In a study reported in the Journal of Clinical Oncology, Farrar et al developed a long noncoding RNA (lncRNA) transcription scoring system that distinguished survival outcomes in pediatric patients with acute myeloid leukemia (AML).
As stated by the investigators, “Risk classification in pediatric AML relies on detection of translocations and gene mutations. lncRNA transcripts have been shown to associate with and mediate malignant phenotypes in AML but have not been comprehensively evaluated in pediatric AML.”
Study Details
To identify lncRNA transcripts associated with outcomes, the annotated lncRNA landscape was analyzed by transcript sequencing of 1,298 pediatric and 96 adult AML specimens. Upregulated lncRNAs identified in the pediatric AML training set were used to construct a Cox regression model of event-free survival using a 37 lncRNA signature (lncScore). The pediatric AML cohort was separated into a training cohort (n = 780) and validation 1 (n = 260) and validation 2 (n = 258) cohorts.
Key Findings
Training cohort cases with positive lncScores had 5-year event-free survival and overall survival rates of 26.7% (hazard ratio [HR] = 2.48, 95% confidence interval [CI] = 2.05–3.00, P < .001) and 42.7% (HR = 3.16, 95% CI = 2.47–4.04, P < .001), respectively, compared with 56.9% and 76.3% with negative lncScores.
In the two pediatric validation cohorts, negative vs positive lncScores were also associated with improved 5-year event-free survival (HRs = 2.38, P < .001, and 2.36, P < .001) and overall survival (HRs = 2.87, P < .001, and 3.0, P < .001). In the adult AML cohort, negative scores were associated with improved 5-year event-free survival (HR = 2.36, P < .001) and overall survival (HR = 3.21, P < .001).
On multivariate analysis, lncScore was independently prognostic for event-free survival (HR = 1.84, P < .001) and overall survival (HR = 2.07, P < .001) in the training cohort and in the combined pediatric validation cohorts (HR for event-free survival = 1.56, P = .001; HR for overall survival = 1.75, P = .001).
Subgroup analysis indicated that lncScore furnished additional outcome information in heterogeneous subgroups currently classified as indeterminate risk.
Concordance analysis showed that lncScore added to overall classification accuracy with a predictive performance at least comparable to current risk-stratification methods that rely on multiple assays.
The investigators concluded, “Inclusion of the lncScore enhances predictive power of traditional cytogenetic and mutation-defined stratification in pediatric AML with potential, as a single assay, to replace these complex stratification schemes with comparable predictive accuracy.”
Jason E. Farrar, MD, of the Department of Pediatrics, Arkansas Children’s Research Institute, University of Arkansas for Medical Sciences, Little Rock, is the corresponding author for the Journal of Clinical Oncology article.
Disclosure: The study was supported by National Cancer Institute grants, St. Baldrick’s Foundation, and others. For full disclosures of the study authors, visit ascopubs.org.
