Long-term outcome of a DLI-based treatment strategy in patients with relapsed AML after TCD allogeneic stem cell transplantation is hampered by GvHD and late relapse.
BACKGROUND Relapse of acute myeloid leukemia (AML) after allogeneic stem cell transplantation (alloSCT) remains a major cause of treatment failure, with poor long-term survival. Donor lymphocyte infusion (DLI) can induce a graft-versus-leukemia effect but may also lead to graft-versus-host disease (GvHD) associated morbidity and mortality. In this study we evaluated the 5-year overall survival of a treatment strategy combining cytoreduction, dose-escalated DLI and interferon-α in patients with relapsed AML after T-cell depleted alloSCT. METHODS In this single-center, retrospective cohort study we analyzed 84 adult patients with relapsed AML after first alloSCT with in vitro T-cell depletion (TCD), and in a proportion of patients combined with in vivo TCD. The study was performed at Leiden University Medical Center between 2005 and 2020, with follow-up through January 2025. Patients with low-burden morphologic relapse (≤10% bone marrow blasts) received DLI without prior cytoreduction; patients with higher-burden morphologic relapse received cytoreductive therapy followed by DLI three weeks later. In the absence of GvHD three weeks after DLI, interferon-α was administered to augment the alloimmune response. Quality of life in long-term survivors was assessed exploratively. RESULTS Six patients received supportive care only, leaving 78 patients treated per protocol. Patients with low-burden morphologic relapse (n=8) received DLI without cytoreduction and achieved a 5-year OS of 60% (95% CI 22-98). In seven additional patients, higher-burden morphologic relapse was unexpectedly diagnosed at the time of a pre-scheduled prophylactic or preemptive DLI. Cytoreduction was deferred to await the effect of this DLI, yet none of these patients survived beyond 18 months. Among 63 patients with higher-burden morphologic relapse who received re-induction therapy with intent to proceed to DLI, 44 (70%) ultimately received DLI, of whom 25 (57%) developed GvHD. The 5-year OS for the re-induction cohort was 7% (95% CI 1-13). Of 44 patients who reached DLI, 24 (55%) achieved complete remission. In this subset, 2- and 5-year RFS were 35% and 19%, respectively, with 5-year cumulative incidences of relapse and non-relapse mortality of 25% (95% CI 12-45) and 42% (95% CI 25-51). Transplantation in first complete remission (HR 0.40, p=0.017) and relapse occurring more than 6 months after alloSCT (HR 0.44, p=0.036) were associated with improved OS. Long-term survivors reported good global mental health but persistent GvHD-related symptoms and impaired social and functional well-being. CONCLUSION A DLI-based strategy can induce durable remission in patients with low-burden morphologic AML relapse after TCD alloSCT, even without prior cytoreduction. In higher-burden morphologic relapse, however, effective cytoreduction followed by DLI and interferon-α is essential, yet long-term outcomes remain poor, constrained by GvHD-associated non-relapse mortality as well as recurrent leukemia. These results define the long-term boundaries of therapeutic DLI and provide a contemporary reference standard for novel post-transplant relapse interventions.