Background Therapy-related myelodysplastic syndromes (t-MDS) represent a heterogenous group of myeloid neoplasms with distinctive clinical, molecular, and cytogenetic features associated with less favorable clinical outcomes relative to their primary counterpart (p-MDS). The original International prognostic scoring system (IPSS) and its revised version IPSS-R development cohorts did not include t-MDS patients. The Molecular IPSS (IPSS-M) is a newly developed clinical-molecular risk stratification model for myelodysplastic syndromes (MDS) that expands on the original IPSS-R by including data on somatic oncogenic mutations. Here we aimed to assess the validity of this tool in the risk stratification of t-MDS using a large cohort of patients treated at a tertiary referral center. Methods A total of 475 t-MDS patients with annotated clinical and molecular data treated at Moffitt Cancer Center were analyzed. For estimation of the IPSS-M score, R version 4.2.1 was used to run batch-calculation using syntax code provided by the Papaemmanuil lab at MSKCC. The mean IPSS-M score was used as reference. Time-to-event analyses were estimated using the Kaplan-Meier method and groups were compared by the log-rank test. We used Cox proportional hazards models for survival endpoints. Results We identified at least 737 driver point mutations involving up to 98 genes across 475 patients. We identified at least one gene mutation in 81% of patients (n=389), and 2 or more in 43% (n=206). Median overall survival (OS) was 1.5 years (95%CI, 1.3-1.8). Median leukemia-free survival (LFS) was 1.1 years (0.9-1.3). Median follow-up was 3.9 years (2.9-4.8). 475 patients (100%) were classified. The most prevalent mutations were TP53 (n=182, 38%), TET2 (n=79, 17%), DNMT3A (n=67, 14%), SF3B1 (n=43, 9%) and U2AF1 (n=35, 7%). Using the IPSS-M risk stratification schema, pts were classified as Very Low (2%, n=8), Low (13%, n=63), Moderate Low (14%, n=64), Moderate High (10%, n=48), High (22%, n=105) and Very High (39%, n=187) (Fig 1). The IPSS-M categories showed significant separation across all examined endpoints. Median LFS was NR, 3.7, 2.1, 1.7, 1.1 and 0.5 years from Very Low to Very High-risk categories (Fig 1). Conversely, median OS was NR, 4.4, 2.9, 3.5, 1.5 and 0.9 years from VL to VH subgroups respectively (Fig 1). A five-to-five mapping between the IPSS-R and IPSS-M risk subgroups (by merging moderate-low (ML) and moderate-high (MH) into moderate), resulted in the restratification of 45% of patients (n=209) (Fig 2). More than 3/4 of patients classified as IPSS-R very-low (VL) were upstaged (majority to IPSS-M low). More than half of patients classified as IPSS-R intermediate shifted: 50% (n=38) were reclassified as IPSS-M high/very high (H/VH). Similarly, 40% of patients classified as IPSS-R high were upstaged to IPSS-M VH. More than 20% of VH risk patients were downstaged (majority to high risk). Most significant was the reclassification of 9.4% (n=44) of the cohort from IPSS-R lower and intermediate risk categories into IPSS-M higher risk strata. Contextually, the restratification of patients from IPSS-R intermediate to higher risk strata in IPSS-M was associated with notable differences in survival outcomes which were as poor as 0.9 yrs and as high as 4.4 yrs for those reclassified as LR (relative to 2.1 yrs). There was significance across all IPSS-M risk strata between t-MDS and p-MDS. Stage to stage t-MDS was associated with worse outcomes compared to p-MDS, except for the very low risk IPSS-M subgroup. Conclusion To our knowledge this is the first external validation of IPSS-M in a large cohort of t-MDS patients. Use of the IPSS-M model for the risk stratification of patients with t-MDS led to reclassification of almost half of MDS patients and refined prognosis estimates. In general, outcomes of t-MDS are worse than p-MDS even when compared stage to stage using IPSS-M. The model can clearly discriminate patients with more indolent or aggressive disease course compared to the IPSS-R and better guide treatment decisions. Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal
Read full abstract