Durability of spleen response may affect outcome of ruxolitinib-treated patients with myelofibrosis: results from a multicentre study on 284 patients

Background: Ruxolitinib (RUX) is a JAK1/2 inhibitor able to control myelofibrosis (MF)-related splenomegaly, with around 60-80% of patients (pts) achieving a reduction in spleen length by at least 25%. The achievement of a single spleen response was not found to correlate with survival in previous analysis. Also, no data are available on the dynamics of spleen response over time and on the impact of RUX-response feature on outcome. Aims: To report the dynamics of spleen response on outcome of 284 evaluable MF pts treated with RUX according to prescribing information. Methods: A clinical database was created in 23 European Hematology Centres including retrospective data of 462 MF pts treated with RUX from Jan 2011 to July 2017. 284 pts (61.5%) received at least 4 evaluations of spleen length by palpation for a period of at least 12 months and were included in the analysis. Spleen response was defined as reduction of spleen length ³25% by palpation. Overall Survival (OS) was calculated from the start of RUX. Progression-free survival included progression to acute leukemia (AL) and death, whichever came first. Event-free survival (EFS) included also RUX discontinuation for any cause. Results: A total of 284 (52.1% primary, 47.9% post Polycythemia Vera/Essential Thrombocythemia) MF-patients were analyzed. After a median follow-up of 25.8 months from RUX-start, 48 deaths, 18 progressions to AL, and 85 RUX discontinuation were recorded. Overall, 195 pts (68.7%) achieved a spleen response at 6 months. By landmark analysis at 6 months, OS of pts in spleen response at 6 months was comparable to non-responders at that time-point, after adjustment for DIPSS risk (p=0.12), while spleen responders had significantly better PFS and EFS (p=0.02, p=0.005 respectively) (Figure 1a,b,c). Pts were classified into three groups of response: a) pts with stable response (SR) at all evaluations; b) pts with fluctuating results (FR); c) pts with no response (NR). Compared to pts with FR/NR, pts with stable response were less frequently at high DIPSS risk (OR 0.36, 95% CI 0.14-0.94, p=0.04), carried less frequently a large (>10 cm) splenomegaly (OR 0.20, 95% CI 0.09-0.45, p<0.001), and started RUX more frequently <2 yrs from MF diagnosis (OR 0.47, 95% CI 0.26-0.83, p=0.01). Among the three groups of response, we observed 172 (60.6%) SR, 50 (17.6%) FR, and 62 (21.8%) NR. Using a Cox model adjusted for DIPSS category, OS and PFS were significantly worse for pts with NR compared to pts with SR/FR (p=0.04 and p=0.05) (Figure 1d,e). 3y-EFS according to different group of response was 73.2%, 58.3% and 45.2% for SR, FR and NR, respectively (p=0.004) (Figure 1f). Summary and Conclusions: This large study failed to detect a direct correlation between spleen response at a specific time-point (6-month) and OS. Additionally, the absence of a spleen response for a period ≥12 months significantly correlated with worse OS and PFS. Noteworthy, achieving a FR correlated with lower EFS, mainly due to a higher rate of drug discontinuation, but did not hamper OS and PFS compared to stable responders. Overall, the study shows that around 40% of pts did not achieve a stable response to RUX, and points out that the dynamics of response, more than a single evaluation of response, correlates with OS.