Serrated polyps <10 mm cannot reliably be characterized by i-Scan without magnification at routine colonoscopy.

Abstract

Background: Colorectal lesions (CRLs) <10 mm found at colonoscopy tend towards "diagnose-and-leave" or "resect-and-discard" strategies based on real-time Kudo glandular pit-pattern's assessment using i-Scan. However, i-Scan has not yet been validated for Kudo's classification. We aimed to assess whether, in routine colonoscopy, i-Scan without magnification and optical enhancement (M-OE) reliably differentiates hyperplastic polyps (HPs) from other serrated lesions (SLs) and conventional adenomas (CAs), and, among SLs, HPs from sessile serrated lesions (SSLs) and traditional or unknown serrated adenomas (TSAs, USAs), in Kudo type II CRLs<10 mm, according to ASGE Preservation and Incorporation of Valuable endoscopic Innovations (PIVI) recommended negative predictive value (NPV) threshold for adenomas.

Methods: Prospectively recorded CRLs over 12 months, classified according to Kudo pit-pattern using i-Scan, were retrospectively compared with histology.

Results: Overall, 898 ≤5-mm and 704 6- to 9-mm CRLs were included. Type II pit-pattern was found in 76.6% and 38.7% of HPs and SSLs-TSAs/CAs (P<0.000001), and in 84.1% and 26.6% of SLs and CAs (P<0.000001). Among SLs, it was found in 81.9% and 86.6% of HPs and SSLs-TSAs. In CRLs≤5 mm, HPs were prevalent over other SLs (P=0.00001); in CRLs 6-9 mm, CAs were prevalent (P<0.000001). About 77% of SLs in right colon were SSLs-TSAs; 82% in left colon were HPs. PIVI ≥90% NPV threshold for adenomas was reached for CRLs 6-9mm (92.1%), nearly achieved for CRLs≤5 mm (88.2%), and not reached for SLs independently on the size.

Conclusions: A strategy of "diagnose-and-leave" or "resect-and-discard" cannot be recommended for SLs<10 mm with Kudo type II pit-pattern using i-Scan, especially in right colon, if M-OE unavailable.