BACKGROUND: Patients with higher risk myelodysplastic syndromes (MDS) according to the International Prognostic Scoring System (IPSS) have poor life expectancy making accurate prediction of overall survival (OS) a critical issue for optimal and personalized patients' management.PURPOSE: The primary objective of this study was to develop a patient-based prognostic index for higher risk-disease, that would include self-reported fatigue into the widely used IPSS classification. A secondary objective was to examine whether this patient-based index would provide more accurate OS prediction than the standard IPSS classification.PATIENTS AND METHODS: Analysis is based on 280 newly diagnosed patients with MDS classified with an intermediate-2 or high-risk score (i.e., higher-risk) according to the IPSS. Patients were recruited in an international prospective cohort observational study involving 37 centers. OS was defined as the date of diagnosis of IPSS intermediate-2 or high risk MDS up to death for any cause. Patients were censored at the date of last follow up if not dead at the time of analysis. Before treatment start, all patients completed the EORTC QLQ-C30 questionnaire and the fatigue scale was a priori selected for possible inclusion into the IPSS. Among all observed values of the fatigue score (range: 0-100) we looked for a threshold defining four risk groups, formed by patients reporting either low or high fatigue respectively in intermediate-2 and high risk IPSS groups. The final prognostic index was developed based on univariate and multivariate Cox models. Differences among Kaplan-Meier OS estimation of new risk categories were assessed by log-rank test. Sensitivity analyses were performed, 1) assessing differences in patients’ baseline characteristics among risk groups, by Kruskal-Wallis and Fisher’s exact tests 2) accounting for several potential confounding factors (baseline and time-dependent) in a multivariate extended Cox model, including treatment received after baseline assessment and further evolution into acute myeloid leukemia, 3) performing a bootstrap resampling procedure to internally validate the final prognostic index. Discrimination and calibration of the new index were evaluated. For all analyses, α=0.05.RESULTS: With a median follow- up of 15 months (IQR 8-27) we observed 113 deaths. The median OS of the overall population was 17 months (95% CI, 15-19). The majority of patients (N=165, 59%) received treatment with hypomethylating agents. The final cut-off value selected for the EORTC QLQ-C30 fatigue scale was 45 points discriminating between patients with low (<45 points) or high fatigue (≥45 points). A new risk score classification was then developed, namely, the Fatigue(FA)-IPSS(h), enabling to distinguish three risk group categories (i.e., risk-1, risk-2 and risk-3) in contrast to the two categories of the IPSS (intemediate-2 and high-risk). Patients with the most favorable prognosis according to the FA-IPSS(h) (i.e., risk-1) had a median OS of 23 months (95% CI, 19-29), those with risk-2 had a median survival of 16 months (95% CI, 12-17) and those with the least favorable prognosis (risk-3) had median OS of 10 months (95% CI, 4-13). In contrast, median OS was 20 months (95% CI, 17-24 months) and 13 months (95% CI, 9-16 months) for patients with an IPSS intermediate-2 and high risk scores, respectively. Survival rates at 6 months, one and two years were markedly different amongst the three groups of the FA-IPSS(h). For example, one-year OS for patients with the most favorable prognosis (risk-1) was 80.2% (95% CI, 73.4-87.8) and only 37.5% (95% CI, 23.8-59.1) for those with the poorest prognosis (risk-3). Sensitivity analyses supported our findings. The FA-IPSS(h) index showed very good predictive performance (bootstrap-corrected c-index=0.911).CONCLUSION: The FA-IPSS(h) is a new prognostic index that integrates patient's self-reported fatigue into the well established IPSS index. Its use might enhance physicians' ability to more accurately predict OS in higher-risk MDS. Also, implementation of this index into standard practice might have important implications to elicit a more active patient participation during initial consultations. DisclosuresEfficace:Seattle Genetics: Consultancy; Bristol Myers Squibb: Consultancy; TEVA: Consultancy, Research Funding; Lundbeck: Research Funding. Gaidano:Gilead: Consultancy, Honoraria, Speakers Bureau; Janssen: Consultancy, Honoraria, Speakers Bureau; Novartis: Consultancy, Honoraria, Speakers Bureau; Karyopharm: Consultancy, Honoraria; Roche: Consultancy, Honoraria, Speakers Bureau; Morphosys: Consultancy, Honoraria. Bonnetain:Roche: Consultancy, Honoraria, Research Funding; Novartis: Consultancy, Honoraria, Research Funding; Celgène: Consultancy, Honoraria; Amgen: Consultancy, Honoraria; Janssen: Consultancy, Honoraria; Ipsen: Consultancy, Honoraria; Integragen: Consultancy, Honoraria; Nestle: Consultancy, Honoraria; BMS: Consultancy, Honoraria; Chugai: Consultancy, Honoraria; Merck Serono: Consultancy, Honoraria; Bayer: Consultancy, Honoraria. Fianchi:Novartis: Honoraria; Celgene: Honoraria; Janssen: Honoraria. Breccia:Novartis: Consultancy, Honoraria; Celgene: Honoraria; Bristol Myers Squibb: Honoraria; Ariad: Honoraria; Pfizer: Honoraria. Platzbecker:Celgene Corporation: Honoraria, Research Funding; Janssen-Cilag: Honoraria, Research Funding; Novartis: Honoraria, Research Funding; Amgen: Honoraria, Research Funding; TEVA Pharmaceutical Industries: Honoraria, Research Funding.