Predicting Outcomes in Esophageal Squamous Cell Carcinoma Using scRNA-Seq and Bulk RNA-Seq: A Model Development and Validation Study

Background: Altered glucose metabolism is a critical characteristic from the beginning stage of esophageal squamous cell carcinoma (ESCC), and the phenomenon is presented as a pink-color sign under endoscopy after iodine staining. Therefore, calculating the metabolic score based on the glucose metabolic gene sets may bring some novel insights, enabling the prediction of prognosis and the identification of treatment choices for ESCC.

Methods: A total of 8, 99, and 140 individuals from The Gene Expression Omnibus database, The Cancer Genome Atlas database, and the Memorial Sloan Kettering Cancer Center, respectively, were encompassed in the investigation. Patients diagnosed with ESCC after surgery were enrolled for further validation.

Results: A total of 13 kinds of cell clusters were screened, and the squamous epithelium was identified with the highest score. And 558 differential genes were selected from the single-cell RNA sequencing (scRNA-seq) dataset. Four glucose metabolism-related genes, namely, SERP1, CTSC, RAP2B, and SSR4, were identified as hub genes to develop a risk prognostic model. The model was validated in another external cohort. According to the risk score (RS) determined by the model, the patients were categorized into low- and high-risk groups (LRG and HRG). Compared with LRG, HRG indicated poor survival and decreased drug sensitivity. Additionally, the immune microenvironment and pathway enrichment were different between the two groups. Immunohistochemical staining revealed that hub genes were expressed differently in ESCC tissues, high- and low-grade intraepithelial neoplasia, and adjacent normal tissues.

Conclusion: Four hub genes (SERP1, CTSC, RAP2B, and SSR4) screened based on glucose metabolism developed a predictive model in ESCC patients. The RS was established as an independent risk factor for predicting prognosis. These findings may enhance understanding of ESCC's molecular profile and serve as a new prognostic tool for better patient stratification and treatment planning in clinical practice.