Acute Leukemia Relapse after Hematopoietic Stem Cell Transplantation: The Good, the Bad, and the Ugly of Isolated Extramedullary Relapse in a Latin American Population

Hematopoietic stem cell transplantation (HSCT) is an effective therapy for acute leukemia (AL). Relapse represents the main cause of mortality. Isolated extramedullary relapse (iEMR) is atypical and has been related to better outcomes. Here we describe the clinical characteristics and outcomes of AL relapse after HSCT in our study population and analyze the impacts of different types of relapse on survival outcomes. This retrospective, multicenter study included 124 patients age ≥15 years with AL who underwent HSCT between 2004 and 2019. At diagnosis, 66.1% of the patients had lymphocytic AL, 19.7% presented with high-risk features, and 18.5% had extramedullary disease (EMD). At HSCT, 83.1% of the patients were in complete remission (CR), and 44.8% had negative measurable residual disease (MRD). The vast majority of donors were related (96%), including 48.4% HLA-matched and 47.6% haploidentical. Myeloablative conditioning was provided to 80.6% of patients. The median overall survival (OS) was 15 months (95% confidence interval [CI] 9.9 to 20.1 months). Factors associated with improved OS were adolescent and young adult (AYA) patient (P = .035), first or second CR (P = .026), and chronic graft-versus-host disease (GVHD) (P < .001). Acute GVHD grade III-IV (P = .009) was associated with increased mortality. The median relapse-free survival was 13 months (95% CI, 7.17 to 18.8 months); early disease status (P = .017) and chronic GVHD (P < .001) had protective roles. Sixty-eight patients (55%) relapsed after HSCT, with a median time to relapse of 6 months (95% CI, 3.6 to 8.4 months). iEMR was reported in 16 patients (23.5%). The most commonly involved extramedullary sites were the central nervous system and skin. Compared to patients with bone marrow relapse, all patients with iEMR had a diagnosis of acute lymphoid leukemia (P = .008), and 93.8% belonged to the AYA group; regarding pre-HSCT characteristics, iEMR patients had higher rates of negative MRD (P = .06) and a history of EMD (P = .009). Seventy-seven percent of relapsed patients received additional treatment with curative intent. The median OS after relapse (OSr) was 4 months (95% CI, 2.6 to 5.4 months). Factors related to increased OSr included lymphoid phenotype (P = .03), iEMR (P = .0042), late relapse (≥6 months) (P = .014), receipt of systemic therapy including second HSCT (P < .001), and response to therapy (P < .001). Rates of relapse and iEMR were higher than those previously reported in other studies. Advanced disease, reduced-intensity conditioning, and a diminished graft-versus-leukemia effect were factors influencing these findings. At relapse, presenting with iEMR after 6 months and receiving intensive therapy with adequate response were associated with better outcomes. Our results strongly suggest that a personalized approach to treating patients with HSCT is needed to counterbalance specific adverse factors and can positively impact clinical outcomes.