Pine-cone and villi patterns are endoscopic signs suggestive of ulcerative colitis-associated colorectal cancer and dysplasia.

ABSTRACT

BACKGROUND AND AIMS: The appropriate site for targeted biopsy during surveillance colonoscopy for ulcerative colitis (UC) is still unclear. We aimed to clarify key endoscopic findings suggestive of neoplastic lesions for targeted biopsy in UC. METHODS: First, we created 769 stereomicroscopic pictures (509 neoplastic, 260 non-neoplastic) mimicking magnifying colonoscopic images from surgically resected specimens, including areas surrounding 25 neoplastic lesions in 15 patients with colitis-associated cancer at a single referral center. Second, we validated the results by using 113 magnifying endoscopic images (64 neoplastic, 49 non-neoplastic) from 39 lesions in 26 patients. Two evaluators, blinded to the pathologic diagnosis, independently classified them according to Kudo's pit pattern and surface morphology, such as pine-cone/villi patterns. The correlation between stereomicroscopic and pathologic findings (neoplastic vs non-neoplastic) for each image was investigated. The interobserver agreement was assessed using kappa statistics. RESULTS: In the stereomicroscopic analysis, neoplastic pit patterns (types III-V) were significantly correlated with the presence of neoplasia (sensitivity 77.4%, specificity 89.5%, kappa value 0.677). Pine-cone/villi patterns also showed high specificity (96.8%) but low sensitivity (21.4%, kappa value 0.625) for neoplasia. Endoscopic validation showed similar trends. A revision of the endoscopic findings of flat dysplasia with non-neoplastic pit patterns revealed that a reddish area may facilitate the identification of such lesions. CONCLUSIONS: Targeted biopsies are recommended, especially for lesions showing pine-cone/villi patterns in addition to neoplastic pit patterns. For flat "non-neoplastic pit patterns," a reddish area may be an indication for a biopsy.