Validation of the Revised International Prognostic Scoring System in 2582 Patients with Myelodysplastic Syndrome: A Multicenter Study By the Hellenic MDS Study Group

Introduction - Aims: Several prognostic scoring systems have been developed for patients with myelodysplastic syndromes (MDS), including the International Prognostic System (IPSS), the WHO Prognostic Scoring System (WPSS) and the Revised IPSS (IPSS-R). We evaluated the prognostic value of the IPSS-R on an independent group of 2,582 Greek patients with MDS, registered in the Hellenic National MDS Registry. The aim of this multicenter study was to validate the IPSS-R as a predictor for leukemia-free survival (LFS) and overall survival (OS), in newly-diagnosed MDS patients and to compare its prognostic significance with that of IPSS and WPSS. Moreover, to investigate the predictive value of IPSS-R in association with other recognized prognostic variables, such as patient's age, baseline serum lactate dehydrogenase (LDH), and ferritin concentrations, IPSS, WPSS, Eastern Cooperative Oncology Group (ECOG) performance status, transfusion dependency, and response to first-line treatment. Methods: Clinicopathological data from 2,582 MDS patients, diagnosed between 1/2000 - 1/2015 and registered in the Hellenic National MDS Registry were analyzed. Patients with MDS/MPN were excluded. Data included age, gender, date of diagnosis, clinical characteristics, WHO-2008 classification, laboratory parameters, transfusion dependency, bone marrow aspirate and biopsy morphology, cytogenetic findings, and type of treatment. LFS was calculated from the date of initial diagnosis of MDS until bone marrow blast increased to ≥20% [transformation to acute myeloid leukemia (AML), according to the WHO classification], or last contact. OS was defined as the time from MDS diagnosis to death, or last contact. Patients alive and not having developed AML until last follow-up were censored for OS and LFS, respectively. Kaplan-Meier survival analysis and Cox regression analysis were performed with regard to LFS and OS. Differences between Kaplan-Meier curves were evaluated using the Mantel-Cox (log-rank) test. All significant variables identified by univariate Cox regression analysis and clinical factors important for MDS were used to build the multivariate Cox regression models. Multivariate Cox regression analysis included only those patients for whom the status of all variables was known, and comprised age, serum LDH, and ferritin levels, transfusion dependency, response to first-line treatment, IPSS, WPSS, and IPSS-R. Confidence intervals (CI) were estimated at the 95% level; all tests were two-sided, accepting p<0.05 as indicative of a statistically significant difference. All statistical analyses were performed with the statistical software SPSS (version 21). Results: 1,623 male (62.9%) and 959 female MDS patients with a median age of 74 years at diagnosis were included in the current study. Complete follow-up information was available for 2,376 patients. The estimated median OS was 58 months (95% CI = 52.9 - 63.1 months). For 1,974 patients, data used in the calculation of all three scoring systems were complete, thus allowing risk score calculation and comparison of the three risk assessment systems. Median OS was significantly different in patient subgroups classified according to IPSS, WPSS, and IPSS-R, as shown by the Kaplan-Meier survival analysis (p<0.001). Fig. 1 shows Kaplan-Meier OS curves of MDS patients stratified according to IPSS-R (p<0.001). Moreover, the comparison of the prognostic value of the IPSS, WPSS, and IPSS-R revealed that the IPSS-R was significantly superior to both, WPSS and IPSS (p<0.001 in all cases). Multivariate Cox regression analysis demonstrated that the high prognostic value of IPSS-R, in terms of LFS and OS, was independent of patient's age, serum LDH, and ferritin concentration, ECOG performance status, and transfusion dependency (p<0.001). Interestingly, besides IPSS-R, patient age and transfusion dependency retain their small - yet significant - prognostic impact in the multiparametric models, thus implying that these two parameters could add prognostic value to the IPSS-R. Conclusions: Our data support the notion that all three prognostic scores are very useful predictors for both, LFS and OS in MDS, yet IPSS-R is superior to IPSS and WPSS as a prognostic tool, with regard to OS. Disclosures No relevant conflicts of interest to declare.