AI-Driven 3D Liver Modeling Enhances Prediction of Liver Resection Complexity from Preoperative CT Scans

Highlight

– Development of an AI-based automatic 3D liver modeling pipeline that integrates organ, tumor, and blood vessel segmentation from preoperative CT scans.
– Introduction of the Hepatic Central Zone (HCZ) as a novel anatomic frame of reference for assessing liver resection complexity.
– The AI model predicts intraoperative liver resection complexity (LRC) with higher accuracy and AUC than experienced surgeons relying solely on CT imaging.
– This novel AI tool holds promise for enhancing surgical planning, optimizing patient referral, and potentially reducing morbidity and mortality in liver cancer surgery.

Study Background