Background: Prognostication of patients with myelodysplastic syndromes (MDS) using IPSS-R is not possible in administrative database analysis. In 2014, a claims-based prognostic scoring system, the SEER-Medicare MDS Risk Score (SMMRS) was developed to risk-stratify MDS patients for overall survival using ICD9 codes {Uno, 2014}. In 2015, ICD9 codes were replaced by ICD10 codes. In this study, we updated SMMRS with inclusion of both ICD-9 and -10 codes, and studied its predictive value for overall survival, leukemia-free survival and initiation of treatment among MDS patients. Methods: Patients with MDS were identified through ICD-O-03 codes in the SEER-Medicare database between 2007 and 2017. SMMRS was calculated using MDS subtype, age, comorbidities, cytopenias, transfusion-dependence, and recent hospital admissions. We analyzed the association of SMMRS with overall survival, leukemia-free survival (time from MDS diagnosis to leukemia), and initiation of treatment (hypomethylating agents /stem cell transplant) in Cox proportional hazards models while adjusting for potential confounders. Results: Patients with MDS (N = 12,710) had a median age of 80 years, half were males, and most were Caucasians. SMMRS stratified patients into low-risk (n = 3155), intermediate-1 risk (n = 3255), intermediate-2 risk (n = 3121), and-high risk (n = 3209) categories at diagnosis. The overall survival, leukemia-free survival and treatment initiation for patients in each SMMRS category is summarized in Table. The 5-year overall survival for the entire cohort was 20%. At 5 years, 10% of the MDS patients had transformed to leukemia, and 1/3rd had started MDS treatment. Conclusions: SMMRS based on ICD-9 and -10 codes predicts overall survival, leukemia-free survival and MDS treatment initiation by stratifying MDS patients into 4 risk-categories. It can be utilized to design/analyze future observational studies.

* Unable to estimate as few patients were alive at this time-point # Adjusted for Gender, Race, Ethnicity, MDS treatment ̃ Adjusted for Gender, Race, Ethnicity