Beyond TP53: Molecular IPSS-genes multihit status revisited

Abstract

Abstract Introduction Molecular profiling has improved risk assessment in myelodysplastic syndromes (MDS), leading to the development of the Molecular International Prognostic Scoring System (IPSS-M). This model incorporates 31 genes but only considers multihit status (MS) for TP53. We aim to explore the prognostic relevance of MS in genes included in IPSS-M among patients from three GESMD centers and to validate the findings in our cohort using the original IPSS-M database. Methods We conducted a multicenter, retrospective study including patients diagnosed with MDS between 2011 and 2024. A Next-generation sequencing (NGS) panel covering 39 myeloid genes—including all IPSS-M genes—was used, with a target depth of x1400. In 81 cases, we applied a 40-gene panel missing only four IPSS-M genes (BCORL1, ETNK1, GNB1, PPM1D). Genes mutated in at least 5% of patients were analyzed. Allelic status was defined following Bernard et al. (Nat Med, 2021). To establish the independent prognostic value, we applied the same two criteria used in the pivotal IPSS-M study : statistically significant impact on leukemia-free survival (LFS) in both univariate and multivariate models (the latter adjusted for age, sex, and therapy-related MDS). Significant results were validated using the public IPSS-M dataset. Results We analyzed 301 patients with a median follow-up of 17 months (IQR, 9–44); 17% progressed to leukemia. The median age was 73 years (IQR 62–81), 46% of whom were female. Median baseline values were hemoglobin 9.6 g/dL, leukocytes 3.8 ×10⁹/L, platelets 120 ×10⁹/L, and bone marrow blasts 4%. According to IPSS-M, 41% of patients were classified as high or very high risk. The genes studied included ASXL1, CBL, DNMT3A, EZH2, RUNX1, SF3B1, SRSF2, STAG2, TET2, TP53, U2AF1, and ZRSR2. Patients with MS in CBL and SRSF2 had significantly shorter LFS compared with single-hit cases (11 vs 81 months for CBL, with a p value of 0,034; 14 vs 55 months for SRSF2, with a p value of 0,041. MS in DNMT3A and STAG2 showed a trend toward worse LFS without reaching significance. In multivariate analysis, CBL retained independent prognostic value, while SRSF2 approached significance. In the international IPSS-M cohort, SRSF2-MS was associated with shorter LFS and remained significant in multivariate models, while the negative impact of CBL-MS was not confirmed. Conclusions MDS patients with multihit mutations in SRSF2 and CBL experience shorter LFS compared with those carrying single-hit alterations. Additionaly, the adverse effect of SRSF2-MS was validated in the IPSS-M cohort. Assessing allelic status in these genes may enhance prognostic stratification in MDS.