Development of a Prognostic Risk Model for Prostate Cancer Patients (Based on Patients with Gleason Score 9-10)

Abstract

<h3>Purpose/Objective(s)</h3> Prostate cancer patients with Gleason scores (GS) 9-10 have high disease malignancy and poor prognosis. In this study, we will construct a prognostic risk model for prostate cancer patients starting from the molecular characteristics of patients with GS 9-10. <h3>Materials/Methods</h3> (1) Firstly, using the information of prostate cancer patients in the TCGA database (N=499), the differential genes were screened for specific screening criteria: prostate cancer tumor tissue, GS 9-10 score, and differential genes in patients with fast-progressing prostate cancer (Log2FC≥1, adjusted P-value <0.05), and these three components were taken as intersection. (2) Then, the GO, KEGG and GSEA enrichment analysis of the differential genes was carried to analyze the function of these genes. (3) And the analysis of protein interaction network for the differential molecules was carried in Metascape website. (4) We next brought the differential risk markers into lasso regression, calculating the risk model formula and risk scores. (5) Besides, we combined clinical factors (such as age, T, N stage, etc.) to establish a nomogram to predict the progress-free survival (PFS). (6) Finally, according to the risk formula, we analyzed the relationship between risk scores and immune infiltration of patients, providing clues for immunotherapy. <h3>Results</h3> (1) Differential gene intersections were obtained for three parts, and 83 genes was in total, further screening WASIR1, KPTAP5-1, TLX1, KIF4A, IQGAP3 as candidate molecules for risk markers. (2) The GO, KEGG and GSEA enrichment analysis showed that these genes were associated with biological processes such as cell cycle, steroid hormone metabolism, and hormone activity and so on, suggesting that these may be biological in prostate cancer patients with poor prognosis processes. (3) We constructed a risk formula for PFS of prostate cancer patients based on these factors as: Riskscore = (0.4851) *WASIR1+(0.4418) *KRTAP5-1+(0.6872) *TLX1+(0.1664) *KIF4A+(0.1699)* IQGAP3, which had high potency for predicting PFS, 10-year AUC=0.805, 95% CI (0.652-0.795); (4) Immune infiltration analysis showed a decrease in CD4+ and CD8+ T cells and an increase in macrophage infiltration in tumor tissues of high-risk patients. <h3>Conclusion</h3> Our study screened the risk markers and formula for predicting the prognosis of prostate cancer patients by TCGA database, which can calculate the risk score of patients according to the risk formula and provide some guidance for prognosis prediction and clinical treatment.