Patient selection for various therapies in myelodysplastic syndromes (MDS) is based on prognostic factors, scoring systems and the individual life expectancy. However, most established risk scores include mainly disease-related parameters and thus focus on leukaemia-transformation rather than survival.
PATIENTS AND METHODS:
To establish a risk score optimized for prediction of survival, we analysed international prognostic scoring system (IPSS)-related and IPSS-independent variables in 400 patients with primary MDS (median age: 71 years; range 18-91) of the Austrian MDS platform. Patients were randomly split into a learning sample (60%) and validation sample (40%). External validation was performed on 93 patients from the Heinrich Heine University (Duesseldorf/Germany).
By multivariate analysis, IPSS, ferritin, age and comorbidities were found to be independent predictive variables concerning survival. Based on weighing these prognostic parameters against each other, we established a novel survival score employing IPSS, ferritin (< 900 ng/mL = 0; ≥ 900 ng/mL = 1), age (< 70 years = 0; 70-79 years = 1; ≥ 80 years = 1·5) and HCT-CI comorbidity (low/intermediate = 0; high = 0·5). Using this score, four prognostic risk groups could be discriminated in the validation sample, with highly significant differences in life expectancy [median survival: LowS (score 0), not reached; Int-1S (score 0·5-1·0), 3·84 years; Int-2S (score 1·5-2·0): 2·72 years; and HighS (score > 2·0): 0·80 years; P < 0·0001].
Our newly proposed score may be a useful tool for survival prediction in MDS and helpful in patient selection for various therapies in daily practice and clinical trials.