Long-Term Effects of Ruxolitinib on Bone Marrow Morphology in Patients With Myelofibrosis and Comparison to Best Available Therapy

Abstract

Context: Ruxolitinib is a JAK1/JAK2 inhibitor that reduced splenomegaly and symptoms and was associated with improved survival in patients with myelofibrosis relative to placebo or best available therapy (BAT) in phase 3 studies. Objective: To evaluate the effects of long-term ruxolitinib treatment on bone marrow (BM) morphology in patients with myelofibrosis. Design and Patients: Trephine biopsies were obtained at baseline, 24 (68 patients), and 48 (18 patients) months from ruxolitinib-treated patients enrolled in a phase I/II trial (NCT00509899; Verstovsek, NEJM 2010). Two of the authors (JT, HMK) independently quantified BM fibrosis and a third (CB-R) reviewed the findings. Reviewers were blinded to patient characteristics and outcomes. BM fibrosis grading was determined in prospectively collected specimens from 139 patients (160 biopsies) treated for 24 (97 patients) and 48 (63 patients) months receiving BAT from an independent, multicenter, observational database. BAT included hydroxyurea (47%), interferon-alpha (7%), assorted sequential therapies (25%), or supportive-only therapy (21%). Outcome Measure: BM fibrosiswasgraded per the World Health Organization (WHO) criteria (scale: 0-3). Changes in BM fibrosis grade versus baseline were categorized as improvement, stabilization, or worsening. Results: At baseline, 21% of ruxolitinib-treated patients had grade 1, 53% grade 2, and 26% grade3BM fibrosis;approximately50%ofpatientshadosteosclerosis at baseline (grade 1: 33%; grade 2: 9%; grade 3: 9%). Collagen accumulation was observed in 32 cases (47%): 30% grade 1, 18% grade 2/3. There were no differences in the distribution of baseline BM fibrosis grades between ruxolitinib and BAT (p¼0.441; CochraneManteleHaenszel test). A greater percentage of ruxolitinib-treated versus BAT-treated patients showed stabilization (24 months: 57% versus 54%; 48 months: 56% versus 46%) or improvement (24 months: 15% versus 6%; 48 months: 22% versus 2%) in BM fibrosis. Worsening (excluding patients with baseline grade 3 BM fibrosis) was more prevalent with BAT (24 months: 37% versus62%;OR[95%CI]:0.40[0.18,0.87],48months:25%versus 75%; OR [95% CI]: 0.11[0.03, 0.43]). Conclusions: Long-term ruxolitinib therapy may meaningfully delay BM fibrosis progression. A comparable effect was not seen with long-term BAT, which supports earlier observations with hydroxyurea (Kvasnicka ASCO