Diagnostic Accuracy of Referral Biopsy Compared to Optical Biopsy in Large Non-pedunculated Colorectal Polyps

Background Endoscopic mucosal resection (EMR) of large non-pedunculated colorectal polyps (LNPCPs) offers excellent efficacy and safety. Referral biopsies are commonly obtained prior to EMR despite their potential adverse effects. We aimed to assess the diagnostic accuracy of referral biopsy as compared with optical evaluations and the polyp's final pathology. Methods We analyzed a cohort from two centers that included adult patients referred for EMR of LNPCPs between 2017 and 2022. The NICE classification system was used for optical classification. We compared procedural outcomes of lesions with or without a referral biopsy and diagnostic accuracy of referral biopsy and optical evaluation to the final histopathology of the resected polyp. Results Of 605 EMR procedures, 398 (65.8%) had referral biopsies. Polyp size was larger in the biopsy group. No significant differences were observed in en-bloc resection rate, adequate lifting, procedural bleeding, or surveillance recurrence rates. Optical biopsies had higher diagnostic concordance with final histology (Cohen's Kappa 0.62 vs. 0.55) and enhanced sensitivity for sessile serrated polyps/hyperplastic polyps (SSP/HP) and cancer as compared to referral biopsies (0.75 vs. 0.72, p < 0.01, and 0.19 vs. 0.0, p < 0.01, respectively). Optical evaluation was more accurate in non-biopsied lesions. Conclusion Referral biopsies do not offer additional diagnostic accuracy and may reduce the accuracy of optical evaluation for LNPCPs.