Propensity Score Matched Comparison of Dasatinib and Nilotinib As a Frontline Therapy in Newly Diagnosed CML with Chronic Phase

Background: Dasatinib (DAS) and nilotinib (NIL) are standard frontline therapy for chronic myeloid leukemia, chronic phase (CML-CP) based on randomized trials compared to imatinib. However, DAS and NIL have not been compared directly. The purpose of this study is to analyze efficacy, long-term outcome and toxicity of DAS and NIL as a front line therapy in newly diagnosed CML-CP.

Method: Newly diagnosed patients (pts) with CML-CP, who received front-line therapy by either one of the phase II trials conducted almost in parallel (DAS: NCT00254423, N = 102 and NIL: NCT00129740, N = 104) are matched with caliper matching by the propensity score (PS) to adjust pre-treatment confounding factors. DAS was given orally by either 50mg twice daily (N = 30) or 100mg daily (N = 77). NIL was given 400mg orally twice daily. Toxicity was recorded according to the CTCAE ver. 4.0.

Result: PS matching resulted in 87 pts from each trial to be matched for pre-treatment characteristics including age, Sokal score, lab, and organ function (Table). The median observation duration was 50.9 months (95% CI: 40.1-61.7) vs. 43.0 months (95% CI: 35.3-50.7) (DAS vs. NIL, P = 0.56). Response rate at 3, 6, and 12 months as well as cumulative (best) response are shown in Table. There were no significant differences in measures of response throughout the study period except for a higher rate of complete molecular response at 6 months with NIL (NIL vs. DAS, 11% vs. 3%, P = 0.04). However, at 12 months, this difference was not retained; there was also no difference in the rate of optimal response at 3 months. There was no statistical difference in cumulative response between 2 groups. No statistical difference was observed between 2 groups in any of the survival endpoints at 3 years (overall, event-free, failure-free, and transformation-free survival). Treatment discontinuation was observed in 16 (18%) vs. 17 (19%) pts with (DAS vs. NIL, P = 0.82). Reason for the discontinuation was; 1) toxicity (8 vs. 8, P = 1.00), 2) resistance (5 vs. 8, P = 0.39), and 3) financial (4 vs. 1, P = 0.37) (all presented as DAS vs. NIL, respectively). Adverse event (AE) was observed in 40 (46%) vs. 42 (48%) pts (DAS vs. NIL, P = 0.76), whereas grate 3 or more AE was observed in 19 (22%) vs. 15 (17%) pts (DAS vs. NIL, P = 0.44).

Conclusion: In PS matched cohort of newly diagnosed CML-CP pts, the outcome observed with both treatment options (DAS and NIL) is excellent with no clear difference in response or long-term survival endpoints. Incidence of clinically significant AEs was similar between DAS and NIL.