Fms-like tyrosine kinase 3 internal tandem duplication mutations (FLT3-ITD+) occur in 25% of newly diagnosed patients with AML. In patients younger than age 60 with normal karyotype AML, FLT3-ITD mutations consistently have been shown to be associated with high relapse rates and short time to relapse, particularly, in patients who receive chemotherapy alone as post-remission treatment. Remission rates after relapse are low and consequently survival is poor.1, 2 Accordingly, allogeneic transplantation has been employed in appropriate patients to attempt to mitigate the negative prognostic impact of FLT3-ITD mutation.

Whether allogeneic transplant for FLT3-ITD+ AML actually improves overall survival over traditional chemotherapy has not been specifically studied in a randomized prospective fashion. Instead, one must infer these results from a number of pediatric and younger adult AML trials, where AML patients in first remission with matched siblings received myeloablative transplants.3, 4, 5 Although initial results of donor versus no-donor analyses from these trials were conflicting and analyses complicated by low rates of transplant among patients with available donors, subsequent data with higher protocol adherence suggest rather substantial protection from relapse among FLT3-ITD+ patients. The magnitude of this benefit is sufficient to expect an improved survival from the approach and transplant is now an accepted frontline therapy for FLT3-ITD+ AML patients. Still, even within the FLT3-ITD+ population, relapse risk is variable. Although based upon a number of clinical and leukemia-specific factors, selection of optimal FLT3-ITD+ patients for transplant remains challenging. In this issue, Song and colleagues6 report addresses these questions; their report is both welcome and topical.

Song and colleagues6 report is notable not only for being among the largest single-center experiences reporting transplant outcomes in the FLT3-ITD+ population but also for its inclusion of patients transplanted with both myeloablative and reduced intensity conditioning, inclusion of unrelated donors as well as matched siblings and a wide range of patient ages from pediatric to older adults. From this broad experience, the authors show not only an impressively favorable tolerability of transplant among FLT3-ITD+ patients but a 3 year overall survival of only 38% with a strikingly high relapse rate (63%) post transplant. Such high relapse rates are concerning, and the authors convincingly show FLT3-ITD+ status at diagnosis—once controlled for a number of confounders—is an independent risk factor for relapse and mortality following transplant. Leukemic relapse was far and away the dominant cause of mortality among FLT3-ITD+ patients (26/28 of the post-transplant deaths due to relapse, 4% non-relapse mortality). Overall, the data unfortunately suggest purported survival benefits from transplant may be modest when applied to a broader population than that enrolled to multicenter prospective trials noted above.

Similarly, Song and colleagues' data describe generally higher relapse rates and worse post-transplant survivals for FLT3-ITD+ AML than that of large registry reports, from EBMT,7, 8 and CIBMTR.9 These registry studies report relapse rates after first remission transplant in the range of ~30%, with expected cure rate of ~50% for transplanted patients. Notably, outside of a recent EBMT analysis of transplant for normal karyotype AML that included patients up to 71 years of age and both myeloablative and reduced intensity conditioning,8 prior registry reports examining the merits of transplant for FLT3-ITD+ AML focused upon considerably younger patients who received T-cell replete, myeloablative allografts from HLA-matched siblings and/or 8/8 antigen-matched unrelated donors. The higher relapse rates reported by Song and colleagues may in part relate to the older age of their patients (median age of 55, with 28% older than 60 years), variations in preparative regimens or other transplant-specific treatments. Additionally, leukemia-specific contributors to relapse reasonably could account for the differences in outcome. Among the reported high-risk features of Song’s study population include a sizeable number (20%) of the FLT3-ITD+ patients who were not in remission at the time of transplant. Although not annotated in Song and colleagues' report, other leukemia-specific negative prognostic factors likely contributed to relapse risk among patients transplanted in remission. These include DNMT3A mutations,10, 11, 12 detectable minimal residual disease,13, 14 high diagnostic FLT3-ITD:WT allelic ratio15, 16 or distal ITD insertion site.15

Although at first glance, the relapse rates from Song and colleagues' data set are disappointing and not as good as those obtained from registry studies listed above, the non-randomized survivals seen here should be placed in the context of that seen in the absence of transplant. FLT3-ITD+ patients under the age of 60 who receive only intensive chemotherapy as post-remission therapy show cumulative relapse rates of 70–94% on cooperative group studies and 5-year survival of 21–31%.3, 17 As might be expected, older FLT3-ITD+ patients treated with chemotherapy-only regimens fare even worse with 79% relapsing and a 3-year overall survival of only 14%.18 These statistics provide a sobering comparator to Song and colleagues' data.

Thus, even the generally poor leukemia-free survival reported here may still hide a rather substantial clinical benefit for transplant over non-transplant consolidation. This is particularly important when assessing potential benefits of transplant among patients over the age of 60, where prospective post-remission comparisons that include transplant are scarce. Although clearly an imperfect solution, allografting of fit older patients in first remission likely still provides significant additional anti-leukemic benefit, albeit with significant associated risk and toxicity. Whether this benefit can be confirmed by prospective studies in older FLT3-ITD+ AML patients is unclear but should be studied, if feasible

Can outcomes for FLT3-ITD+ patients be further improved? Additional or alternate low toxicity interventions certainly appear warranted to reduce risk of post-transplant relapse. As FLT3-ITD encodes a constitutively activated tyrosine kinase, a number of drugs have been designed to inhibit its oncogenic signal transduction. Although early reports of FLT3 inhibitors such as lestaurtinib or midostaurin as single agents were disappointing, these earliest developed drugs were poorly optimized for target selectivity, potency or pharmacokinetic properties. Indeed, more bioavailable, potent, and selective drugs such as quizartinib (AC220), crenolanib and gilteritinib (ASP2215) have been recently tested in phases 1 and 2 trials and appear to generate far more significant activity in the clinic with limited toxicity.19, 20, 21 As well, a recent report from a randomized placebo-controlled trial in newly diagnosed FLT3-mutated AML patients suggested that the addition of midostaurin to frontline therapy significantly improved overall survival.22

In addition, FLT3 inhibitor maintenance therapy may prove to be a critical component in reducing post-transplant relapse rates. Encouraging early data have been reported for sorafenib, midostaurin and quizartinib, when given as post-transplant maintenance with single-arm studies showing remarkably low relapse rates.23, 24, 25 Phase 3 trials of current generation, selective FLT3 inhibitors as salvage therapies for patients with relapsed/refractory FLT3-ITD+ AML, and randomized studies of selective, potent inhibitors combined with frontline chemotherapy as well as post-transplant maintenance are underway or planned. If confirmatory, these data may prove transformative for this high-risk population.