Treatment-free remission (TFR) in patients with chronic myeloid leukemia (CML) following the discontinuation of tyrosine kinase inhibitors.

Abstract

7050 Background: Tyrosine kinase inhibitors (TKIs) discontinuation in patients (pts) with CML is increasingly considered. We evaluated the outcome of pts with CML who discontinued TKIs and determined the factors associated with differences in the success rates of TFR. Methods: We reviewed data from 284 pts with CML treated with TKIs at our institution between October 1999 and February 2017 and who subsequently discontinued therapy. Major molecular response (MMR) was defined as a BCR-ABL1/ABL1 transcripts ratio ≤0.1% as determined by real time (RT)-PCR, MR4 as a ratio ≤0.01% IS, and MR4.5 as a ratio ≤0.0032%. TFR failure was defined as the loss of MMR on any single test. We analyzed TFR rates according to duration and depth of response and conducted a multivariate analysis for factors associated with loss of MMR. Results: Median age was 63 years (range, 25-93). 199 pts (70%) had electively discontinued their TKI while 70 pts (24%) stopped therapy because of adverse events. 92 pts (32%) had switched ≥1 TKI prior to discontinuation due to drug intolerance or resistance. The median time from the initiation of frontline TKI to discontinuation was 117 months (range, 16-242). The median duration of MR4 and MR4.5 before TKI discontinuation was 74 months (range, 2-207) and 64 months (range, 0-207), respectively. At a median follow-up of 36 months (95% CI, 32-40) after TKI discontinuation, 53 pts (19%) lost MMR, translating into a 5-year TFR rate of 79%. 50 pts (94%) resumed TKI therapy and, among 47 evaluable pts, all but one pt regained MMR, with 41 pts (88%) achieving MR4.5. The estimated 5-year TFR rates were 91%, 76% and 70% in pts achieving MR4.5 for ≥6 years, between 5 and 6 years, and < 5 years, respectively (P < 0.0001). The estimated 5-year TFR rates were higher with MR4 and MR4.5 ≥5 years, compared with MR4 < 5 years (87% vs 92% vs 64%; P < 0.0001). Pts who remained on their frontline TKI at the time of discontinuation had a 5-year TFR rate of 82%, compared with 75% and 72% among pts who switched to a second line TKI or beyond because of intolerance or resistance, respectively (P = 0.417). TFR rates did not vary according to the type of frontline TKI used (P = 0.761). By multivariate analysis, only durations in MR4 or MR4.5 ≥5 years before stopping treatment were associated with a lower risk of loss of MMR, with hazard ratios of 0.37 (95% CI, 0.18-0.76; P = 0.007) and 0.20 (95% CI, 0.09-0.45; P < 0.0001), respectively. We evaluated the impact of the frequency of molecular monitoring on the success rate of TFR. The estimated 5-year TFR rate was 79% for pts monitored monthly compared with 85% for pts monitored every 6-8 weeks following discontinuation (P = 0.263). Conclusions: Our findings suggest that achieving MR4 for ≥5 years was associated with a very high probability of maintaining TFR, and that less frequent molecular monitoring could be more cost-effective without any negative impact on outcomes.