Genetic alterations at diagnosis predict transplant outcomes for older patients with AML
Click Here to Manage Email Alerts
ORLANDO — Genetic alterations at diagnosis predicted OS and leukemia-free survival for older individuals with acute myeloid leukemia who underwent allogeneic hematopoietic stem cell transplantation in first remission, according to study results presented at ASH Annual Meeting and Exposition.
Researchers created an integrated genetic and clinical model that identified distinct risk groups within this population
“Our data suggest patients classified as very high risk would perhaps benefit from strategies aimed at mitigating their high rate of relapse, either through conditioning intensity or novel maintenance strategies,” H. Moses Murdock, research fellow at Dana-Farber Cancer Institute and student at Perelman School of Medicine at University of Pennsylvania, said during a presentation. “Conversely, even in a cohort in which a majority of patients received reduced-intensity conditioning, patients we classified as low risk have very low rates of relapse and would benefit from approaches aimed at minimizing therapy-related toxicity.”
Older patients with AML often have a high prevalence of genomically poor-risk disease, often reflective of antecedent myelodysplastic syndrome. Allogeneic HSCT is a potentially curative treatment for AML; however, older age is associated with poorer outcomes.
Murdock and colleagues assembled a retrospective cohort of adults aged 60 years or older who underwent transplantation in first complete remission and examined how the genomics of AML at the time of diagnosis affected outcomes. Patients with acute promyelocytic leukemia or isolated extramedullary disease were excluded.
The cohort included 300 patients (median age, 66 years; range, 60-76; 59.3% men) from six institutions. About 41% of patients had Hematopoietic Cell Transplantation-specific Comorbidity Index scores of 3 or higher.
Nearly one-third of patients (30.3%) had secondary AML, and 40% of those patients received prior therapy with hypomethylating agents.
Per 2017 European LeukemiaNet criteria, 28.7% of patients had intermediate risk and 50.6% had adverse risk. Most patients (85%) received intensive induction chemotherapy.
Half (49.3%) of patients received transplant from matched unrelated donors, whereas 10.3% had unmatched unrelated donors, 18% had matched related donors and 2% had unmatched related donors. About one in five patients (19.7%) had alternative donors.
Two-thirds (65.7%) of patients received reduced-intensity conditioning, and only 9.3% of patients received myeloablative conditioning.
More than three-quarters (78%) of patients achieved complete remission with hematologic recovery, whereas 22% achieved complete remission with incomplete hematologic recovery.
Researchers reported median leukemia-free survival of 2.8 years, with median follow-up for survivors of 3.75 years.
Relapse was the primary mode of failure after transplantation. At 3 years, researchers reported a 36% cumulative incidence of relapse and a 23% cumulative incidence of nonrelapse mortality.
To link genetics to transplant outcomes, researchers performed targeted sequencing of 112 genes from diagnostic leukemia samples. They identified approximately 1,200 mutations in 63 genes.
Nearly all patients (96%) had at least one mutation. Eleven percent had TP53 mutations, 43% had secondary-type mutations in genes associated with prior myelodysplastic syndrome and 22% had NPM1 mutations. More than half of patients with NPM1 mutations had FLT3-internal tandem duplications.
Six percent of patients had DDX41 mutations. These are associated with germline predisposition to myeloid malignancies with long latency, Murdock said.
Murdock and colleagues used univariable and multivariable Cox models to evaluate the impact of gene mutations on leukemia-free survival, and they developed a hierarchical model of three molecular genetic risk groups.
They characterized patients with TP53 mutations, JAK2 mutations or FLT3-internal tandem duplication in the absence of NPM1 mutations as high risk.
Patients who had DNMT3A, DDX41 or NP1 mutations in the absence of FLT3-internal tandem duplication were characterized as intermediate risk.
All other patients were considered low risk.
Based on the molecular model, rates of 3-year leukemia-free survival were 8% in the high-risk group, 47% in the intermediate-risk group and 65% in the low-risk group (P < .001).
