SPECIFIC GENE SET SCORES IN PROGNOSTIC PREDICTION OF MUSCLE INVASIVE BLADDER CANCER

Yansheng Kan,Qing Zhang,Xiaozhi Zhao,Hongqian Guo

DOI: https://doi.org/10.1097/ju.0000000000000911.08

2020-01-01

Abstract:INTRODUCTION AND OBJECTIVE: Bladder cancer is one of the most common urogenital tract tumor worldwide. The convenient access to The Cancer Genome Atlas (TCGA) database enables us to explore the genes that have a predictive role in overall survival through genome-wide expression profiles. The prediction may provide help for easier classification of subtypes, so as to make better diagnosis and prognosis prediction. METHODS: Tumor microenvironments, including immune cells and stromal cells, have been shown to be closely related to tumor progression and prognosis. By leveraging the ESTIMATE algorithm, we could calculate the immune and stromal scores in the tumor. To better understand the genes' role in tumor prognosis in immune and stromal sets, patients with muscle invasive bladder cancer (MIBC) in TCGA database were divided into immune and stromal high/low groups, and the differentially expressed genes (DEGs) were obtained. Then Gene Set Variation Analysis (GSVA) was used to comprehensively score these DEGs to verify their predictive effect on the prognosis of bladder cancer. Finally, these screened DEGs and scores were used to verify their universality in an external database in GEO database (GSE48277GPL6947). RESULTS: Gene expression profiles and clinical data of 393 patients with MIBC were downloaded from the TCGA database. Utilizing the ESTIMATE algorithm, we screened 27 DEGs in the groups with highand low-score in immune and stromal score set. Under KaplanMeier survival analysis, ACTC1, SFRP2 and OMD could individually predict the prognosis with statistical significance (log-rank test p<0.05). The scores of combination of the DEGs through GSVA showed significant differences in prognostic prediction (log-rank test p<0.05). Also, DEG scores could distinguish some subtypes. Luminal-papillary subtype had the lowest score, and Luminal subtype usually had higher score (Fig. 1A). In the GEO dataset of 73 patients with MIBC, similarly using GSVA, ACTC1, OMD group and SFRP2, OMD group had significant differences in predicting prognosis (logrank test p[0.0295 and 0.0052, respectively) (Fig. 1B). CONCLUSIONS: By synthetic analysis of individual DEGs and DEG sets, we obtained tumor microenvironment-related genes and gene sets that could predict the prognosis of bladder cancer. Source of Funding: This study was funded by the National Natural Science Foundation of China (81802535), China postdoctoral fund(223427), Nanjing Medical Science and technique Development Foundation (YKK 18064)