Efficacy and Toxicity of Donor Lymphocyte Infusions in Relapsed Acute Myeloid Leukemia Post Allogeneic Hematopoietic Cell Transplantation: Single-Center Experience

Introduction: Allogeneic hematopoietic cell transplantation (alloHCT) is a potentially curative treatment in acute myeloid leukemia (AML) and is pursued as a means to decrease the hazard of relapse. However, relapse of the underlying disease following both conventional and reduced intensity HCT still remains an unresolved problem highlighting the necessity to investigate novel and different approaches to target it. Over the last decades, Donor Lymphocyte Infusions (DLIs) have been proposed as a major immunotherapeutic strategy in relapsed patients. In the era of emerging modified immunological approaches, efficacy and toxicity of DLIs remain to be clarified. The present retrospective analysis aimed to assess efficacy and toxicity of DLIs in relapsed AML patients post alloHCT.

Methods: We retrospectively studied relapsed patients undergone alloHCT for AML between January 1990 and June 2016. Until 2007, our center's immunotherapeutic strategy consisted of immediate cessation of immunosuppression, administration of IL-2/ΙFNa-2β if chimerism reduction was evident and escalated doses of DLIs depending on the time period and the donor type (sibling or unrelated). Since then, we have stopped the administration of immunomodulatory agents due to inefficacy and complications shown in interim analysis.

Results: In total, 266 AML patients underwent transplantation from 162 identical sibling, 8 relative, 8 haploidentical, 88 matched unrelated donors; 89 patients relapsed. Fifty two patients aged 40 (range 7-63) years old received DLIs. They had received peripheral (47) or bone marrow (5) grafts, post myeloablative (44) or reduced intensity (8) regimens. Twenty out of 52 patients were transplanted in complete remission and 32 in relapsed or refractory disease. Their donors were identical siblings (42), matched unrelated(5), non-identical siblings (3) and haploidentical donors (2). With a median follow-up of 11.8 (2.6-181) months, relapse was documented 4.6 months (0.4-28.5) post alloHCT. The majority of patients (45) received induction and/or salvage chemotherapy or azacytidine because of morphological remission (blasts ≥ 20 %). Twenty four patients received chemotherapy and cryopreserved graft infusion, with a dose of CD34⁺>1×10⁶/kg body weight, to prevent neutropenia. Median DLIs were 2 (1-6) and CD3⁺ were 1.9 (0.5-4.5) x10⁸/kg in sibling and 0.9 (0.5-1.43) x10⁶/kg in unrelated transplants. Second transplant from the original relative donor (1 bone marrow /2 peripheral blood grafts) was performed in 3 patients early after first HCT (5-9 months) but it failed due to disease progression. Complete remission was observed in 17 patients (34%); whereas acute graft-versus-host-disease (GVHD) in 27 patients (52%) and chronic in 13 patients (25%). In total, 43 patients died; 35 of disease, 6 of DLI-related mortality and 2 of other causes. One-year Disease-free survival (DFS) post DLIs was 50% and 2-year 34.4%. The presence of GVHD was associated with higher DFS (74% vs 4%, p=0.015). In multivariate analysis, independent favorable factors for higher DFS were the presence of GVHD (p<0.001/beta=4.086), the myeloablative conditioning (p=0.011/beta=3.087) and the longer time to relapse (p=0.004/beta=0.921). No association with survival was found for the other factors studied (age, gender, disease phase, type of transplant and donor, HLA histocompatibility).

Conclusion: Our experience of more than 2 decades with DLIs aiming at improving disease free survival and lowering relapse has shown acceptable toxicity and efficacy of immunotherapy in relapsed AML patients. We have also demonstrated the favorable impact of the graft-versus-leukemia effect that was shown to be stronger with myeloablative conditioning and the presence of GVHD. Despite encouraging results, there is an emerging need of alternative strategies. Systematic prophylactic use of DLIs may be useful, but should be balanced against the risk of severe or fatal GVHD. Furthermore, early administration of biologically targeted therapies, may be an attractive treatment in certain specifically defined subgroups of patients. In this context, the innovative strategies would be ideally more effective and less toxic in treatment of relapse.