[Allogeneic hematopoietic stem cell transplantation for 510 patients with active relapsed/refractory acute myeloid...

Objective: To investigate the long-term survival outcomes of salvage allogeneic hematopoietic stem cell transplantation (R/R AML) . Methods: This was a retrospective real-world study. We conducted a retrospective analysis of 510 patients with R/R AML who underwent allo-HSCT at Aerospace Center Hospital between July 2012 and December 2024. All patients had active disease before transplantation, defined as a bone marrow blast proportion of at least 5%. Overall survival (OS) and disease-free survival (DFS) were assessed using the Kaplan-Meier method, whereas the cumulative incidence of relapse (CIR) and non-relapse mortality (NRM) were evaluated by competing-risk models. Univariate analysis was performed using the log-rank test, and multivariable analysis was conducted using Cox regression for survival outcomes and Fine-Gray regression for competing-risk outcomes to identify independent prognostic factors. Results: In this cohort of 510 patients, the median follow-up duration was 50 months (range, 0-151 months). The estimated 10-year overall survival (OS) and disease-free survival (DFS) rates were 35.5% (95% CI: 31.2% -40.5% ) and 33.2% (95% CI: 29.1% -38.0% ), respectively. Multivariable analysis demonstrated that European LeukemiaNet (ELN) risk stratification was a core prognostic factor. Patients in the intermediate-risk (HR=1.74, P=0.002) and adverse-risk (HR=1.97, P<0.001) categories had markedly lower OS and DFS rates. Mild-to-moderate chronic graft-versus-host disease (cGVHD) served as a protective factor for OS and DFS (both P<0.001). A diagnosis-to-transplant interval longer than 10 months (HR=1.56, P<0.001) was correlated with poor survival outcomes. A higher infused CD34(+) cell dose (≥4.3×10(6)/kg) was independently associated with superior OS (HR=0.64, P<0.001) and DFS (HR=0.62, P<0.001). In competing-risk analysis, intermediate and adverse ELN risk stratification (HR=2.73 and 2.90, respectively; both P<0.001) and elevated bone marrow blast percentage (≥50% ) (HR=1.62, P=0.037) were independently linked to increased relapse risk. Mild, moderate, and severe cGVHD, compared with no cGVHD (HR=0.38, 0.13, and 0.24; P<0.001, P<0.001, and P=0.015, respectively), and grade Ⅲ-Ⅳ acute graft-versus-host disease (aGVHD) (HR=0.54, P=0.033) were associated with a lower CIR. Regarding non-relapse mortality (NRM), grade Ⅲ-Ⅳ aGVHD (HR=2.36, P<0.001) and severe cGVHD (HR=2.74, P<0.001) were major risk factors, and advanced age (HR=1.47, P=0.033) was also associated with higher NRM. Conclusions: This study indicates that long-term disease-free survival may be attained through salvage allo-HSCT in patients with R/R AML despite active disease at transplantation. Long-term outcomes are mainly determined by disease biology and post-transplant immune effects, with ELN risk stratification and GVHD playing dominant roles, whereas pre-transplant tumor burden mainly affects relapse risk but is not an independent determinant of long-term survival. These findings suggest that greater emphasis should be placed on risk stratification and transplantation timing in clinical practice, rather than solely pursuing complete remission prior to transplantation.