Necrosis-Related lncRNAs: Biomarker Screening and Prognostic Prediction for Hot and Cold Tumours of Prostate Cancer.

Abstract

Background: The objectives of this work were the search for novel prognostic biomarkers for the diagnosis of prostate cancer (PCa) and the improvement of therapy outcomes in cases with a poor prognosis and the failure of immunotherapy.

Methods: The GTEx (Genotypic Tissue Expression Project) and TCGA (The Cancer Genome Atlas) databases were used to find out the co-expressed long non-coding RNAs (lncRNAs) associated with necrosis status based on statistics and univariate Cox regression tests. IncRNA associated with necrosis was screened by least absolute shrinkage and selection operator (Lasso) analysis, and the predictive model was further verified by Kaplan-Meier analysis, Receiver Operating Characteristic (ROC) curve analysis, Cox regression, nomogram and calibration curve. Also, immune analysis, principle component analysis, gene set enrichment analysis and prediction of semi-maximum inhibitory concentration in risk groups were conducted.

Results: The model successfully identified 16 necrosis-related lncRNA models, demonstrating good consistency among the calibration map and prognosis expectation. The ROC curve's area under the curve (AUC) for 1-year overall survival was 0.726, 0.763 and 0.770. The risk groups identified by the model could guide systematic treatment due to significant differences in semi-inhibitory concentrations. The study also demonstrated that the model could differentiate amongst hot and cold tumours and provide accurate mediation, with cluster 2 recognised as a hot tumour and likely to benefit from immunotherapy drugs.

Conclusions: In conclusion, the given study supports the potential of necrosis-related lncRNAs as a biomarker for predicting the prognosis and personalised treatment for PCa.