Construction and validation of a prognostic model for tongue cancer based on three genes signature

Tongue squamous cell carcinoma (TSCC) has a poor prognosis and destructive characteristics. Reliable biomarkers are urgently required to predict disease outcomes and to guide TSCC treatment. This study aimed to develop a multigene signature and prognostic nomogram that can accurately predict the prognosis of patients with TSCC. We screened differentially expressed genes associated with TSCC using The Cancer Genome Atlas dataset. Based on this, we developed a new multi-mRNA gene signature using univariate Cox regression, Least Absolute Shrinkage and Selection Operator regression, and multivariate Cox regression. We used the concordance index to evaluate the accuracy of this new multigene model. Moreover, we performed receiver operating characteristic and Kaplan-Meier survival analyses to assess the predictive ability of the new multigene model. In addition, we created a prognostic nomogram incorporating clinical and pathological characteristics, with the aim of enhancing the adaptability of this model in practical clinical settings. We successfully developed a new prognostic model based on the expression levels of these 3 mRNAs that can be used to predict the prognosis of patients with TSCC. This prediction model includes 3 genes: KRT33B, CDKN2A, and CA9. In the validation set, the concordance index of this model was 0.851, and the area under the curve was 0.778 and 0.821 in the training and validation sets, respectively. Kaplan-Meier survival analysis showed that regardless of whether it was in the training or validation set, the prognosis of high-risk patients was significantly worse than that of low-risk patients (P < .001). Multivariate Cox regression analysis revealed that this model was an independent prognostic factor for patients with TSCC (P < .001). Our study suggests that this 3-gene signature model has a high level of accuracy and predictive ability, is closely related to the overall survival rate of patients with TSCC, and can independently predict the prognosis of TSCC patients with high accuracy and predictive ability.