BACKGROUND: In 'real-life', the Nordic score guides Erythropoietic stimulating agent (ESA) use in lower-risk myelodysplastic syndrome (MDS) with predicted response rates of 25% or 74%. As new treatments emerge, a more discriminating score is needed. OBJECTIVES: To validate existing ESA predictive scores and develop a new score that identifies non-responders. METHODS: ESA-treated patients were identified in 3 MDS registries in Italy and Canada (FISM 555, GROM 233, and MDS-CAN 208). Clinical and disease-related variables were captured. Nordic, MDS-CAN, and IPSS-R-based ESA scores were calculated and documented ESA responses compared. RESULTS: 996 ESA-treated patients were identified. Overall response rate (ORR) was 59%. The database was randomly divided into balanced derivation (n = 463) and validation (n = 462) cohorts. By multivariate analysis, transfusion independence, erythropoietin (EPO) level <100 IU/L, and IPSS low-risk were independently predictive of response. Assigning a score of 1 to each resulted in a scoring system of 0-3 with response rates of 23%, 43%, 67%, and 85%. ORR was concordant in the validation cohort. The 'ITACA' score had the highest discriminating power of response. CONCLUSION: ITACA is an internally-validated predictive SS of ESA response in real-life 'good risk' MDS patients derived from a large international dataset that surpasses others. The incorporation of biologic markers to better identify non-responders is still needed.
BACKGROUND: In 'real-life', the Nordic score guides Erythropoietic stimulating agent (ESA) use in lower-risk myelodysplastic syndrome (MDS) with predicted response rates of 25% or 74%. As new treatments emerge, a more discriminating score is needed. OBJECTIVES: To validate existing ESA predictive scores and develop a new score that identifies non-responders. METHODS: ESA-treated patients were identified in 3 MDS registries in Italy and Canada (FISM 555, GROM 233, and MDS-CAN 208). Clinical and disease-related variables were captured. Nordic, MDS-CAN, and IPSS-R-based ESA scores were calculated and documented ESA responses compared. RESULTS: 996 ESA-treated patients were identified. Overall response rate (ORR) was 59%. The database was randomly divided into balanced derivation (n = 463) and validation (n = 462) cohorts. By multivariate analysis, transfusion independence, erythropoietin (EPO) level <100 IU/L, and IPSS low-risk were independently predictive of response. Assigning a score of 1 to each resulted in a scoring system of 0-3 with response rates of 23%, 43%, 67%, and 85%. ORR was concordant in the validation cohort. The 'ITACA' score had the highest discriminating power of response. CONCLUSION: ITACA is an internally-validated predictive SS of ESA response in real-life 'good risk' MDSpatients derived from a large international dataset that surpasses others. The incorporation of biologic markers to better identify non-responders is still needed.