Factors associated with deep submucosal infiltration in superficial colorectal carcinoma and development of a prediction nomogram: a retrospective study

The global burden of colorectal cancer(CRC)is rising. Early diagnosis and treatment are crucial for improving survival. Although advances in endoscopic techniques have improved the evaluation of tumor characteristics and invasion depth, no single method can accurately determine lesion depth in all cases. Therefore, this study aims to develop a more accurate approach for identifying patients suitable for endoscopic treatment. A total of 339 patients with 387 superficial CRC lesions were enrolled at Liaocheng Hospital, China. Lesions were classified as superficial cancer (high-grade dysplasia, mucosal carcinoma, or submucosal invasion < 1000 μm) or deep submucosal cancer (> 1000 μm). Demographic, tumor, treatment, and complication data were collected. For patients with multiple lesions, the single most advanced lesion (the index lesion), prioritizing depth of invasion and tumor size, was selected for analysis. Associations between potential risk factors and invasion depth were assessed using univariate and multivariate logistic regression analyses. A nomogram was then developed to construct and validate a predictive model for deep submucosal invasion. Among the 339 patients, altered bowel habits (25.37%) and abdominal pain (22.42%) were the most frequently reported clinical symptoms. Significant differences in tumor size, gross morphology, JNET classification, and pit pattern classification were observed between the superficial cancer group and the deep submucosal cancer group ( P < 0.05). Multivariate logistic regression analysis demonstrated that a tumor diameter > 20 mm ( P = 0.001, OR = 8.652, 95% CI: 2.415–30.981), an infiltrative JNET pattern ( P < 0.001, OR = 6.843, 95% CI: 2.215–21.142), and an infiltrative pit pattern ( P = 0.002, OR = 7.125, 95% CI: 1.985–25.568) were independently associated with deep submucosal invasion in superficial CRC. A nomogram incorporating these factors was developed as an exploratory predictive model. Lesions displaying a tumor diameter > 20 mm, infiltrative JNET, and infiltrative Pit pattern are strongly associated with deep submucosal infiltration in superficial colorectal carcinoma. The exploratory nomogram incorporating these three factors shows potential as a non-invasive screening tool; however, caution should be exercised, and external validation is required before broad clinical application.