Murky waters for MRD detection in AML: flightyFLT3/ITDs

A guiding principle in the application of MRD monitoring in leukemia has been that the mutant genes present at diagnosis and causally implicated in disease pathogenesis will be harbingers of molecular relapse. This principle has been thoroughly validated, for example, in the context ofBCR/ABL in CML. But 2 articles in this issue indicate that FLT3/ITD may not be a suitable marker for MRD in a subset of AML patients. Shih et al (page 2387) and Kottaridis et al (page 2393) report the fascinating observation that someFLT3/ITD-positive patients at diagnosis do not have detectable FLT3/ITDs at relapse, and the converse may also be true. Furthermore, a cohort of patients had severalFLT3/ITD variants detectable at diagnosis, with selection for one of the variants at time of relapse.

These observations have several important implications. First, use ofFLT3/ITD for MRD detection must be employed with caution and will not be of value in detecting molecular relapse in all AML patients that are FLT3/ITD positive at diagnosis. Second, the data indicate that FLT3/ITD is probably a secondary event in clonal evolution of at least some AML. Viewed from another perspective, these data provide further support for 2-hit pathogenesis of AML in which activating mutations in FLT3 confer a proliferative and/or survival benefit to leukemia clones. Third, the data raise the question of whether FLT3/ITD inhibition using small molecule inhibitors, analogous to imatinib, will be an effective therapy for AML or will simply select for a parent clone with an alternative second mutation that confers a proliferative signal. Indeed, one patient with FLT3/ITD at diagnosis had “substituted” a clone with an activating mutation in RAS for theFLT3/ITD. It seems likely (and comes as no surprise given the more modest response rates to imatinib for CML blast crisis) thatFLT3/ITD inhibitors will need to be combined with other agents for effective therapy of AML. Finally, FLT3,RAS, and KIT may each be activated by mutation in AML and confer proliferative and survival signals, but collectively account for perhaps one-half of all AML cases. In those cases in which FLT3, RAS, or KITare not mutant, it seems likely that other mutations must confer a proliferative signal to cells. Identification of these genes will be important in the future, as they may also be targets for small molecule inhibition.