Supplementary Figure 8. Impact of TP53 Mutation and Complex Karyotype on Multivariable Coefficients and Collinearity.

Supplementary Figure 17. Comparison of Risk Classification Logic Across Models.

Supplementary Figure 16. Effect of Allo-SCT and IDH1/2 inhibitors on OS.

Supplementary Table 4. Statistical Significance of Risk Reclassification Between Moffitt and Comparator Models.

Supplementary Figure 13. Risk Group Reclassification Patterns Between Moffitt Model and Comparator Models.

Supplementary Table 8. Baseline clinical and genetic data for the study cohort.

Venetoclax (VEN) combined with hypomethylating agents (HMA) improves outcomes for patients with newly diagnosed acute myeloid leukemia (AML) who are ineligible for intensive chemotherapy, yet overall survival (OS) remains variable. We analyzed 506 patients with AML treated with first-line HMA/VEN at Moffitt Cancer Center to develop a genetics-based prognostic model. In multivariate analysis, mutations in TP53, KRAS, JAK2, U2AF1, CBL, and cytogenetic lesions del(7q)/-7, del(17p)/-17/i(17q), del(20q), and MECOM rearrangements predicted inferior OS, whereas IDH1/2 mutations were favorable. A point-based system stratified patients into low-, intermediate-, and high-risk groups with median OS of 54.2, 22.3, and 7.5 months, respectively [P < 0.0001; concordance index (C-index) 0.648]. External validation (n = 126) retained prognostic separation (median OS 24.7, 17.4, and 4.3 months, P = 0.0005; C-index 0.626). Compared with existing HMA/VEN-specific models, our model demonstrated superior low- versus intermediate-risk discrimination (31.9-month separation, P = 0.002; hazard ratio = 0.45, P = 0.003), with a comparable C-index. Our model supports personalized risk stratification for HMA/VEN-treated AML, pending broader validation. SIGNIFICANCE:This study identifies key mutational and cytogenetic markers associated with treatment response and OS in patients with AML receiving HMA/VEN therapy. By integrating these genetic markers, our new prognostic model offers improved risk stratification, guiding personalized treatment strategies for patients ineligible for intensive chemotherapy. See related commentary by Lachowiez and Loghavi, p. 339.

Supplementary Figure 4. Total cycles of HMA/VEN treatment.

Supplementary Figure 5. Frequency of genetic alterations in the Moffitt AML cohort.

Supplementary Table 6. Demographic and Clinical Characteristics of Patients in the Validation cohorts.

Supplementary Table 5. Association Between Risk Upclassification and Adverse Cytogenetics.

Supplementary Figure 1. Study Flow and Model Development Pipeline.

Supplementary Figure 11. Risk Discrimination Between Low/Favorable and Intermediate Risk Groups Across Prognostic Models.

Supplementary Figure 6. OS outcomes based on clinical factors.

Supplementary Figure 7. OS outcomes based on TP53 mutation, allelic status, and VAF.

Supplementary Figure 15. Effect of Genetic Risk, AML Subtype, Prior HMA Exposure, and Allo-SCT on OS.

The World Health Organization and international consensus 2022 classifications have proposed lowering the absolute monocyte count threshold to ≥ 0.5 × 109/L for the diagnosis of chronic myelomonocytic leukemia (CMML). This was based on reports describing patients with oligomonocytic CMML (O-CMML) (0.5 × 109/L to 1.0 × 109/L) as a CMML variant. We identified patients among the myelodysplastic syndrome (MDS) database with O-CMML who meet the new criteria for CMML and compared them to MDS and CMML (≥1.0 × 109/L). Among 1861 patients with MDS, 468 (25%) were O-CMML meeting new CMML criteria. The genomic landscape of O-CMML was more comparable to MDS. Classical CMML somatic mutations (TET-2, ASXL-1 and SRSF2) frequencies were closer to MDS. Proliferative CMML (P-CMML) mutations were less common. DNMT3A and TP53 mutations were more commonly observed in MDS and O-CMML compared to CMML. SF3B1 SM was observed in 29% of O-CMML compared to 16% in MDS, 8% dysplastic CMML (D-CMML) and 5% P-CMML (P < .005). Thirteen patients (2.8%) progressed from O-CMML to CMML. The rate of acute myeloid leukemia transformation was less than MDS and P-CMML. The hazard ratio for overall survival was 1.2 (95% confidence interval [CI], 1.03-1.4; P = .018) for O-CMML, 1.3 (95% CI, 1.07-1.58; P = .008) for D-CMML and 1.87 for P-CMML (95% CI, 1.57-2.4; P = .005) compared to MDS after adjusting for molecular international prognostic scoring system. Although O-CMML may be a unique entity, the current classification does not enrich CMML-like variants by all clinical measures. A comprehensive analysis of clinical, molecular and immunophenotype is needed for better classification.

