P722: IS THE SECOND-GENERATION TYROSINE KINASE INHIBITOR A BETTER THERAPY THAN IMATINIB FOR PERSONS WITH CHRONIC MYELOID LEUKEMIA PRESENTING IN ACCELERATED-PHASE?

Background: Although most patients with chronic myeloid leukemia (CML) were presenting in the chronic phase at diagnosis, approximately 5% presented with features of accelerated phase. The features of accelerated phase were variably defined by different criteria, such as MD Anderson and European LeukemiaNet (ELN) criteria most often used. Regardless of these criteria used to defined accelerated phase, only scant data were available regarding the prognostic co-variates for outcome. Moreover, evidence that patients could benefit more from the second generation (2G-) tyrosine kinase inhibitor (TKI) than imatinib as initial therapy was still extremely limited. Aims: Explore whether patients with chronic myeloid leukemia (CML) presenting in accelerated-phase could benefit from initial second-generation tyrosine kinase inhibitor (2G-TKI) therapy. Methods: We interrogated data from consecutive patients with CML presenting in accelerated phase at diagnosis by the MD Anderson criteria (n = 325) or ELN criteria (n = 275). Propensity score matching (PSM) analysis was used to compare cytogenetic and molecular responses and outcomes in those receiving initial imatinib versus a 2G-TKI. COX multi-variable regression model was performed to identify the prognostic co-variates. Results: In the cohort defined by MD Anderson criteria, PSM analysis showed that patients receiving initial 2G-TKI (n = 89) had higher probability of achieving a CCyR (p = 0.01), MMR (p = 0.04) and MR4.5 (p = 0.004) than those receiving imatinib (n = 156), however, probabilities of FFS, PFS and survival were similar. Multi-variable analyses showed that lower platelets and higher blasts in blood or bone marrow were significantly-associated with poor FFS, PFS and/or survival (Figure A). In an attempt to determine whether choice of therapy contributed to the differences in responses and outcomes, platelets < 600 × 10E+9/L and blood / bone marrow blasts ≥ 5% were identified as the prognostic categorical co-variates by ROC analyses. The entire cohort were classified into low- (possessing none of these co-variates, n = 57), intermediate (possessing any 1 of these co-variates, n = 127) or high-risk (possessing all of these co-variates, n = 61) subgroups. There were significant differences in probabilities of outcomes among these risk subgroups (all p values < 0.001, Figure B). In each risk subgroup, patients receiving initial 2G-TKI had the similar outcomes to those receiving imatinib except higher incidences of responses including CCyR, MMR and MR4.5. The same results were observed in the cohort defined by ELN criteria. Image:Summary/Conclusion: Initial 2G-TKI lead to similar outcomes as imatinib in patients with CML presenting in accelerated-phase except higher therapy responses.