Extended validation of the RR6 prognostic model at 6 and 12 months in primary and secondary myelofibrosis patients treated with ruxolitinib

Abstract Introduction Ruxolitinib has improved outcomes in both primary and secondary myelofibrosis (MF), though resistance or intolerance often develop, adversely affecting overall survival (OS). Existing prognostic scores such as DIPSS and MYSEC-PM are not tailored for patients undergoing ruxolitinib therapy. The RR6 model, introduced in 2022, was developed to identify early treatment failure after six months of ruxolitinib. This study aims to validate RR6 in an independent cohort of patients with MF and assess its prognostic performance also at 12 months. Methods We conducted a retrospective, single-center study on patients diagnosed with primary or secondary MF (post-PV or post-ET) per WHO/ICC 2022 criteria and treated with ruxolitinib for ≥6 months at the Hematology Clinic of Ancona (2011–2024). Demographic, clinical, laboratory, molecular, and cytogenetic data were collected at baseline, therapy start, and at 3, 6, and 12 months. RR6 was calculated using the online tool (https://rr6.eu/), which integrates ruxolitinib dose (<20 mg BID), spleen size reduction ≤30%, and transfusion dependence at each timepoint. Prognostic performance was compared to DIPSS (PMF) and MYSEC-PM (SMF) using Kaplan-Meier estimates, Cox regression, and concordance index (C-index). Results A total of 128 patients were enrolled (52.3% male; median age 70 years [range 23–92]); 49.2% had PMF and 50.8% had SMF (25.8% post-PV, 25% post-ET). JAK2 V617F mutations were found in 72.7% of cases, CALR in 11.7%, MPL in 1.6%, and 14% were triple-negative. High-risk mutations (e.g., ASXL1, U2AF1, IDH2) were detected in a minority by next-generation sequencing. Initial ruxolitinib dosing was most commonly 20 mg BID (42.2%). For prognostic evaluation, 109 patients with ≥6 months of follow-up were included (median follow-up: 32.6 months). Based on RR6 at 6 months, patients were classified as low-risk (16.5%), intermediate-risk (41.3%), and high-risk (50.5%). OS was significantly worse in the high-risk group (p < 0.01). RR6 showed good correlation with DIPSS and MYSEC-PM but was superior in early identification of poor responders. When applied at 12 months, RR6 remained prognostically relevant, especially in further stratifying intermediate-risk patients. On univariate Cox analysis, RR6 score, age >65 years, baseline anemia (Hb <10 g/dL), persistent transfusion dependence, and absence of spleen response at 12 months were independently associated with inferior OS. The RR6 model demonstrated strong discrimination power via C-index analysis. Conclusions Our study confirms the prognostic value of the RR6 model in a real-world cohort of patients with PMF and SMF treated with ruxolitinib. RR6 enables early identification of high-risk patients after six months and retains its relevance at 12 months, supporting timely clinical decisions regarding therapy modification or transplant referral. Its use may enhance personalized treatment strategies in MF.