Outcomes of hypomethylating agent failure in patients with myelodysplastic syndromes

Abstract Background: Hypomethylating agents (HMA) are the current standard-of-care for patients with myelodysplastic syndromes (MDS). Though HMA improve cytopenia and delay progression of disease, most patients eventually lose response to these agents with dismal survival outcomes. There are no FDA-approved therapeutic options in this setting, thus posing significant treatment challenges after HMA failure (HMAF) and an increased risk of transformation to acute myeloid leukemia (AML). Here, we assess the outcomes of pts with HMAF MDS. Methods: We retrospectively evaluated all patients with MDS seen at a single tertiary cancer center from July 2017 to July 2021 and identified those presented with or later developed HMAF. Patient characteristics, laboratory values, and bone marrow (BM) data, including cytogenetics and mutations, were assessed at the time of HMAF. Survival data was updated in July 2025. Results: Of 700 patients who received HMA, 287 patients (41%) acquired HMAF MDS. The median age was 71 years (range: 21, 89) with 186 males (65%). Patients were at risk for adverse outcomes, with 176/244 patients (72%) higher-risk by IPSS-R, 240 (84%) higher-risk by IPSS-M, 80 (28%) with therapy-related MDS, 106/248 (43%) having high- or very high-risk cytogenetics, and 69/207 (33%) TP53-mutated at the time of HMAF. The median time from diagnosis to HMAF was 11.9 months (range: 0.7, 133.6). With a median follow-up time of 25.5 months (95% confidence interval [CI]: 23.3, 36.2), the median overall survival (mOS) from the time of HMAF was 8.7 months (95% CI: 7.5, 11.0). Both Revised International Prognostic Scoring System (IPSS-R) and Molecular IPSS (IPSS-M) were able to stratify patients (p<0.001). The patients with therapy-related MDS at baseline had inferior mOS than those with de novo MDS (3.7 [95% CI: 2.9, 5.5] vs 12.3 months [95% CI: 9.5, 16.7], p<0.001). By univariable analysis of variables at the time of HMAF, including age, gender, relevant laboratory values, cytogenetic risk score, and mutations present in at least 10 patients, increased BM blasts (p<0.001); thrombocytopenia (p=0.006); higher-risk cytogenetics (p<0.001); the presence of NRAS (p=0.029), PTPN11 (p<0.001), and TP53 mutations (p < 0.001); and the absence of mutations in IDH2 (p=0.014), SF3B1 (p=0.023), and TET2 (p=0.040) predicted worse outcomes. After multivariable analysis, only BM blasts (hazard ratio [HR] 1.03, p=0.002), increased absolute neutrophil count (HR 1.05, p=0.004), and TP53 mutations (HR 3.00, p<0.001) were associated with inferior survival. We further grouped patients by type of HMAF, including transformation to acute myeloid leukemia (AML; n=58 [21%]), blast progression (n=150 [53%]), and worsening cytopenia (n=73 [26%]). When analyzing the 223 patients with HMAF MDS, mOS from HMAF was 9.3 months (95% CI: 8.1, 13.3) for the blast progression subgroup and 18.8 months (95% CI: 11.2, 24.2) for the cytopenia subgroup (p = 0.088). The median lines of therapy received after HMAF was 1 (range: 0, 10). Subsequent frontline treatment was known in 217 patients, including HMA monotherapy (n=16, 7%), HMA with venetoclax (n=50, 23%), AML-like regimens (n=23, 11%), immunotherapy (n=30, 14%), targeted therapy (n=14, 6%), direct allogeneic stem cell transplantation (SCT; n=2, 1%), supportive care alone (n=37, 17%), and other regimens (n=45, 21%). Survival differences were observed based on subsequent therapy, with immunotherapy-based and venetoclax-based regimens having the worst mOS (6.7 months [95% CI: 5.4, 13.3] and 9.0 [95% CI: 6.6, 13.2], respectively) and targeted therapy at 20.0 months (95% CI: 8.1, not estimable [NE]) and SCT at 26.1 months (95% CI: 26.1, NE) with the best outcomes (p=0.025). There were no survival differences with HMA retreatment (p=0.347) or use of venetoclax (p=0.454). Twenty-two patients (9%) underwent SCT after HMAF with improved survival outcomes when compared to those who did not (mOS 30.5 [95% CI: 26.1, NE] vs 9.0 months [95% CI: 8.0, 12.3]; p < 0.001). Progression to AML at the time of or after HMAF was observed in 116 patients (40%) with mOS from AML diagnosis of 3.2 months (95% CI: 2.8, 4.4). Conclusions: HMAF MDS remains a significant therapeutic challenge with poor survival outcomes. SCT and targeted therapies resulted in improved survival. Novel investigative options are urgently needed in this patient population. These data potentially serve as the basis for future clinical trial design.