A dynamic prognostic model for polycythemia vera long-term outcomes in patients treated with ruxolitinib

Ruxolitinib is widely used in polycythemia vera (PV) following hydroxyurea failure. However, validated response criteria to predict long-term outcomes in ruxolitinib-treated patients are lacking, complicating clinical decision-making. We investigated predictors of event-free survival ([EFS]; including progression to post-PV myelofibrosis, thrombosis, hemorrhages, or death) after 6 months of ruxolitinib in 178 patients with PV enrolled in the observational PV-ARC study. After a median follow-up from ruxolitinib start of 3.50 years, 15 patients died, 7 had a thrombosis, 9 had a hemorrhage, and 21 progressed to myelofibrosis. Overall, 5-year EFS was 70.4%. Leukocytosis, thrombocytosis, phlebotomy need, lack of spleen length reduction ≥50% (SR50), and ruxolitinib dose <10 mg twice daily at 3 time points (baseline and months 3 and 6), were tested for association with EFS. Multivariable analysis identified 3 independent risk factors: (1) ruxolitinib dose <10 mg twice daily at ≥1 time point (hazard ratio [HR], 1.94; P = .047), (2) no SR50 at months 3 and 6 (HR, 2.65; P = .009), (3) phlebotomy requirement at ≥2 time points (HR, 2.11; P = .039). Points were assigned as follows: 1 to phlebotomies at 1 time point; 2 to phlebotomies at ≥2 time points and to ruxolitinib <10 mg twice daily at ≥1 time point; and 2.5 to lack of SR50. Based on cumulative scores, we developed the PV-response to ruxolitinib after 6 months (PV-RR6) prognostic model, identifying 3 risk categories: low (score 0: 5-year EFS, 89.4% [n = 63]), intermediate (score 1-2.5: EFS, 71.0% [n = 82]), and high (score >2.5: EFS, 38.9% [n = 33]). PV-RR6 enables early identification of patients at risk of poor outcomes, supporting timely treatment optimization in ruxolitinib-treated patients. This trial was registered at www.clinicaltrials.gov as NCT06134102.