INTRODUCTION:Mutation of the FMS-like tyrosine kinase 3 ( FLT3) gene is one of the most common genetic alterations in acute myeloid leukemia (AML), while FLT3 internal tandem duplication (ITD) is a predictor of poor disease prognosis in patients. Timing of FLT3 mutation testing and the pattern of FLT3 clonal evolution are critical issues for ensuring that patients with FLT3 mutation-positive AML receive targeted therapy promptly, but data from real-world clinical practice are limited.AIM:This retrospective cohort study aimed to assess the turnaround time of FLT3 testing in two patient cohorts with non-M3 primary AML: newly diagnosed (ND) patients and those with relapsed or refractory disease (R/R). A further aim in the R/R cohort was to assess the pattern of FLT3 clonal evolution.METHODS:Eligible adult patients had evidence of new diagnosis of AML or R/R AML between 1 January 2009 and 31 December 2019 and had undergone at least one FLT3 test at National Taiwan University Hospital. Patients were followed up until either the end of the data period, loss to follow up (no records for 6 months), or death. We analyzed the time from the date of ordering FLT3 test to receipt of results (both cohorts) and the acquisition (negative at diagnosis, positive at R/R), loss (positive at diagnosis, negative at R/R), or maintenance (no change in status) of FLT3 mutation at R/R compared with status at diagnosis. We also explored the association between acquisition/loss/maintenance of FLT3 mutation at R/R and overall survival (OS) in the R/R cohort.RESULTS:From a total AML dataset of 1,789 patients, 659 and 223 were included in the ND and R/R cohorts, respectively ( Table 1). Demographics and clinical characteristics of the ND and R/R cohorts were similar. The most common type of AML was recurrent genetic abnormalities in both ND and R/R cohorts (50.7% and 67.7%, respectively); 45.8% and 48.0% of patients, respectively, had a normal karyotype; and NPM1 (21.1% and 16.6%, respectively), and IDH1/ IDH2 (18.5% and 15.2%, respectively) were the most common types of co-mutations detected. There was no difference in the turnaround time of FLT3 testing between the ND and R/R cohorts ( Table 1). Since 2020, FLT3 testing protocols have been further optimized to meet a 7-day turnaround time; additionally, almost all chemotherapy-fit patients with FLT3 mutation receive midostaurin in combination with 3+7 chemotherapy after reimbursement of midostaurin was implemented in Taiwan in February 2020.Of 164 patients with FLT3 tests at both diagnosis and R/R, most (75.6%, n=124) were FLT3 mutation-negative at diagnosis ( Table 1). At R/R, 27/124 (21.8%) acquired an FLT3 mutation and 12/40 (30.0%) patients lost their FLT3 mutation ( Table 1). Risk of mortality was increased in patients who acquired an FLT3 mutation (HR [95% CI] 1.59 [0.97, 2.62]), maintained their FLT3 mutation (1.52 [0.58, 2.56]), or lost their FLT3 mutation (1.22 [0.58, 2.56]), compared with those who were FLT3-mutation negative at both diagnosis and R/R ( Figure 1).CONCLUSIONS:This study provided information on the real-world landscape of FLT3 testing among patients with AML in Taiwan. The turnaround time of FLT3 testing in the current study was longer than the time reported in clinical trials, likely owing to different procedures in sampling arrangement and report delivery between real world and clinical trial settings.The distribution of FLT3 mutation and pattern of clonal evolution were largely consistent with previously published clinical studies. The higher prevalence of FLT3 mutation in the R/R cohort versus the ND cohort is in accordance with previous research showing that FLT3-ITD mutation is predictive of relapse. Patients with FLT3 mutation at R/R had a shorter OS, which indicates the importance of FLT3 testing at R/R diagnosis to optimize outcomes through targeted therapy.