We aimed to compare the ability of recently developed prognostic indices for myelodysplastic syndromes to identify patients with poor prognoses within the lower-risk (low and intermediate-1) categories defined by the International Prognosis Scoring System (IPSS).
We included patients with de-novo myelodysplastic syndromes diagnosed between Nov 29, 1972, and Dec 15, 2011, who had low or intermediate-1 IPSS scores and were in the Spanish Registry of Myelodysplastic Syndromes. We reclassified these patients with the new prognostic indices (revised IPSS [IPSS-R], revised WHO-based Prognostic Scoring System [WPSS-R], Lower Risk Scoring System [LRSS], and the Grupo Español de Síndromes Mielodisplásicos [Spanish Group of Myelodysplastic Syndromes; GESMD]) and calculated the overall survival of the different risk groups within each prognostic index to identify the groups of patients with overall poor prognoses (defined as an expected overall survival <30 months). We calculated overall survival with the Kaplan-Meier method.
We identified 2373 patients. None of the prognostic indices could be used to identify a population with poor prognoses (median overall survival <30 months) for the patients with low IPSS scores (1290 individuals). In the group with intermediate-1 scores (1083 individuals), between 17% and 47% of patients were identified as having poor prognoses with the new prognostic indices. The LRSS had the best model fit with the lowest value in the Akaike information criteria test, whereas the IPSS-R identified the largest proportion of patients with poor prognoses (47%). Patients with intermediate-1 scores who were classified as having poor prognoses by one or more prognostic index (646 [60%] individuals) had worse median overall survival (33·1 months, 95% CI 28·4-37·9) than did patients who were classified as having low risk by all prognostic indices (63·7 months, 49·5-78·0], HR 1·9, 95% CI 1·6-2·3, p<0·0001) INTERPRETATION: Recently proposed prognostic indices for myelodysplastic syndromes can be used to improve identification of patients with poor prognoses in the group of patients with intermediate-1 IPSS scores, who could potentially benefit from a high-risk treatment approach.