Abstract
TK and JPB Co first; RK and AMZ are Co senior authors. Introduction The IPSSM incorporates molecular features to improve outcome prediction for pts with MDS. However, IPSSM requires assessment of 31 genetic mutations to classify pts into 6 risk groups. Importantly, IPSSM accounts for missing data and provides best, average, and worst IPSSM scores. Here, we used a large cohort of pts with MDS to analyze the performance of IPSSM if molecular data were missing to better understand the performance of the tool in this frequently encountered clinical setting. Methods For this study, we combined data from a publicly available dataset (Kewan et al. 2023 Nat com) with our multicenter VALIDATE database to increase sample size. Only pts with available molecular data were included and IPSSM was calculated twice: a. using TP53 mutation ( TP53 MT) only and assuming that all other molecular mutations are missing (missing molecular data [IPSSM MM]) and b. using the full molecular panel (available molecular data [IPSSM AM]). TP53 MT were included in all IPSSM calculations given frequency and significant impact on survival. Time-to-event analysis from the time of diagnosis was conducted using Kaplan-Meier estimator. The performance of different scores was evaluated by Harrell's c-index. This study was supported by an independent research grant from AbbVie. Results Of 2,789 pts, 2,489 had molecular data and were included. Median age was 72 years (IQR: 65-78). MDS with excess blast 1/2 (42%) was the most common subtype. Overall, 39% of pts treated with HMA and 15% of underwent transplantation. In total, 16% of pts had complex karyotypes. The most prevalent mutations were SF3B1 (21%), ASXL1 (20%), TP53 (15%), SRSF2 (15%), DNMT3A (12%), and RUNX1 (10%), 39% of the pts had >1 gene mutation. Based on IPSSR, pts were classified as very low (13%), low (47%), intermediate (13%), high (13%), and very high (14%) risk. First, we calculated IPSSM score assuming that all molecular data were missing (IPSSM MM) except TP53 MT. Accordingly, pts were classified as very low (4%), low (33%), moderate low (19%), moderate high (14%), high (17%), and very high (14%) risk. IPSSM MM resulted in the re-stratification of 1623 (66%) pts of which 1104 (68%) pts were up-staged, and 519 (32%) pts were down-staged. When applying actual molecular data (IPSS-M AM), pts were classified as very low (14%), low (30%), moderate low (12%), moderate high (10%), high (17%), and very high (17%) risk. When compared with IPSSR, IPSSM AM resulted in re-stratification of 1218 pts (50%) of which 868 pts (71%) were upstaged ( Panel A). When comparing IPSSM MM with IPSSM AM, 1243 pts (51%) were assigned to the same risk group and 1186 pts (49%) were assigned to different risk groups. Amongst reclassified pts, 431 (36%) were upstaged and 755 (64%) were down staged. Only 109 pts (5%) were reclassified with more than one shift. Median follow-up time was 24 (IQR:10-77) months (mo) with a median overall survival (OS) of 50 mo (95% CI: 47-55). The probability of OS based on IPSSM MM and IPSSM AM risk groups was significantly different (p-value: <0.0001 for both), Panel B. Median OS (mo) based on IPSSM MM vs. IPSSM AM were as follows: very low (not reached vs. 125), low (103 vs. 82), moderate low (63 vs. 58), moderate high (46 vs. 43), high (27 vs. 24), and very high (13 vs.15). Median OS based on IPSSR were: very low (not reached), low (78), intermediate (36), high (24), and very high (14). The median LFS (n=1100 pts) based on IPSSM MM vs. IPSSM AM were as follows: very low (41 vs. 78), low (63 vs. 54), moderate low (39 vs. 33), moderate high (26 vs. 29), high (18 vs. 15), and very high (11 vs. 13). IPSSM MM showed comparable performance to IPSSM AM with c-index (95%CI): 0.713 (0.697 -0.728) vs. 0.714 (0.699-0.729) for OS and 0.645 (0.622-0.669) vs. 0.623 (0.599-0.647) for LFS. When used as continuous scores, IPSSM MM continues to be comparable to IPSSM AM for OS (c-index:0.721 vs. 0.730) and LFS (c-index:0.655 vs. 0.641). IPSS-R had a lower c-index for OS (0.688) and LFS (0.622). Conclusions Our study supports prognostic and clinical value of IPSSM even if most molecular data are missing, confirming that IPSSM adjusts well for missing molecular data and can be used in clinical practice if molecular data are largely missing. While molecular testing remains optimal for accurate risk stratification in MDS, our data suggest that clinical, pathological and cytogenetic data continue to be the main determinant of outcome prediction for pts with MDS.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.