Characteristics and survival of secondary acute myeloid leukemia arising from myelodysplastic syndrome (MDS-sAML).

e19075 Background: sAML is a heterogenous category including therapy-related AML, AML arising from MDS or myeloproliferative neoplasm, and other biologically distinct groups. Combining these as one entity likely masks important clinical and therapeutic differences. We defined the clinical profile and survival for MDS-sAML. Methods: The population consisted of patients (pts) with MDS identified in SEER-Medicare (2007-17) by ICD-O-3 histology codes and continuous Medicare parts A+B coverage. Transformation to leukemia was confirmed by at least 2 AML codes during follow up. Established algorithms identified MDS type and risk group, Charlson comorbidity index (CCI), and transfusion dependence for red blood cells (RBC) and platelets. Logistic regression identified risk factors for transformation to MDS-sAML. Kaplan-Meier estimation and multivariable Cox regression were used to analyze overall survival (OS). Results: Of 15,227 MDS pts, 1,871 transformed to AML. MDSsAML incidence varied by MDS risk group and histology. Time to transformation was shorter with higher risk groups. Younger age (odds ratio [OR] = 1.6, 95% confidence interval [95 CI]: 1.4-1.9), lower CCI (OR = 1.6, 95 CI: 1.4-1.9), higher risk (OR = 2.3, 95 CI: 2.0-2.5), transfusion dependence for RBC (OR = 1.2, 95% CI: 1.1-1.3) or platelets (OR = 2.5, 95 CI: 2.1-3.0), were associated with transformation (all respective p < 0.05). Median OS (mOS) after transformation was 3.3 months. The 1- and 2-year OS rates for MDS-sAML were 25% and 12%. MDS-sAML survival varied by antecedent MDS histology and risk group, being longer for pts with antecedent low risk MDS (p < 0.001). Sex, race, and CCI did not affect survival after transformation. Older age (hazard ratio [HR] = 1.5, 95 CI: 1.2-2.0), higher MDS risk (HR = 1.5, 95 CI: 1.2-1.9), RBC transfusion dependence (HR = 1.4, 95 CI: 1.3-1.6), and platelet transfusion dependence (HR = 1.5, 95 CI: 1.3-1.7), were associated shorter MDS-sAML survival (all respective p < 0.05). Conclusions: Risk of MDS-sAML is substantial, with 5% AML transformation even in lower risk MDS histology. Real-world survival of MDS-sAML pts is dismal and lower than reported in clinical trials. Worse MDS risk group before transformation is associated with lower OS after developing AML. These data support a unique pathophysiology of MDS-sAML. [Table: see text]