P-0003 Quantitative Tissue Phenotype (QTP) as a New Diagnostic Test in Barrett’s Esophagus Dysplasia

Abstract

ABSTRACT Introduction Barrett’s esophagus (BE), defined by intestinal metaplasia in squamous-lined esophagus, is a precursor lesion that confers an increased risk of esophageal adenocarcinoma. Progression occurs through a metaplasia–dysplasia–carcinoma sequence. However, the risk of progression is variable and histological grading of BE dysplasia has shown significant inter-pathologist variability. QTP is a revolutionary method in the current era of digital pathology that uses high-resolution image analysis to assess nuclei both quantitatively (morphometry) and qualitatively (DNA texture). We aimed to evaluate QTP as a maker of dysplasia and risk progression in BE. Methods QTP image analysis was carried out by an in-house automated quantitative system on Feulgen-Thionin stained sections of esophageal biopsies with BE. A nuclear phenotype score (NPS) was generated initially for a training set of 10 cases each of BE with no dysplasia and invasive adenocarcinoma. NPS is a combination of nuclear morphometric features that best discriminate “normal” nuclei (selected from BE with no dysplasia) from “abnormal” nuclei (selected from BE with invasive adenocarcinoma). NPS was then determined for the test set including 10 cases each of BE with indefinite for dysplasia, low grade dysplasia, high grade dysplasia and intramucosal adenocarcinoma In the cases with indefinite and low grade dysplasia, half progressed to cancer, and half did not progress after 2 years of follow up. Results A progressive linear correlation was found between NPS and BE grades of dysplasia. The best-fit model for NPS included a combination of nuclear morphometric features (sphericity, compactness, area) with DNA textural features (DNA content, chromatin organization) and yielded the best rate of correct classification. High-grade dysplasia showed a significant increase in NPS compared with low-grade dysplasia and indefinite for dysplasia. There was a difference between NPS of classical low risk lesions that progressed to high risk lesions and those that did not. Conclusion These data suggest that QTP is a very promising platform test in digital pathology, which can aid pathologists in the grading of dysplasia, reducing the significant interobserver variability between pathologists and has the potential to facilitate the identification of BE at high risk of progression into invasive adenocarcinoma. Further larger studies are required for validation.