Validation of the Revised International Prognostic Scoring System (R-IPSS) for Patients with Myelodysplastic Syndromes: Therapeutic Implications.

Abstract

AbstractAbstract 2816▪▪This icon denotes a clinically relevant abstract Background:The International Prognostic Scoring System (IPSS) was recently revised under the auspices of MDS foundation as a collaborative international effort. The proposed R-IPSS is suggested to refine the prognostic value of the IPSS. Instead of the 4 original IPSS categories, 5 categories are proposed by R-IPSS. To validate this prognostic model and examine its utility for therapy decisions, we tested the new risk model in a large external single institution patient cohort. Methods:Data were collected retrospectively from the Moffitt Cancer Center (MCC) MDS database and chart review. The primary objective was to validate the new risk model. The R-IPSS score was calculated as reported. Patients were divided into 5 prognostic categories (very low, low, intermediate, high and very high risk). The Kaplan–Meier method was used to estimate median overall survival. Log rank test was used to compare Kaplan–Meier survival estimates between the groups. Results:The MCC MDS database captured 1157 patients. Complete data was available for 1029 patients to calculate the R-IPSS score. Median age was 68 years, and the most common WHO subtype was RCMD (29%). Two thirds of patients were low/int-1 IPSS risk, and 44% were int-2 or high risk MDAS. (Table-1). Among those, 729 patients (77%) were RBC transfusion dependent (TD), and 264 (26%) had serum ferritin >1000 ng/l. Six hundred eighteen patients (60%) received hypomethylating agent (HMA). The median duration of follow up was 68 months (mo). Median OS according to IPSS risk score was 90 mo (95%CI 75–105), 44 mo (95%CI 39–46), 18 mo (95%CI 15–21), and 14 mo (95%CI 11–17), for low, int-1, int-2, and high risk categories, respectively (p < 0.005). According to MD Anderson risk Score, the median OS was 108 mo (95%CI 91–126), 55 mo (95%CI 50–60), 25 mo (95%CI 22–28), and 14 mo (95%CI 12–16), for low, int-1, int-2, and high risk respectively (p < 0.005).Using the R-IPSS, 106 (10%), 311 (30%), 247 (24%), 201 (20%), and 164 (16%) were classified as very low, low, int, high, and very high risk. The median OS was 82 mo (95% CI 64–100), 57 mo (95% CI 46–68), 41 mo (95% CI 33–49), 24 mo (95% CI 20–28), and 14 mo (95% CI12–16) for each of the corresponding R-IPSS groups (p <0.005). Table-2 summarizes reclassification of each IPSS risk group by R-IPSS and expected OS accordingly.Among those patients who received HMA, the median OS from time of diagnosis was 76 mo, 55 mo, 42 mo, 25 mo, and 16 mo for very low, low, int, high, and very high risk respectively (p < 0.005). A survival benefit for HMA therapy was only statistically significant in patients with very high risk R-IPSS, with a corresponding median OS of 16 mo with HMA versus 7 mo with no HMA (p< 0.005). OS in patients with very high or high R-IPSS who underwent Allogeneic Stem cell transplant (ASCT) was improved compared to corresponding patients who received non-ASCT management. Patients who had very low, low, and int risk R-IPSS had no apparent OS benefit with ASCT. (Table-3). Conclusion:Our data validates the prognostic value of the proposed R-IPSS, but refines prognostic discrimination only for intermediate risk group of IPSS. Both the R-IPSS and IPSS were valid prognostic models for patients treated with HMA. The benefit of ASCT was restricted to patients with high and very high R-IPSS groups. The utility of the R-IPSS as a tool for therapeutic decisions should be further examined before wide adaptation.Table-1N=1029Age> 60 years798 (78%)GenderMale671 (65%)RaceWhite950 (92%)WHORA114 (11%)RARS93 (9%)RCMD303 (29%)Del5q32 (3%)RAEB-1218 (21%)RAEB-2171 (17%)MDS-U26 (3%)CMML23 (2%)MDS/MPN-U45 (4%)AML4 (<1%)IPSSLow179 (17%)Int-1513 (50%)Int-2254 (25%)High83 (8%)MD Anderson ScoreLow181 (18%)Int-1411 (40%)Int-2221 (22%)High213 (20%)Table-2IPSSR-IPSSMedian OS (mo)PLowVery low47 (26%)860.3n=179Low112 (63%)99Int20 (11%)70Int-1Very low59 (12%)800.003N=511Low192 (38%)49Int178 (35%)34High74 (14%)29Very High8 (1%)29Int-2Low7 (3%)370.005N=254Int46 (18%)30High112 (44%)18Very High89 (35%)14HighInt2 (2%)420.06N=83High14 (17%)14Very High67 (81%)14Table-3R-IPSSAllo SCTMedian OS (mo)PVery lowYes (n=4)690.8n=106No (n=102)86LowYes (n=14)420.8N=311No (n=297)57IntermediateYes(n=36)460.4n=246No (n=210)40HighYes (n=44)420.004n=200No (n=156)21Very HighYes (n=33)31<0.005n=164No (n=131)12 Disclosures:No relevant conflicts of interest to declare.