Relapse, Survival in AML Could Be Predicted With Next-Generation Sequencing

December 22, 2017

Next-generation sequencing could be a powerful and independent predictor for relapse and survival among adults with acute myeloid leukemia.

Next-generation sequencing (NGS) assessment of molecular minimum residual disease at complete remission (CR) is a powerful and independent predictor for relapse and survival among adults with acute myeloid leukemia (AML), according to a study (abstract LBA-5) presented at the American Society of Hematology (ASH) Annual Meeting & Exposition.

The persistence at CR of the DNMT3A, TET2, and ASXL1 (or “DTA”) mutations was not associated with the risk of subsequent relapse, but residual non-DTA mutations did predict relapse, the researchers found.

“In an unprecedentedly large AML cohort we showed that assessment of residual gene mutations in CR by NGS is applicable to the great majority of newly diagnosed adults with AML,” said Tim Grob, MD, of the department of hematology at Erasmus University Medical Center in Rotterdam, the Netherlands.

Clonal hematopoiesis–associated mutations had no impact on relapse in patients who achieved CR.

More than 80% of patients with AML achieve a complete morphologic remission after induction chemotherapy, but relapse is the leading cause of treatment failure, Grob noted. There is a sparse evidence base regarding which leukemia-specific mutations at CR predict relapse.

The authors performed NGS to sequence 52 genes at diagnosis for 430 adult patients with AML, and again after induction therapy among patients who achieved CR, using the Illumina TruSight Myeloid Sequencing Panel. No germline genome sequencing was undertaken.

Fifty percent of mutations detected at diagnosis persisted into CR. At least one mutation was detected in 90% of patients. On average, the panel detected 2.9 mutations for each patient. Those mutations represented a wide range of loci, however-illustrating the highly heterogeneous nature of the AML-associated genomic landscape, Grob said.

DTA mutations were the most frequent mutations seen at diagnosis and CR. They did not predict relapse.

However, an analysis of non-DTA mutations in CR showed that they were highly predictive for AML relapse (P = .001). Non-DTA mutations at CR were also highly predictive for overall survival (P = .012).

In multivariable analyses, residual leukemia (the persistence at CR of non-DTA mutations) was associated with relapse (hazard ratio [HR], 1.88; 95% CI, 1.34–2.64; P < .001).

“Risk of recurrence nearly doubles in patients with these mutations in CR,” Grob said.

In the multivariate analysis, overall survival also remained significantly associated with the presence of non-DTA mutations at CR (HR, 1.64; 95% CI, 1.18–2.27; P = .003).

In multivariate analysis including the presence at CR of non-DTA mutations, relapse risk also correlated with white blood cell count at diagnosis, European LeukemiaNet 2017 risk classification (adverse vs favorable), the number of cycles of induction therapy to achieve CR (2 vs 1), and postremission therapy (allogeneic stem cell transplant vs none).