade agent ipilimumab in patients with relapsed or refractory myelodysplastic syndrome or acute myeloid leukaemia in both the post-transplant or transplant naïve settings.

The rationale for this clinical trial was both preclinical and clinical. Clinically there has been some really encouraging data with single agent ipilimumab demonstrating some anti-leukemic activity in the post-transplant AML relapsed setting. This was presented by Matt Davids and Rob Soiffer and what they had found is that there was an encouraging rate of complete remission with ipilimumab at 10mg/kg predominantly in patients with extramedullary or SOL leukaemia cutis. Based on the initial clinical activity we noted also that there was some immune toxicity, including induction of GVHD. We asked whether or not we could build upon this initial clinical data and increase the response without increasing immune toxicity. We also had the biologic question of whether or not CTLA-4 blockade only had a role in the post-transplant setting where there is an alloreactive environment or whether or not patients who are transplant naïve and are not post-transplant, whether or not they would benefit from this type of therapy.

So based on these findings we moved forward and generated a phase I study that is CTAP [?] supported. On this phase I multicentre study we assessed the safety of decitabine plus ipilimumab and we assessed it in two separate arms and both arms received the same therapy so to allow for comparison. Recognising that patients who are post-transplant may have different safety and efficacy compared to those who are transplant naïve, we separated patients based on transplantation status. So patients would receive on this trial a month of decitabine at standard dosing of decitabine 20mg/m2 days 1-5. This is a lead-in period, it allowed for patients that were recently potentially taken off immunosuppressive therapy to make sure that they could be effectively off those therapies without GVHD induction. Then all patients would then receive decitabine plus ipilimumab and the dose of ipilimumab would be dependent on the dose level assigned. Ipilimumab was tested at doses of 3, 5 and 10mg/kg. This was given monthly on day 1 with each decitabine cycle.

For patients that got through four combination cycles safely if they had at least stable disease or better they were allowed to enter what was called maintenance therapy where decitabine was continued monthly as standard of care dosing but ipilimumab was added on day 1 every other cycle.

So the primary objective of this trial was to establish the MTD or recommended phase II dose of ipilimumab when combined with decitabine for patients that were either post-transplant or transplant naïve. Secondary objectives included evaluating for objective responses following IWG criteria. We also wanted to look at the one year overall survival and PFS as we know that patients that relapse, particularly those that relapse post-transplant, the overall survival for these patients is incredibly low as many therapies do not work in this setting. We also wanted to measure the incidence and severity of GVHD.

In terms of the patients that were eligible for this trial, it included patients who had relapsed, refractory or secondary MDS or AML. We did allow prior HMA therapy. Patients that were post-transplant had to have no history of acute GVHD grade 3 or higher and, importantly, they had to be off their systemic immunosuppressive therapy for at least two weeks prior to study start. That was important because we did not want patients that had active GVHD to come onto trial.

Because of our concerns for safety and recognising what had happened in the pilot studies with ipilimumab alone, we had a purposely prolonged dose limiting toxicity window of eight weeks from the first dose of ipilimumab in order to capture serious immune toxicities that were not easily reversed with corticosteroids. We also wanted to capture acute GVHD events and any study treatment related death.

In terms of the baseline characteristics of the study patients, at the time of this presentation we were already in dose expansion, making significant progress, and we presented 47 patients that were enrolled onto trial. In arm A that included 25 patients and in arm B, which is the transplant naïve group, it included 22 patients. The median age of patients was 67 for those that were previously transplanted. 88% of the patients had AML, 68% of patients had prior hypomethylating agent therapy. The patients were pretty heavily pre-treated with a median of three prior regimens prior to coming onto study. Importantly, in arm B, the transplant naïve group, the median age was 75 so comparable; 64% of patients had AML; 64% had prior HMA therapy. So this is a pretty heavily pre-treated group, again having received prior HMA therapy and I think that’s important when thinking about the efficacy that I’ll present later.

