Long-Term Efficacy and Safety in COMFORT-II, a Phase 3 Study Comparing Ruxolitinib with Best Available Therapy for the Treatment of Myelofibrosis: 5-Year Final Study Results

BACKGROUND: Ruxolitinib (RUX) is a potent JAK1/JAK2 inhibitor that has demonstrated rapid and durable improvements in splenomegaly and symptoms as well as improved survival in the two phase 3 COMFORT studies in patients (pts) with myelofibrosis (MF). In COMFORT-II, significantly more pts achieved the primary endpoint (a ≥ 35% decrease in spleen volume from baseline at wk 48) with RUX compared with best available therapy (BAT) (28% vs 0%; P ˂ .0001). The 3-year follow-up confirmed that spleen volume reductions were sustained and RUX treatment remained tolerable with long-term use. Here, we report final study results on longer-term safety and efficacy after 5 years of RUX treatment in COMFORT-II.

METHODS: COMFORT-II is a randomized (2:1), open-label phase 3 study of RUX vs BAT in pts with intermediate-2– or high-risk primary MF, post-PV MF, or post-ET MF. Pts initially received RUX 15 or 20 mg bid based on their platelet counts at baseline (100-200 and > 200 x 10⁹/L, respectively), and doses were individually titrated to maximize safety and efficacy. Pts were allowed to cross over from the BAT arm to receive RUX upon protocol-defined progression (primarily progressive splenomegaly, a ≥ 25% increase in spleen volume from on-study nadir). All pts randomized to BAT had crossed over or discontinued by Nov 2011. The date of final database lock for the study is 20 Apr 2015.

RESULTS: Pts were randomized to RUX (n = 146) or BAT (n = 73). Baseline characteristics were well balanced between arms and have been described previously (Harrison, N Engl J Med, 2012); disease and hematologic characteristics were representative of a population of pts with advanced primary or secondary MF. At study completion (median follow-up, 4.3 years), 39 pts (26.7%) in the RUX arm and 11 of the 45 pts (24.4%) who crossed over from BAT completed 5 years of on-study treatment. Primary reasons for premature discontinuation before 5 years were adverse events (AEs; 24.0%) and disease progression (21.9%) in the RUX arm and withdrawal of consent and other in the BAT arm (12.3% each).

Overall 78 pts (53.4%) in the RUX arm achieved a ≥ 35% reduction in spleen volume from baseline at any time during treatment; the median duration of maintenance of spleen volume reduction was 3.2 years. The K-M estimated probability of maintaining this reduction was 0.51 (95% CI, 0.38-0.62) at 3 years and 0.48 (95% CI, 0.35-0.60) at 5 years. Approximately one-third of evaluable JAK2 V617F-positive pts had a ˃ 20% reduction in allele burden at 3.2 years (38.3%) and 3.7 years (31.0%). With RUX treatment, 23 pts (15.8%) had improved fibrosis (including 4 who improved to grade 0 from baseline fibrosis grades of 1 [n = 1], 2 [n = 2], and 3 [n = 1]), 47 pts (32.2%) had stable fibrosis, and 27 (18.5%) had a worsening at their last assessment.

There was no relevant increase in the incidence of AEs with longer exposure (median: RUX arm, 2.6 years; BAT arm, 0.87 years; RUX after crossover, 1.2 years) compared with previous reports. The most commonly reported AEs in pts who received RUX any time (randomized treatment, extension phase or after cross over from BAT) were thrombocytopenia (52.4%), anemia (49.2%), diarrhea (35.6%), and peripheral edema (33.0%); grade 3/4 AEs included anemia (22.5%), thrombocytopenia (15.2%), pneumonia (5.8%), general physical health deterioration (4.2%), and dyspnea (4.2%). 8 pts (5.5%) and 5 pts (6.8%) developed leukemia in the RUX and BAT arms, respectively. There were no new or unexpected AEs.

Overall, 59 (40.4%) and 35 (47.9%) deaths were reported in the RUX and BAT arms, respectively. Median OS was not reached in the RUX arm and was 4.1 years in the BAT arm. There was a 33% reduction in risk of death with RUX compared with BAT (HR, 0.67; 95% CI, 0.44-1.02; P = .06). The K-M estimated probability of survival at 5 years was 56% with RUX and 44% with BAT. As expected, the confounding effect on OS of crossover from BAT to RUX became apparent in this extended follow up compared with previous analyses; an analysis of OS correcting for crossover will be presented.

SUMMARY/CONCLUSIONS: The immediate benefits of RUX treatment, such as improvements in spleen size, were maintained with long-term therapy. The previously reported OS benefit was maintained, although results are confounded by extensive crossover from the BAT arm following the primary analysis at wk 48, which becomes more apparent with longer follow-up. Long term safety and tolerability was consistent with previous findings.