Predictive factors for successful treatment-free remission in patients with chronic myeloid leukemia: A pooled analysis

Abstract

Abstract Introduction Treatment-free remission (TFR) has become a new and important goal of therapy for patients (pts) with chronic myeloid leukemia (CML). We aim in this analysis to evaluate predictive factors for successful TFR. Methods We retrospectively reviewed pts with CML in chronic phase (CP) or accelerated phase (AP, defined by clonal evolution only) who discontinued tyrosine kinase inhibitor (TKI) therapy at MD Anderson Cancer Center, Texas, USA (n=226) and Léon Berard Center, Lyon, France (n=183). We analyzed the duration of therapy and sustained (≥12 months) deep molecular remission (DMR; BCR::ABL1 transcripts ≤0.01% on the International Scale [IS]) before TKI discontinuation (DC) and their association with molecular relapse-free survival (MRFS). Molecular relapse was defined as the loss of major molecular response (MMR; BCR::ABL1 transcripts >0.1% [IS]) at any time following DC. MRFS was defined as the duration between DC and molecular relapse, death, or last follow-up. Univariate and multivariate analyses were performed to analyze factors associated with MRFS. Results A total of 409 pts were included, 399 in CP (98%); 222 were males (54%). Sokal score was low in 203 pts (50%), intermediate in 150 (37%), and high in 47 (11%); it was not available for 9 pts (2%). The median age was 51 years (range, 15-80) at diagnosis and 61 years (range, 25-92) at DC. First-line TKI was imatinib in 196 pts (48%), nilotinib in 100 (24%), dasatinib in 93 (23%), ponatinib in 16 (4%), and bosutinib in 4 (1%). 309 pts electively discontinued TKI (76%), 72 (18%) discontinued therapy due to adverse events, 7 (2%) for pregnancy, 7 (2%) for financial reasons, 6 (1%) due to the diagnosis of another malignancy, and 6 (1%) for other reasons. The median number of prior therapies received prior to DC was 1 (range, 1-5). Imatinib was the last TKI used at the time of DC in 142 pts (35%), followed by dasatinib in 132 (32%), nilotinib in 121 (30%), bosutinib in 11 (2%), and ponatinib in 3 (1%). At the time of DC, 305 pts (75%) were on the same type of TKI used as first-line therapy. The median duration of therapy before DC was 97 months (range, 27-261). The median duration of sustained MR4 before DC was 60.4 months (range, 12.2-239): <2 years in 22 pts (5%), 2-5 years in 179 (44%), and >5 years in 208 (51%). The median duration of sustained MR4.5 before DC was 43.8 months (range, 12.03-239). Among 359 pts who achieved MR4.5 before DC, 77 (21%) had MR4.5 duration <2 years, 153 (43%) between 2-5 years, and 129 (36%) >5 years. After a median follow-up of 63 months after DC, 147 pts (36%) lost MMR, 94 (64%) within 6 months, 31 (21%) between 6-12 months, and 22 (15%) after 12 months. This translated into a 5-year MRFS rate of 63% overall. All pts achieved sustained MR4 or deeper before DC; their 5-year MRFS rates were 37%, 55%, and 73% for MR4 duration of <2 years, 2-5 years, and >5 years, respectively (p<0.001). 359 of 409 pts achieved sustained MR4.5 before DC; their 5-year MRFS rates were 53%, 54%, and 84% for MR4.5 duration of <2 years, 2-5 years, and >5 years, respectively (p<0.001). We then compared the outcome of pts who received imatinib (n=110), dasatinib (n=83), or nilotinib (n=72), as the only TKI before DC. The 5-year MRFS rates were 65%, 58%, and 77%, respectively (p=0.086 overall and p=0.031 after excluding imatinib). The 5-year MRFS was similar with imatinib (65%) compared to second-generation TKIs, 65% and 67%, respectively (p=0.84). By multivariate analysis, a longer duration of TKI therapy before DC (HR, 0.99; p=0.031) and a duration of sustained MR4 or deeper of >5 years (HR, 0.41; p=0.007) were independently associated with TFR. The type of TKI therapy was not associated with TFR by univariate analysis. Out of the 147 pts who lost MMR, 142 resumed TKI therapy; of them, 138 (97%) regained MMR or deeper response after a median of 2.9 months (range, 0.4-123). Three pts did not regain MMR, and one was lost to follow-up. After a median follow-up of 15 years from the time of diagnosis, the 15-year OS rate was 92% with no deaths attributed to CML. Conclusion Treatment DC is a safe approach in pts with CML who achieve a prolonged DMR, with no increase in the risk of disease progression or CML-related deaths. A sustained duration of DMR of more than 5 years was associated with the highest rates of MRFS and TFR.