Identification of key lncRNAs as prognostic prediction models for colorectal cancer based on LASSO.

Colorectal cancer (CRC) is one of the most common malignancies, with varying prognoses and a high mortality. There is an urgent need to establish a new prediction model to predict the survival risk of CRC patients. The long non-coding RNAs (lncRNAs) expression profiles and corresponding clinical information of CRC patients were obtained from The Cancer Genome Atlas, TCGA. We identified a total of 1,176 lncRNAs differentially expressed between 480 CRC and 41 normal tissues. In the training test, we combined these differentially expressed lncRNAs with overall survival of CRC patients. Six lncRNAs (AL356270.1, LINC02257, AC020891.2, LINC01485, AC083967.1 and RBAKDN) were finally screened out by using LASSO regression mode to establish a novel prediction model as a prognostic indicator for CRC patients. The area under the curve (AUC) of 3- and 5-year ROC analysis in CRC were 0.6923 and 0.7328 for training set, and were 0.6803 and 0.7035 for testing set, respectively. K-M analysis revealed a significant difference between high risk and low risk in the training set (P-value = 5.0e-05) and testing set (P-value = 0.00052), respectively. Our study shows that the six lncRNAs model can improve the survival prediction mechanism of patients with CRC and provide help for patients through personalized treatment. IJCEP