A new endoscopic tumor grading for rectal neuroendocrine tumors: Correlation of vascular pattern with histopathology

Introduction: The clinical utility of applying endoscopic vascular patterns in rectal neuroendocrine tumor (NET) remains unknown. The aims of this study were to develop a system for utilizing vascular patterns for tumor grade and submucosal invasion depth.

Methods: We retrospectively included patients diagnosed as well-differentiated rectal NET between March 2015 and February 2024. The entire dataset was divided into evaluation and validation sets. In the evaluation set, the relationship between endoscopic tumor features and tumor grade was explored using a multivariable regression model. Based on the endoscopic vessel characteristics, vascular patterns were established and classified into V1 and V2 types. In the validation set, 104 lesions were used to assess the diagnostic performance of the vascular patterns.

Results: Multivariate analysis demonstrated good association between tumor grade and vascular pattern (13.65 odds ratio; 95 % confidence interval, 2.06-90.58). Vascular patterns exhibited almost perfect inter- and intra-observer agreement (kappa = 0.957). The sensitivity, positive predictive value and accuracy of V1 for predicting grade 1 were 94.5 %, 97.7 and 93.3 %, respectively. Histopathological analysis revealed significantly deeper submucosal invasion depth for V2 than that for V1 lesions (p < 0.001).

Conclusions: Vascular patterns can help identify tumor grade and assess submucosal invasion depth.