Endoscopic detection of esophageal low-grade squamous dysplasia: How to predict pathologic upgrades before treatment?

ABSTRACT

OBJECTIVE: In this study we aimed to predict the risk factors related to histopathologic upgrade after endoscopic submucosal dissection (ESD) in patients with pre-ESD esophageal squamous low-grade intraepithelial neoplasm (LGIN). METHODS: A training cohort of 201 patients with biopsy-confirmed esophageal squamous LGIN and underwent ESD at a tertiary medical center between January 2017 and July 2019 were included. Risk factors for histological upgrade were identified using the least absolute shrinkage and selection operator (LASSO) regression. A nomogram was then established. Internal validation was evaluated by discrimination, calibration plot, and decision-curve analysis. Another cohort of 48 patients were prospectively collected from July 2019 to June 2021 for external validation of the nomogram. RESULTS: The rate of histological upgrade was 34.8% (70/201) and 27.1% (13/48) in the training and validation sets, respectively. LASSO regression identified that tumor area (mm2 ) per biopsy, Lugol's staining pattern, background coloration, and the circumferential range of the lesion were significantly associated with histological upgrade. The final nomogram attained favorable prediction efficacy in the training cohort (area under the receiver operating curve [AUROC] 0.96, 95% confidence interval [CI] 0.94-0.98) and validation cohort (AUROC 0.92, 95% CI 0.79 -0.99). This model generated well-fitted calibration and clinical-decision curves in both cohorts. CONCLUSIONS: The nomogram may better guide clinical decision on whether performing EDS or follow-up for suspicious lesions in patients with biopsy-confirmed esophageal squamous LGIN.