Validation of the parkland grading scale in predicting the critical view of safety during laparoscopic cholecystectomy: A prospective cohort study with implications for future artificial intelligence ground truth establishment
- Author:
Hoai Kim NGUYEN
1
;
Thien Lai VO
;
Ho TRAN
;
Thanh Phuoc BUI
;
Xuan Binh DAU
;
Phuoc Cong Thanh NGUYEN
Author Information
- Publication Type:Original Article
- From: Annals of Hepato-Biliary-Pancreatic Surgery 2026;30(2):212-219
- CountryRepublic of Korea
- Language:English
-
Abstract:
Background:s/Aims: The critical view of safety (CVS) prevents bile duct injury, but severe inflammation hinders its achievement.This study evaluated the parkland grading scale (PGS) as a real-time intraoperative predictor of CVS attainment in a Southeast Asian population, establishing standardized clinical data for future artificial intelligence applications.
Methods:This prospective observational study (January–June 2025) included 88 consecutive patients undergoing laparoscopic cholecystectomy by a single surgeon. PGS was assessed upon initial laparoscopic inspection. We analyzed correlations between PGS, CVS attainment (Strasberg’s criteria), operative time, and bail-out procedures using Spearman’s correlation and multivariate logistic regression.
Results:Severe inflammation (PGS grade 4–5) was observed in 31.8% of patients. PGS exhibited a strong negative correlation with CVS score (p = −0.652; p < 0.001) and strong positive correlations with operative time and blood loss. A PGS threshold of 4 accurately predicted CVS failure (area under the curve = 0.863). Bail-out procedures were necessary in 11 cases (12.5%), all occurring in the PGS ≥ 4 group. Advanced age and diabetes mellitus were independent risk factors for CVS failure.
Conclusions:The PGS serves as a precise intraoperative early warning system. A score ≥ 4 indicates a significant risk of CVS failure, prompting safer bail-out strategies. Additionally, this study provides a standardized dataset vital for training future autonomous surgical risk assessment models.
