Discovery of lncRNA-Based ProsRISK Score in Serum as Potential Biomarkers for Improved Accuracy of Prostate Cancer Detection

Circulating lncRNAs have emerged as promising biomarkers for the diagnosis of various cancers. This study aimed to establish an accurate risk prediction model based on serum lncRNAs to facilitate the detection of prostate cancer (PCA). RT-qPCR was used to analyse the levels of candidate lncRNAs, and four lncRNAs (NEAT1, ARLNC1, FOXP4-AS1 and DSCAM-AS1) were identified to be differently expressed in serum from 190 PCA patients, 140 benign controls, and 170 healthy controls. A ProsRISK score based on four lncRNAs and prostate-specific antigen (PSA) was established in the training set. ROC analysis in the validation set revealed that the ProsRISK demonstrated more powerful capacity in discriminating PCA from healthy and benign controls, with an AUC of 0.926 (95% CI: 0.882-0.970) and 0.837 (95% CI: 0.770-0.904), which were significantly higher than those of the lncRNA panel or PSA alone (all at p < 0.05). Moreover, the ProsRISK showed good diagnostic performance for PCA I-II patients compared with healthy and benign controls, and the corresponding AUCs were 0.905 (95% CI: 0.843-0.968) and 0.819 (95% CI: 0.732-0.907). Our findings indicated that the constructed ProsRISK could be a reliable risk stratification model and have great potential for clinical use to improve the precision surveillance for PCA.