AMSTERDAM – Patients with acute myeloid leukemia who were in morphological complete remission prior to allogeneic hematopoietic cell transplant but had genomic evidence of a lingering AML variant had worse posttransplant outcomes when they underwent reduced-intensity conditioning, rather than myeloablative conditioning, investigators reported.
Among adults with AML in remission after induction therapy who were randomized in a clinical trial to either reduced-intensity conditioning (RIC) or myeloablative conditioning prior to transplant, those with known AML variants detected with ultra-deep genomic sequencing who underwent RIC had significantly greater risk for relapse, decreased disease-free survival (DFS), and worse overall survival (OS), compared with similar patients who underwent myeloablative conditioning (MAC), reported Christopher S. Hourigan, DM, DPhil, of the Laboratory of Myeloid Malignancies at the National Heart, Lung, and Blood Institute in Bethesda, Md.
The findings suggest that those patients with pretransplant AML variants who can tolerate MAC should get it, and that investigators need to find new options for patients who can’t, he said in an interview at the annual congress of the European Hematology Association.
“If I wasn’t a lab investigator and was a clinical trialist, I would be very excited about doing some randomized trials now to try see about novel targeted agents. For example, we have FLT3 inhibitors, we have IDH1 and IDH2 inhibitors, and I would be looking to try to combine reduced-intensity conditioning with additional therapy to try to lower the relapse rate for that group at the highest risk,” he said.
Previous studies have shown that, regardless of the method used – flow cytometry, quantitative polymerase chain reaction, or next-generation sequencing – minimal residual disease (MRD) detected in patients with AML in complete remission prior to transplant is associated with both cumulative incidence of relapse and worse overall survival.
Measurable, not minimal
Dr. Hourigan contends that the word “minimal” – the “M” in “MRD” – is a misnomer and should be replaced by the word “measurable,” because MRD really reflects the limitations of disease-detection technology.
“If you tell patients ‘you have minimal residual disease, and you have a huge chance of dying over the next few years,’ there’s nothing minimal about that,” he said.
The fundamental question that Dr. Hourigan and colleagues asked is, “is MRD just useful for predicting prognosis? Is this fate, or can we as doctors do something about it?”
To get answers, they examined whole-blood samples from patients enrolled in the, which compared survival and other outcomes following allogeneic hematopoietic stem cell transplants (allo-HSCT) with either RIC or MAC for pretransplant conditioning in patients with AML or the myelodysplastic syndrome.
The trial was halted early after just 272 of a planned 356 patients were enrolled, following evidence of a significantly higher relapse rate among patients who had undergone RIC.
“Strikingly, over half the AML patients receiving RIC relapsed within 18 months after getting transplants,” Dr. Hourigan said.
Relapse, survival differences
For this substudy, the National Institutes of Health investigators developed a custom 13-gene panel that would detect at least one AML variant in approximately 80% of patients who were included in aof genomic classification and prognosis in AML.
They used ultra-deep genomic sequencing to look for variants in blood samples from 188 patients in BMT CTN 0901. There were no variants detected in the blood of 31% of patients who had undergone MAC or in 33% of those who had undergone RIC.
Among patients who did have detectable variants, the average number of variants per patient was 2.5.
In this cohort, transplant-related mortality (TRM) was higher with MAC at 27% vs. 20% with RIC at 3 years, but there were no differences in TRM within conditioning arms for patients, with or without AML variants.
Relapse rates in the cohort studied by Dr. Hourigan and his colleagues were virtually identical to those seen in the full study set, with an 18-month relapse rate of 16% for patients treated with MAC vs. 51% for those treated with RIC.
Among patients randomized to RIC, 3-year relapse rates were 57% for patients with detectable pretransplant AML variants, compared with 32% for those without variants (P less than .001).
Although there were no significant differences in 3-year OS by variant status among patients assigned to MAC, variant-positive patients assigned to RIC had significantly worse 3-year OS than those without variants (P = .04).
Among patients with no detectable variants, there were no significant differences in OS between the MAC or RIC arms. However, among patients with variants, survival was significantly worse with RIC (P = .02).
In multivariate analysis controlling for disease risk and donor group among patients who tested positive for an AML variant pretransplant, RIC was significantly associated with an increased risk for relapse (hazard ratio, 5.98; P less than .001); decreased DFS (HR, 2.80; P less than .001), and worse OS (HR, 2.16; P = .003).
“This study provides evidence that intervention for AML patients with evidence of MRD can result in improved survival,” Dr. Hourigan said.
Questions that still need to be addressed include whether variants in different genes confer different degrees of relapse risk, whether next-generation sequencing positivity is equivalent to MRD positivity, and whether the 13-gene panel could be improved upon to lower the chance for false negatives, he said.
The study was supported by the NIH. Dr. Hourigan reported research funding from Merck and Sellas Life Sciences AG, research collaboration with Qiagen and Archer, advisory board participation as an NIH official duty for Janssen and Novartis, and part-time employment with the Johns Hopkins School of Medicine.
SOURCE: Hourigan CS et al. EHA Congress, Abstract LB2600.