Risk factors for leukemic transformation in myeloproliferative neoplasms and prognostic factors in secondary acute myeloid leukemia: an exploratory analysis of the SEER database.

To identify risk factors for secondary acute myeloid leukemia (sAML) transformation in myeloproliferative neoplasms (MPN), determine prognostic factors, and construct a reliable prognostic risk model, this study performed a retrospective analysis using the US Surveillance, Epidemiology, and End Results (SEER) database. Multivariable competing risk regression revealed that female (HR = 0.84, P = 0.012), polycythemia vera (PV, HR = 0.29, P < 0.0001) or essential thrombocythemia (ET, HR = 0.26, P < 0.0001) were protective factors against sAML transformation. In contrast, American Indian/Alaska Native/Asian Pacific Islander (AIAN/API) race (HR = 1.31, P = 0.034), household income <$70,000 (HR = 1.21, P = 0.022), chemotherapy history (HR = 1.80, P < 0.0001), and MPN diagnosis during 2011–2022 (HR = 1.29, P = 0.0001) were risk factors for leukemic transformation. Multivariate Cox proportional hazards regression showed that MPN-phase chemotherapy (HR = 1.25, P = 0.002), MPN diagnosis at age 60–75 years (HR = 1.34, P = 0.012), and sAML diagnosis at age ≥ 75 years​ (HR = 1.83, P = 0.001) were independent risk factors for post-progression survival (PPS) in patients with MPN-sAML. while sAML-phase chemotherapy was an independent protective factor (HR = 0.39, P < 0.001). Subgroup analysis revealed heterogeneity in prognostic factors among different sAML subgroups, but the protective effect of sAML-phase chemotherapy was consistently observed across all subgroups. Further multivariable Cox proportional hazards subgroup analysis suggested heterogeneity in the benefit of sAML-phase chemotherapy across age groups (P for interaction = 0.041), with attenuated benefit observed in patients aged 60–75 years (HR = 0.41) and ≥ 75 years (HR = 0.42) compared to younger patients aged 20–60 years (HR = 0.17). The prognostic risk model constructed based on independent prognostic factors and MPN subtypes can effectively stratify patients and provide a reference for risk stratification in patients with MPN-sAML.