Comparative Study on Efficacy and Risk of Pathological Upgrading between Endoscopic and Surgical Resection for Colorectal High-Grade Intraepithelial Neoplasia: A Retrospective Analysis

Background Accurate management of colorectal high-grade intraepithelial neoplasia (HGIN) is challenging due to the limited accuracy of diagnostic endoscopic biopsy and the difficulty in choosing between endoscopic therapy and radical surgery. Methods We retrospectively analyzed 69 patients with preoperative biopsy-diagnosed colorectal HGIN (2019–2023). Data included demographics, lesion features, and treatments, focusing on preoperative-postoperative pathological concordance. Results Fifty-six and thirteen patients underwent endoscopic resection (ER) and additional surgery resection (ASR), respectively. The baseline characteristics were balanced between the two groups, with no statistically significant differences observed in age, sex, clinical presentation, or lesion number (P > 0.05). Complication rates were similar (16.1% vs. 15.4%, P = 0.723). The biopsy accuracy was 81.2%; 18.8% of lesions were upgraded to T1 colorectal cancer (CRC). Lesions in the ASR group were larger than in the ER group (2.2 ± 0.51 cm vs. 1.9 ± 0.43 cm, P = 0.015). T1 lesions were larger than HGIN lesions (P = 0.002). Rectosigmoid location (P = 0.032) and solitary lesions (carcinoma rate: ASR 62.5% vs. ER 4.76%; OR = 33.33, P < 0.001) conferred higher cancer risk. For solitary lesions, the carcinoma rate was significantly higher in the ASR group than in the ER group (62.5% vs. 4.76%, OR = 33.33, P < 0.001). Although sessile morphology was associated with the decision for surgical intervention (P = 0.003), its efficacy in discriminating between HGIN and T1 carcinoma was limited (P = 0.057). Lymph node metastasis (LNM) occurred only in T1 patients (27.3%; 28.6% in the ASR). High-grade tumor budding (Bd > 1) and lymph vascular invasion (LVI) were enriched in the ASR group (OR = 43.33, P < 0.001) and strongly associated with T1 cancer (P < 0.001). The R0 resection rate was comparable between HGIN and T1 lesions (P = 0.264). Conclusion Improving endoscopic biopsy accuracy is essential. ER is safe and feasible, providing effective local control for HGIN and low-risk T1 CRC without LNM.