performed to investigate mutation status of KIT and 32 myeloid genes.Results: The multivariate model identified five independent predictors of survival: age > 60 years (HR 2.4, CI 1.4-5.0, P < .003), hemoglobin < 10 g/dL (HR 2.0, CI 1.3-3.0, P = .002), platelets < 100 x 10 9 /L (HR 1.7, CI 1.1-2.5, P = .02), presence of one high molecular risk gene mutation (ie, in SRSF2, ASXL1, and/or RUNX1) (HR 2.5, CI 1.6-4.5, P < .0001), and of 2 high molecular risk geme mutations (HR 4.4, CI 2.1-7.3, P < .0001). For assignment of individual scores, we divided the HR value of each variable by the median value of all variables. Accordingly, a weighted score of 1 was assigned to age > 60 years, hemoglobin < 10 g/ dL, platelets < 100 x 10 9 /L, and S/A/R 1 mutation, whereas a score of 2 was assigned to S/A/R 2 mutations. These weighted scores were used to generate three risk groups which comprise the MARS: low-risk, 0 to 1; intermediate risk, 2; high-risk, 3 to 5 with median OS not reached, 3.9 years (95% CI 2.1 to 5.7 years) and 1.9 years (95% CI 1.3 to 2.6 years) (P < .0001) (Figure 1A), respectively. The MARS was independent of the World Health Organization (WHO) classification and was confirmed in an independent validation set (Figure 1B). The MARS was also predictive for leukemia-free survival (P < .0001). On basis of receiver operating characteristic curve analysis, the C-index was 0.42 for the WHO classification and 0.81 for the MARS.
Summary/Conclusion:We conclude that the WHO classification remains the pivotal diagnostic and prognostic tool for subtyping of SM into indolent SM and AdvSM. The MARS is a WHO-independent and complementary tool for the heterogeneous cohort of patients with Ad-vSM by defining three risk groups based on a five-parameter prognostic score which may improve upfront treatment stratification for these rare hematologic neoplasms.