3358 - Single-Cell and Bulk Transcriptomes Reveal a Two-Gene Based Clinically Subtypes for Gleason Grade 5 Prostate Cancer Patients Receiving Radiotherapy
Presenter(s)
F. Wang1, X. Li2, Q. Huang3, X. Gao1, and C. Liu1; 1Department of Radiation Oncology, Peking University First Hospital, Beijing, China, 2Department of Medical Oncology, Peking University First Hospital, Beijing, China, 3Department of Oncology, Renmin Hospital of Wuhan University, Wuhan, China
Purpose/Objective(s): Patients with Gleason Grade 5 prostate cancer have a high risk of distant metastases following radiotherapy, yet accurate prognostic stratification of whether additional intensive therapy is required after radiotherapy remains challenging. We aimed to construct a prognostic classifier of Gleason grade 5 prostate cancer after radiotherapy, thereby optimizing clinical decision-making.
Materials/Methods: We integrated a cohort of single-cell RNA sequencing data of 36,424 cells, and bulk RNA sequencing data from 196 Gleason Grade 5 prostate cancer samples treated with radiotherapy. Differential expression analysis, t test and Cox proportional-hazards models were used to identify prognosis-associated genes. Kaplan–Meier survival analysis and the area under receiver operating characteristic curve (AUC) were performed to select the most efficient model. Functional enrichment analyses, including Gene Ontology, were conducted to explore molecular heterogeneity across prognostic subgroups.
Results: We generated distinct gene expression profiles from a cohort of 196 Gleason Grade 5 prostate cancer samples between different outcome after radiotherapy, revealing that high LAMTOR2 or SERTAD3 expression correlated with improved survival, whereas elevated PDCD11 or NAA50 predicted worse outcomes (all P <0.001). In order to establish a clinical convenient and efficient classifier, an expression signature predictive of overall survival was reduced to a two-gene ratio, PDCD11 and NAA50, with AUC values of 0.88, which outperformed other biomarkers. Moreover, the single-cell RNA sequencing data supported that both genes were highly expressed on the tumor cells. Therefore, we construct a dual-gene prognostic model classified patients into three risk groups: the PDCD11-low/NAA50-low subgroup showed the best prognosis, enriched with response to virus and oxidative stress pathways, while the PDCD11-high/NAA50-high subgroup exhibited the poorest outcome (P = 0.0011), enriched with epithelial cell morphogenesis and regulation of vasculature development pathways.
Conclusion: This study established a two-gene based clinically subtypes leveraging PDCD11 and NAA50 expression to predict survival in Gleason Grade 5 prostate cancer patients treated with radiotherapy. This tool enhances risk stratification, identifies candidates for intensified therapy, and supports personalized clinical management.