Development and validation of the SDLD-score: a simplified tool to predict successful endoscopic papillectomy in ampullary lesions
| Author | Affiliation | |
|---|---|---|
Vu Trung, Kien | ||
| Date | Volume | Issue | Start Page | End Page |
|---|---|---|---|---|
2023-10-12 | 23 | S1 | 81 | 81 |
Poster no. P-08-39
Background: Endoscopic papillectomy (EP) is a standard treatment for most non-invasive ampullary lesions (AL) and early ampullary cancers but advanced lesions require surgery. However, management of AL can be challenging, depends on local expertise and may lead to over- or undertreatment. We developed a score to identify best eligible patients for endoscopic or surgical treatment.
Methods: 575 patients who underwent EP fulfilled the inclusion criteria without missing data. The cohort was split into a training set (416) and a validation set (159) by random. In the training set predictors for incomplete resection (R1) were analysed by logistic regression and incorporated to a score of four items. The score was validated in a separate independent cohort. Performance was estimated by the area-under-the-receiver-operating-characteristic-curve (AUROC).
Results: Size >30mm (odds ratio [OR] 4.3 (95%CI 1.8-10.3); p=0.001), high-grade dysplasia or invasive cancer (OR 7.1 (95%CI 3.9-13.0); p<0.001)), laterally spreading lesion (OR 3.4 (95%CI 1.4-8.5); p=0.009) and dilation of bile or pancreatic duct (OR 3.0 (95%CI 1.7-5.5); p<0.001) were identified as independent factors for incomplete resection (R1) and used to develop the SDLD score. AL not exceeding one item (0-1 points) had the highest R0-rate (training: 90.7%; validation: 89.5%). By fulfilling at least two criteria R1-rate was significantly increased (training: 55.7%; validation: 51.4%; p<0.001). The AUROC was 0.834 in the training cohort and 0.763 in the validation cohort.
Conclusion: The SDLD score accurately predicts incomplete resection in patients undergoing EP and may help to optimize the decision process for endoscopic or surgical resection of AL.