Prognostic impact of persistent and treatment-emergent cytopenia in 879 myelofibrosis patients treated with ruxolitinib: The “RUX-MF” study

Abstract

Abstract Introduction Cytopenia, such as anemia and thrombocytopenia, are associated with inferior overall survival (OS) in patients (pts) with myelofibrosis (MF). Ruxolitinib (RUX), a JAK1/2 inhibitor, remains the standard of care for symptomatic, higher-risk MF, but it does not improve cytopenia and may worsen them. Previous studies have shown that both baseline cytopenia and the development or worsening of anemia during the first 12 weeks of RUX negatively impact patient outcomes (Palandri F, et al. Cancer 2023; Kuykendall et al, ASH 2024). However, the 12-week timepoint may be too early to guide clinical decisions. Also, hemoglobin (Hb) improvement during the first 24 weeks of momelotinib therapy was linked to prolonged OS (Palandri F. et al, EHA 2025). Here, we evaluated the trajectory of cytopenia during the first 6 months of RUX therapy, aiming to investigate the prognostic impact of: (1) treatment-emergent (TE) and persistent cytopenia; (2) TE anemia versus TE thrombocytopenia; and (3) Hb improvement in patients with baseline anemia. Methods This sub-analysis of the RUX-MF study (NCT06516406) included 879 patients treated with RUX for ≥6 months. Anemia and thrombocytopenia were defined as Hb <10 g/dL and platelet count (PLT) <100 x10⁹/L, respectively. Patients were categorized into four groups: (1) no cytopenia at baseline (BL) or 6 months (never cytopenia); (2) no cytopenia at BL but development of ≥1 cytopenia during follow-up (TE cytopenia); (3) cytopenia present at both BL and 6 months (persistent cytopenia); (4) baseline anemia with Hb >10 g/dL at 6 months, without concurrent BL/TE thrombocytopenia (improved anemia). OS was calculated from RUX start to death, last contact, or transplant, using the Kaplan-Meier method; hazard ratios (HRs) were estimated via Cox regression. Results At RUX start, 357 pts (40.6%) had anemia (n=301), thrombocytopenia (n=32), or both (n=24). These patients were older, had a higher prevalence of primary MF, higher risk scores, elevated peripheral blasts, and more severe symptoms (all p<0.001) compared to myeloproliferative pts. Median OS was significantly shorter in cytopenic versus myeloproliferative pts (3.7 vs. 6.7 years; HR 2.11, 95% CI 1.75–2.54; p<0.001). After 6 months of RUX, 317 pts (36.1%) remained cytopenia-free; 273 pts (31.1%) developed ≥1 TE cytopenia (205 previously myeloproliferative, 68 with additional cytopenia); 235 pts (26.7%) had persistent cytopenia; 54 pts (6.1%) demonstrated improved anemia. Compared to never cytopenia pts (median OS 8.07 yrs), median OS of TE cytopenia pts was 5.11 yrs (HR 2.01, 95% CI 1.56–2.58; p<0.001), and 3.68 yrs for persistent cytopenia pts (HR 2.62, 95% CI 2.04–3.35; p<0.001). These results remained significant after adjusting for baseline DIPSS/MYSEC-PM risk (p=0.003). Within the TE cytopenia group, pts developing anemia alone (n=149) had a median OS of 6.15 years, while those with isolated thrombocytopenia (n=94) had a shorter median OS (4.33 years; HR 1.65, 95% CI 1.15–2.34; p=0.006), similar to pts developing both cytopenia (n=30, median OS 5.11 years; p=0.74). Compared with never cytopenia pts, TE anemia conferred an intermediate prognosis (HR 1.63, 95% CI 1.19–2.22; p=0.002), while TE thrombocytopenia (HR 2.51, 95% CI 1.82–3.47; p<0.001) and concurrent TE anemia and thrombocytopenia (HR 2.17, 95% CI 1.34–3.51; p=0.002) were associated with the poorest outcomes. Patients with baseline and persistent anemia only had shorter OS compared to those with baseline anemia who achieved Hb >10 g/dL at 6 months (3.47 vs. 5.18 years; HR 1.46, 95% CI 1.01–2.12; p=0.043). Conclusions Among 879 MF pts treated with RUX, the proportion with cytopenia increased from 40.6% at baseline to 57.8% at 6 months. Both persistent and TE cytopenia—occurring in 26.7% and 31.1% of pts, respectively—doubled the risk of death. Recovery from anemia without thrombocytopenia was uncommon but correlated with improved survival. TE thrombocytopenia emerged as the most adverse prognostic factor and may warrant inclusion in future dynamic risk models during JAK inhibitors therapy.Patients who remained cytopenia-free had the best survival, representing the most favorable setting for continuous RUX therapy. Achieving Hb >10 g/dL was validated as a favorable prognostic marker during JAK inhibitors therapy. These results underscore the prognostic heterogeneity of cytopenia during RUX therapy and the need for individualized therapeutic strategies.