Development and validation of a pathomics model to predict SOX11 expression and prognosis in hepatocellular carcinoma

SOX11 is overexpressed in hepatocellular carcinoma (HCC), in which it promotes tumorigenesis and is associated with poor prognosis. This study aims to assess the prognostic value of SOX11 expression in HCC patients and to develop a predictive pathomics model based on this expression.

We analyzed data from 335 HCC patients in The Cancer Genome Atlas (TCGA) database, including pathological images and gene expression profiles. Patients were stratified into high and low SOX11 expression groups based on RNA sequencing data for prognostic and survival analysis. A pathomics model was constructed using the light gradient boosting machine (LightGBM) algorithm. Its predictive performance was evaluated in terms of discrimination, calibration, clinical utility, and interpretability. Potential pathological mechanisms were investigated through differential gene expression, enrichment analysis, immune cell infiltration profiling, and gene mutation analysis.

SOX11 expression was significantly elevated in tumor tissues compared to normal tissues. High SOX11 expression was significantly associated with reduced overall survival (OS) in the HCC population, and this association remained significant after adjusting for multiple clinical covariates. Nine pathomics features were selected to construct a pathomics signature (PS) for predicting SOX11 expression. The PS demonstrated strong discriminative ability, with area under the curve (AUC) of 0.928 [95% confidence interval (CI): 0.895-0.959] in the training cohort, the mean AUC of the model after 5-fold cross-validation was 0.869 (95% CI: 0.824-0.895), and 0.830 (95% CI: 0.699-0.879) in the validation cohort. It also showed good calibration across all cohorts, and decision curve analysis (DCA) confirmed its clinical utility. Kaplan-Meier analysis indicated that high PS expression was associated with poorer OS.

SOX11 expression may influence the prognosis of HCC patients. The PS can accurately and robustly predict both SOX11 expression and patient prognosis, making it a potentially valuable tool for clinical application.