Abstract
BACKGROUND: The prognosis of non-small cell lung cancer (NSCLC) patients has been comprehensively studied. However, the prognosis of resectable (stage I-IIIA) lung squamous cell carcinoma (LUSC) has not been thoroughly investigated at genomic and transcriptional levels. METHODS: Data of genomic alterations and transcriptional-level changes of 355 stage I-IIIA LUSC patients were downloaded from The Cancer Genome Atlas (TCGA) database, together with the clinicopathological information (training cohort). A validation cohort of 91 patients was retrospectively recruited. Data were analyzed and figures were plotted using the R software. RESULTS: Training cohort was established with 355 patients. TP53 (78%), TTN (68%), CSMD3 (39%), MUT16 (36%) and RYR2 (36%) were genes with the highest mutational frequency. BRINP3, COL11A1, GRIN2B, MUC5B, NLRP3 and TENM3 exhibited significant higher mutational frequency in stage III (P < 0.05). Patients with stage III also exhibited significantly higher tumor mutational burden (TMB) than those with stage I (P < 0.01). The mutational status of 10 genes were found to have significant stratification on patient prognosis. TMB at threshold of 25 percentile (TMB = 2.39 muts/Mb) also significantly stratified the patient prognosis (P = 0.0003). Univariate and multivariate analyses revealed TTN, ADGRB3, MYH7 and MYH15 mutational status and TMB as independent risk factors. Further analysis of transcriptional profile revealed many significantly up- and down-regulated genes, and multivariate analysis found the transcriptional levels of seven genes as independent risk factors. Significant factors from the multivariate analyses were used to establish a Nomogram model to quantify the risk in prognosis of individual LUSC patients. The model was validated with a cohort containing 91 patients, which showed good predicting efficacy and consistency. CONCLUSION: The influencing factors of prognosis of stage I-III LUSC patients have been revealed. Risk factors including gender, T stage, cancer location, and the mutational and transcriptional status of several genes were used to establish a Nomogram model to assess the patient prognosis. Subsequent validation proved its effectiveness.