Novel DNA Repair–Related Gene Prognostic Model for Patients With Breast Cancer
Posted: Monday, January 25, 2021
According to research published in JAMA Network Open, a novel prognosis model may be a reliable predictor for overall survival in patients with breast cancer. Developed by Yiche Li, MD, of Hebei Medical University in China, and colleagues, the model constructs a new signature for breast cancer using eight DNA repair–related genes.
“We hope that it [the novel 8 DNA repair-related gene signature] can be applied in clinical treatments or research studies as a potential prognostic biomarker of breast cancer,” the investigators commented.
Between October 2019, and February 2020, the prognostic study gathered gene-expression profiles and clinical data on 1,096 women with breast cancer from The Cancer Genome Atlas database and the Gene Expression Omnibus. Patients were divided into a training cohort, which used data from the Atlas, and two validation cohorts, which used data from the Omnibus. Prognostic biomarkers from primary patient screening were evaluated via Cox proportional hazards regression analysis. The DNA repair–related gene signatures and the clinical characteristics of patients were used to construct the novel prognostic nomogram.
The eight prognostic biomarkers that formed the new signature included MDC1, RPA3, MED17, DDB2, SFPQ, XRCC4, CYP19A1, and PARP3. An analysis of the time-dependent receiver operating characteristic curve found that this signature was likely to have “good predictive accuracy.” The AUC values were comparable across cohorts, at 0.708 for 3-year survival and 0.704 for 5-year survival in the training cohort; 0.717 and 0.772, respectively, in the first validation cohort; and 0.691 and 0.718, respectively in the second validation cohort. The primary DNA repair–related gene signature engagement came from selecting regulation pathways of vascular endothelial cell proliferation.
Disclosure: The study authors reported no conflicts of interest.