Background: Endoscopic optical diagnosis is crucial to the therapeutic strategy in early gastrointestinal cancer. It accurately (> 85%) predicts pT category based on micro surface (SP) and vascular patterns (VP). However, inter-observer variability is a major problem. We have visualized and digitalized the graded irregularities based on bioinformatically enhanced quantitative endoscopic image analysis (BEE) of high definition white light images.
Methods: In a pilot study on 26 large colorectal lesions (LCL, mean diameter 39 mm), we retrospectively compared BEE variables with corresponding histopathology of the resected LCL.
Results: We included ten adenomas with low grade intraepithelial neoplasia (LGIN), nine with high grade intraepithelial neoplasia and early adenocarcinoma (HGIN and EAC), and seven deeply submucosal invasive carcinomas. Quantified density (d) and non-uniformity (CU) of vascular and surface structures correlated with histology (rs d VP: -0.77, rs CU VP: 0.13, rs d SP: -0.76, and rs CU SP: 0.45, respectively). A computed BEE-score showed a sensitivity and specifity of 90% and 100% in the group of LGIN, 89% and 41% in the group of HGIN and EAC, and 100% and 95% in the group of deeply invasive carcinoma, respectively.
Conclusions: In this pilot study, BEE has shown promise as a tool for endoscopic characterization of LCL during routine endoscopy. Prospective clinical studies are needed.