Researchers then performed multivariable analysis of clinical variables that affect posttransplant analysis. They identified two clinical variables — complete remission with hematologic recovery vs. complete remission with incomplete hematologic recovery, as well as adverse karyotype status — that were significant and integrated them into their risk model.
Using the integrated risk model, they reported 3-year leukemia-free survival rates of 5% for patients with very high risk (accounting for 16.4% of cohort, 29% for those with high risk (31.5% of cohort), 51% for those with intermediate risk (26.8% of cohort) and 70% for those with low risk (25.2% of cohort).
Nearly 80% of patients in the very high-risk group relapsed, and that was the primary contributor to poor outcomes in that group, Murdock said. In the other groups, outcomes were a composite of relapse and nonrelapse mortality.
“We show that an integrated genetic and clinical model is able to identify distinct risk groups among older adults undergoing allogeneic transplantation in first remission,” Murdock said. “The fact that remission quality was associated with outcome suggests that assessment of measurable residual disease may aid in guiding clinical management. Our work suggests ... that prospective clinical trials should perhaps consider genetic risk as we work to improve outcomes in this patient population.” – by Mark Leiser
Reference: Murdock HM, et al. Abstract 48. Presented at: ASH Annual Meeting and Exposition; Dec. 7-10, 2019; Orlando.Disclosure: Murdock reports no relevant financial disclosures. Please see the abstract for all other authors’ relevant financial disclosures.
Perspective
Back to Top
Joseph Alvarnas, MD
We have been really struggling to figure out how to most effectively treat older individuals with AML. The challenge is that we have some amazing therapeutic choices that include allogeneic stem cell transplantation, but we can also look at a population of older patients and find that the outcome when they receive that form of treatment is very highly variable. We don’t want to overtreat those who can be treated with less intensive therapeutic strategies, and we don’t want to deny those who are going to benefit from treatment an effective strategy even though it may be intensive.
We have risk models that help us make these kinds of decisions, and this abstract adds significantly to our understanding of risk in older patients with AML. We have from European LeukemiaNet a list of cytogenetics that help to characterize patient risk. This adds materially to that by overlaying additional data from genomic testing looking at a number of genes that have prognostic significance among patients with AML. By overlaying those two constructs, and based on targeted sequencing data from 112 genes, the researchers were able to construct a risk model that identified those patients at high molecular risk, intermediate molecular risk and low molecular risk.
The first key deliverable is that we’re able to take data from multigene sequencing and amalgamate it in ways that lend a more robust prism through which we can look at older patients with newly diagnosed AML to plot out a therapeutic strategy.
The other part is that they looked at this group of patients and how they fared undergoing transplant, but they overlay the two risk schemes — the risk scheme from genomic testing and the risk scheme from cytogenetics — to leverage the best of these data in order to figure out treatment approaches that take full advantage of the data richness we might have at the time of diagnosis to plot out a strategy. They were able to identify a 3-year leukemia-free survival for patients in the low-risk group vs. 0% for those with adverse cytogenetics and very high-risk molecular diagnostics. The other patients fall in between.
There are a couple key lessons here. For those at low risk, looking to identify approaches that reduce the toxicity of treatment is a laudable goal. On the other hand, for those who have a 0% 3-year leukemia-free survival, we need to look for meaningful therapeutic clinical trials that might afford them better opportunities moving forward. This takes the best of what we know and translates it into a scheme that is applicable across the United States for a next arc of learning.
I have spent a lot of time looking at risk models for acute leukemia, and there’s definitely splay. You can find characteristics that will portend a significant difference between groups of patients, but the difference between 69% and 0% is dramatic even in light of that. It makes this seem very powerful. If I’m caring for someone in that very high-risk group, my obligation as a clinician is to be much more creative, and we better have something ready to go for that group of patients. That could be a good set of clinical trials or different treatment algorithms. It also might require better preparedness around the issue of supportive care medicine and goals of care. But I think we have to respect the differences we see here.
Click Here to Manage Email Alerts