An Autophagy-Related Long Non-Coding RNA Prognostic Signature for Patients with Lung Squamous Carcinoma Based on Bioinformatics Analysis

Purpose: Lung cancer is the most common and deadly cancer type affecting humans. Although huge progress has been made on early diagnosis and precision treatment, the overall 5 year survival rate remains low. In this study, we constructed an autophagyrelated long non-coding RNA (lncRNA) prognostic signature for guiding clinical practice. Methods: From The Cancer Genome Atlas, we retrieved mRNA and lncRNA expression matrices of patients with lung squamous carcinoma. We then established a prognostic risk model using Lasso regression and multivariate Cox regression. The model generated a risk score to differentiate high-and low-risk groups. An ROC curve and nomogram were used to visualize the predictive ability of the current signatures. Finally, we used Gene Set Enrichment Analysis to determine gene ontology and pathway enrichment. Results: After screening 1248 autophagy-related lncRNAs, we selected seven lncRNAs (LUCAT1, AC022150.2, AL035425.3, AC138976.2, AC106786.1, GPRC5D-AS1 and AP006545.2) for our signature. Univariate (hazard ratio [HR] = 2.147, 95% confidence interval [CI]: 1.681-2.743, P < 0.001) and multivariate (HR = 2.096, 95% CI: 1.652- 2.658, P < 0.001) Cox regression analyses revealed that the risk score is an independent predictive factor for LUSC patients. Further, areas under the receiver operating characteristic curve were 0.622, 0.699, and 0.721, respectively, for the 1 year, 3 year, and 5 year risk scores-indicating a reliable model. Selected lncRNAs were primarily enriched in autophagy, metabolism, MAPK pathway, and JAK/STAT pathway. Further drug sensitivity analysis revealed that low-risk patients were more sensitive to Cisplatin, Docetaxel, Vinblastine, and Vinorelbine. Finally, a multiomics analysis found that lncRNA-linked proteins IKBKB and SQSTM1 were expressed at low levels and significantly correlated in tumor samples, compared with normal tissue. Conclusion: Our prognostic model successfully predicted patient prognosis in lung cancer.