Abstract
BACKGROUND: Ultra-low rectal endoscopic submucosal dissection (ESD) presents technical challenges due to anatomical features. The objective of this research was to determine the risk factors linked to unsuccessful curative resections and to create a nomogram predictive model to assess the likelihood of encountering technical challenges. METHODS: Patients with ultra-low rectal tumors received ESD form June 2017 to December 2022 were retrospectively enrolled. An ESD procedure exceeding 30 min was deemed difficult. A logistic regression analysis was performed to pinpoint important factors and predictors. The effectiveness of the nomogram, which incorporated the identified predictors, was evaluated by employing receiver operating characteristic (ROC) curves, calibration plots, and decision curve analysis (DCA). RESULTS: A total of 300 patients with ultra-low rectal tumors were enrolled, with a curative resection rate of 82.0%. Multivariate logistic regression revealed that poor lifting sign (OR = 3.282, P = 0.026), non-granular type laterally spreading tumors (LST-NG, OR = 2.230, P = 0.042) and procedure time ≥ 60 min (OR = 6.976, P = 0.010) contributed to non-curative resection. Predictors for ESD difficulty included tumor diameter ≥ 30 mm (compared with < 30 mm, 30-50 mm, OR = 2.450, P = 0.044; ≥ 50 mm, OR = 5.047, P = 0.009), ≥ 1/2 circumference involvement (OR = 3.183, P = 0.038); dentate line invasion (OR = 3.881, P = 0.026) and less colorectal ESD experience (OR = 3.415, P = 0.032). The nomogram performed well in both train and validation sets (area under the curve (AUC) = 0.873 and 0.810, respectively). Calibration plots exhibited satisfactory agreement between predicted and observed outcomes, and DCA showed superior clinical benefit of the model than individual predictors. CONCLUSIONS: Poor lifting sign, LST-NG and procedure time ≥ 60 min were associated with non-curative resection for ultra-low rectal ESD. By including factors such as tumor size, location, and the operator's experience with ESD, the nomogram can predict the complexity of the procedure before surgery.