Disagreement between high confidence endoscopic adenoma prediction and histopathological diagnosis in colonic lesions ≤ 3 mm in size

Abstract

Abstract Background Diminutive colorectal polyps resected during colonoscopy are sometimes histologically interpreted as normal tissue. The aim of this observational study was to explore whether errors in specimen handling or processing account in part for polyps ≤ 3 mm in size being interpreted as normal tissue by pathology when they were considered high confidence adenomas by an experienced endoscopist at colonoscopy. Methods One endoscopist photographed 900 consecutive colorectal lesions that were ≤ 3 mm in size and considered endoscopically to be high confidence conventional adenomas. The photographs were reviewed blindly to eliminate poor quality images. The remaining 644 endoscopy images were reviewed by two external experts who predicted the histology while blinded to the pathology results. Results Of 644 consecutive lesions ≤ 3 mm in size considered high confidence conventional adenomas by a single experienced colonoscopist, 15.4 % were reported as normal mucosa by pathology. The prevalence of reports of normal mucosa in polyps removed by cold snare and cold forceps were 15.2 % and 16.0 %, respectively. When endoscopy photographs were reviewed by two blinded outside experts, the lesions found pathologically to be adenomas and normal mucosa were interpreted as high confidence adenomas by endoscopic appearance in 96.9 % and 93.9 %, respectively, by Expert 1 (P = 0.15), and in 99.6 % and 100 %, respectively, by Expert 2 (P = 0.51). Conclusion Retrieval and/or processing of tissue specimens of tiny colorectal polyps resulted in some lesions being diagnosed as normal tissue by pathology despite being considered endoscopically to be high confidence adenomas. These findings suggest that pathology interpretation is not a gold standard for lesion management when this phenomenon is observed.