Promising Efficacy of Allogeneic Stem Cell Transplantation with Postransplantation Bendamustine in Refractory Acute Myeloid Leukemia: A Dose-Ranging Study

Background: In the preclinical models of allogeneic hematopoietic stem cell transplantation (HSCT) in active leukemia substitution of posttransplantation cyclophosphamide with bendamustine (PTbenda) prolonged survival, possibly due to augmented graft-versus-leukemia (GVL) effect. We conducted a pilot dose-finding study to evaluate the potential of PTbenda to prevent graft-versus-host disease (GVHD) and potentiate GVL.

Methods: The prospective (NCT02799147) Phase I de-escalation study evaluated PTbenda doses of 140-100-70 mg/m2 administered on days +3,+4 with or without additional immunosuppressive agents. All patients received myeloablative conditioning with fludarabine 180 mg/m2 and busulfan 12-14 mg/kg per os. Inclusion criteria were diagnosis of acute lymphoblatic (ALL) or acute myeloblastic leukemia (AML), failure of at least two previous induction courses of chemotherapy or immunotherapy, more than 5% of clonal blasts in the bone marrow and Karnovsky ≥70%. The study planned to enroll 30 patients (pts), however enrollment to the 140 mg/m2 group was halted with only 7 patients enrolled based on the protocol stopping rules. The first cohort of 12 patients received signle-agent PTBena prophylaxis, while the rest received combination of PTBena, tacrolimus and MMF. Five pts had acute lymphoblastic leukemia (ALL) and 22 – acute myeloblastic leukemia (AML). Blast count before conditioning was 31±28%. Five pts had matched sibling, 15 – matched unrelated and 7 – haploidentical, 59% had primary refractory disease and 41% - refractory relapse, 33% had complex caryotype. Half of patients received systemic antibiotics and 22% antifungals for the treatment of febrile neutropenia or infection at the time of enrollment. Median age was 38 years (20-56). Histological confirmation was done for all graft-versus-host disease (GVHD) cases

Results: Median follow-up is 12 months (range 1-47). Ninety three percent of patients engrafted, 89% had CR after HSCT, 63% had negative measurable residual disease (MRD) by flow cytometry and PCR (if applicable). 68% of AML pts were MRD-negative vs 40% of ALL pts (p=0.18). BTBenda demonstrated significant activity in GVHD prophylaxis. In the first cohort without additional immunosuppression acute GVHD grade II-IV was observed only in 42% of patients, while in combination cohort in 30% of patients. On the contrary, all long-term survivors of the single-agent cohort had severe chronic GVHD (100% vs 22% in the combination arm, p=0.005). Enrollment to the 140 mg/m2 arm was stopped prematurely due three consecutive cases of non-relapse mortality, severe viral infections (CMV, HHV6, HHV1,2, BK) and cytokine-release syndrome (CRS). Patients in all cohorts (67% vs 73%, p=0.7 after single-agent PTbenda and combination immunosuppression respectively) exhibited CRS which was presented by macrophage activation syndrome with elevated ferritin (mean 20188 ng/ml). The clinical presentation of CRS included liver functional tests elevation in 95%, skin vasculitis was observed in 74%, soft pallet inflammation and esophagitis in 53%, neurologic symptoms in 37%, disseminated intravascular coagulation in 53% and polyserositis in 42%. Three pts had grade 5 CRS, 5 pts – grade 4, 7 – patients grade 3 and 4 pts – grade 1-2. Haploidentical HSCT was associated with higher risk of grade 3-5 CRS compared to matched related and unrelated allografts (88% vs 42%, p=0.030). The presence of grade 3-5 CRS was also associated with higher incidence of “classical” acute GVHD (53% vs 17%, p=0.050). Two-year event-free survival was significantly higher in acute myeloid leukemia patients: 32% vs 0%, p=0.017. AML patients demonstrated higher event-free survival in the 100 mg/m2 group: 44% vs 30% and 25% than in the 140 mg/m2 and 70 mg/m2 groups, although the differences were not significant (p=0.8). The reasons for failure were non-relapse mortality in 44% of patients and relapse in 26%. No relapses beyond 100 days were documented.

Conclusions: This pilot trial demonstrated that PTbenda can cause CRS and potentiate GVL after allogeneic stem cell transplantation in chemorefractory AML, while no significant benefit was observed for ALL. Combination of PTBenda 100 mg/m2 on days +3,+4 , tactolimus and MMF is the most promising for validation study in AML patients. Better approaches are required to control CRS and reduce early mortality with this GVHD prophylaxis, especially in the haploidentical setting.