When we looked at non-hematologic toxicities that were not immune based, for patients that received ipilimumab on this clinical trial we did not see a safety profile that looked to be more excessive than what would be expected from the underlying disease and from decitabine alone. But when we took a deep dive in capturing and sharing the immune toxicities we separated them based on arm A, which is the prior transplantation group, versus arm B, those without prior transplantation history. What we do see is that in the prior transplantation group the immune toxicity rate is about 38% and these included events that were GVHD and no overt GVHD. We did see that most cases of GVHD were acute and were mild. In terms of serious events there was one event of acute GVHD grade 3 in one patient, we called this a DLT, and this patient unfortunately later developed septic shock as a complication of immunosuppression, infection, multi-organ dysfunction. Chronic GVHD was also observed on this trial. There were two cases of moderate chronic GVHD and one case of severe chronic GVHD. All of these were controlled with immunosuppressive therapies and, importantly, systemic corticosteroids.

In terms of the transplant naïve patients what we saw in the 16 patients that received ipilimumab, there was an immune toxicity rate of 44%. It was more notable at the highest dose level of ipilimumab at 10mg/kg. Fortunately most cases were not severe; the one grade 4 case that was severe that we labelled as a DLT was an event of new onset HLA refractory thrombocytopenia that was new for our patient on study. HLA refractory thrombocytopenia does happen to our patients with MDS and AML that require frequent transfusions but as this occurred while on study we called this a DLT. So one DLT was seen in the post-transplant group at 10mg/kg and one DLT was observed in the transplant naïve group at 10mg/kg.

In terms of the preliminary safety we saw that there were responses despite prior HMA therapy. In fact, in arm A the objective response rate was 19%. Arm A, again, is our post-transplant arm. There were three complete remissions and one marrow CR. The median time to response was 1.7 months which is encouraging. Four of these four patients, so 100% of these responders, had prior HMA therapy but, really interesting, only 50% of these responders had any evidence of immune related toxicity.
Among the patients who were transplant naïve in arm B 56% of patients had a clinically meaningful objective response and this included four complete remissions, two CRis, three marrow CRs. The median time to response is a bit longer, 3.9 months. 56% had prior HMA therapy and 22% of these responders had immune related AEs.

When we took a look at the outcomes after ipilimumab in terms of survival the median follow-up time as this trial is still ongoing is only 5.5 months but we saw that the median overall survival overall was 12.8 months. For arm A, or post-transplant patients, the median overall survival was 7.6 months and in arm B it was 18.3 months. Dose expansion is still ongoing to capture safety.

We shared some initial correlative studies and we were able to demonstrate using multiplex immunofluorescence of serial time points among responders and non-responders that there was a gradual influx of cytolytic T-cell infiltrates after ipilimumab that had been observed in single responders. When we looked for this pattern in non-responders this was not as clearly evidenced; although we do see the presence of immune cells, the evidence for cytolytic capacity based on granzyme B costaining was not observed among non-responders. Further follow-up with a larger patient panel and with multiple time points is currently underway in order to look for significance. There was just a descriptive presentation.

In terms of looking at the immunophenotype of the immune cells that were present we performed multiparametric flow cytometry of blood from patients treated with ipilimumab at 10mg/kg which is our highest dose cohort which is our recommended phase II dose. What we see when we compare pre-treatment versus end of decitabine alone versus end of combination cycles with ipilimumab that there is an increase in ICOS positive T-cells that is seen after ipilimumab therapy, a reduction in naïve T-cells and an increase in the memory CD4 T-cell compartment after ipilimumab therapy. This is seen regardless of response, suggesting that ipilimumab does play a role in what we’re seeing in our patients.

So overall we are sharing the data from our dose escalation, dose expansion, study in relapsed/refractory MDS and AML patients that received decitabine followed by decitabine plus ipilimumab therapy. We report the immune toxicity events that were frequent but mostly controllable. We share the objective responses that are observed even in the transplant naïve setting. Importantly, we highlight that in the post-transplant arm, although the GVHD induction rate was common at 29%, only 10% of cases were severe and response may not require GVHD induction. In the arm B, or the transplant naïve group, we highlight that immune toxicity is frequent at 40% but the severe immune related toxicity rate was low at 6% and most events were indeed reversible fairly promptly with steroid intervention.

We did see promising clinical activity observed, including among patients with prior HMA therapy, suggesting that an alloreactive environment may not be required to benefit from CTLA-4 blockade. So next steps for us are to complete the study, we’re completing expansion. Hopefully we’ll be able to do that within the next couple of months. We’re doing an extensive correlative science component to really understand the mechanism of response and to really delineate details of people that are long-term versus short-term responders and to identify patients that potentially might benefit from this therapy most.