Insights From Nonsense-Mediated mRNA Decay for Prognosis in Homologous Recombination-Deficient Ovarian Cancer

Area of Science:

• Oncology
• Genetics
• Molecular Biology

Background:

• Ovarian cancer patients with homologous recombination deficiency (HRD), particularly those with germline BRCA mutations, exhibit variable responses to platinum-based and targeted therapies.
• Identifying predictive biomarkers is crucial for optimizing treatment strategies in this patient subgroup.

Purpose of the Study:

• To investigate the prognostic value of nonsense-mediated mRNA decay (NMD) in ovarian cancer.
• To develop and validate a machine learning-based gene model for predicting prognosis and treatment response in ovarian cancer with HRD or germline BRCA variants.

Main Methods:

• Retrospective analysis of 797 ovarian cancer patients from public and in-house cohorts.
• Development of a prediction algorithm for NMD status (trigger vs. escape).
• Differential gene expression analysis and functional pathway enrichment.
• Construction of an optimized key gene model using integrated machine learning algorithms.

Main Results:

• NMD 'escape' status was associated with a better prognosis.
• An 8-gene model related to the cell cycle demonstrated high predictive accuracy (mean AUC > 0.89) for prognosis in ovarian cancer with HRD or germline BRCA variants.
• Patients classified into low-risk groups showed improved prognosis, enhanced drug response, and higher levels of activated dendritic cells compared to high-risk groups.

Conclusions:

• The developed machine learning model offers superior prognostic value compared to the NMD algorithm alone for ovarian cancer with HRD or germline BRCA variants.
• This study provides a novel perspective using NMD and cell cycle pathways to stratify ovarian cancer subtypes, potentially guiding personalized treatment decisions.