Biopsy-free score for screening myelodysplastic syndrome and related diseases in cytopenic patients

Abstract

Myelodysplastic syndromes (MDS) often present with cytopenia, requiring bone marrow exams for diagnosis, which can be subjective and challenging. This study aimed to develop a non-invasive tool integrating clinical, laboratory, and genetic data to estimate MDS probability in Asian patients with unexplained cytopenia. Data were collected from three University hospitals in Thailand (2018–2023). Participants were adults with unexplained cytopenia. Next Generation Sequencing targeted a 40-gene myeloid panel. The predictive model was validated using bootstrapping. The study enrolled 285 participants, including 210 diagnosed with MDS (MDS: 144, MDS/MPN: 32, AML-MR: 34) and 75 non-MDS controls. The MDS group exhibited more severe anemia (Hb 8.7 vs. 10.2 g/dL; p < 0.001) and lower neutrophil count (1.8 vs. 1.9 × 109/L; p = 0.02) compared to non-MDS. Higher red cell distribution width (RDW ≥ 15%) and ≥ 2 mutations were more common in MDS (75% vs. 44% and 46% vs. 2.7%; p < 0.001). The most common gene mutations were TET2 (20.0%) , SRSF2 (16.7%) , ASXL1 (15.2%) , RUNX1 (11.9%) , and DNMT3A (10.5%) . A univariate analysis showed that Hb < 10 g/dL, ANC<1 × 109/L, platelet<100 × 109/L, RDW ≥ 15%, ≥ 2 mutations, one of six gene mutations ( TET2 , ASXL1 , SRSF2 , RUNX1 , TP53 , SF3B1 ), high-risk mutations ( NRAS , NPM1 , IDH2 , SETBP1 , CEBPA , JAK2 , IDH1 , FLT3 , KRAS ), and other risk mutations ( BCOR , U2AF1 , STAG2 , ZRSR2 , PTPN11 ) were significant predictors. The model containing 13 factors demonstrated an ROC curve area of 0.86. At a probability score of ≥ 45%, sensitivity, specificity, positive predictive value (PPV), and positive likelihood ratio (LR+) for MDS were 90.5%, 42.7%, 81.5%, and 1.6, respectively. A web-based calculator is available at ( https://tsh.or.th/mds/calculator/ ). A biopsy-free score integrating complete blood counts and gene mutations may help predict the probability of MDS and related diseases in patients with cytopenia. At a predicted probability of ≥ 45%, the score may help prioritize patients for further diagnostic evaluation, including bone marrow examination.