Supplementary Figure 10. Nested Bootstrap Analysis.

Supplementary Figure 14. Comparison of the Moffitt Prognostic Model with Existing Risk Classification Systems.

6509 Background: MDS/MPNs are complex diseases that exhibit proliferative symptoms and aggressive clinical courses. Treatment options are limited. Ruxolitinib, a JAK1/JAK2 inhibitor, showed clinical benefit in pts with chronic neutrophilic leukemia (CNL) and atypical CML (aCML); yet survival remained poor (mOS 18.8 mo). Fedratinib (Fed) is a JAK2 inhibitor approved for MF that, compared to ruxolitinib, has a broader kinase inhibition profile which may provide enhanced efficacy in molecularly complex disease. Fed potently inhibits FLT3 and BRD4 and potently suppresses c-Myc, a critical upregulated pathway in MDS/MPN. Here, we present the primary results of NCT05177211 assessing the efficacy of Fed in pts with MDS/MPNs. Methods: NCT05177211 is a phase 2, multi-institutional study that enrolled pts with aCML, CNL, MDS/MPN-unclassifiable (MDS/MPN-U), and MDS/MPN-ring sideroblasts and thrombocytosis (MDS/MPN-RS-T) who have splenomegaly (spleen volume > 450cc) and/or significant symptom burden (MPN TSS ≥ 10). The primary endpoint is overall response defined as complete/partial response or clinical benefit at 24 weeks per MDS/MPN IWG response criteria. Exploratory endpoints include progression-free survival (PFS), overall survival (OS), and duration of response (DOR). Fed was given at a dose of 400 mg daily on day 1-28 of 28-day cycles. Results: The data cutoff was 1/1/2026. Among 25 pts, 6 had aCML, 5 had CNL, 6 had MDS/MPN-RS-T, and 8 had MDS/MPN-U. Median age was 68.8 y. ≥3 mutations were present in 19 (76%) pts. Three (12%) pts remain on study. Median duration of treatment was 7.4 months. Eleven (44%) pts responded at week 24: 6 (30%) spleen responses (SVR35) and 9 (47%) symptom responses (TSS50). Four (25%) pts had both spleen and symptom response. Among 16 pts with splenomegaly treated for ≥ 24 weeks, spleen volume decreased by an average of 30% (+9% to -61%). Among 16 pts with symptomatic disease treated for ≥ 24 weeks, TSS improved by an average of -41% (range +35% to -80%). CSF3R mutations were enriched among pts with SVR35 (n = 4; 67%). Median PFS and OS were 36.9 mo (95% CI 19.3-NR) and 36.9 mo (95% CI 18.4-NR), respectively. Among responders, the duration of response was 7.6 mo (95% CI: 3.72-not reached). When stratified by diagnosis, OS for aCML, CNL, MDS/MPN-RS-T, and MDS/MPN-U was 19.8, 36.9, NR and NR, respectively. SAEs occurred in 13 (52%) pts. SAEs possibly related to Fed included acute kidney injury (n = 2), low thiamine (n = 1), colitis (n = 1), diarrhea (n = 1), arthralgias (n = 1), and edema (n = 1). Any grade nausea, vomiting, diarrhea, and constipation were seen in 52%, 16%, 44%, and 48% of pts, respectively. Conclusions: Fed demonstrates promising clinical efficacy in MDS/MPN and CNL pts with proliferative features. The safety profile is consistent with prior experience. Fed’s unique kinase inhibition profile may provide a mechanism for enhanced effectiveness in this pt population. Clinical trial information: NCT05177211 .