Year: 2025
Type: article
Abstract: Introduction Advancements in surgical techniques have expanded curative options for patients with high hepatic tumor burden. Non-anatomic resections and vascular tumor detachments enable more aggressive interventions but must be carefully balanced with the need to preserve an adequate future liver remnant (FLR). This study aimed to evaluate the utility of artificial intelligence (AI)-assisted imaging technology (Fujifilm Synapse 3D) in preoperative planning for complex liver resections. Method Consecutive patients undergoing major hepatectomy or parenchyma-sparing hepatectomy were included. Preoperative imaging was used to generate AI-assisted, patient-specific 3D liver models. Virtual simulations of the planned procedures were performed on these models, and the standardized FLR was calculated. The primary endpoint was the rate of change in surgical planning based on 3D modeling versus conventional imaging review alone. Secondary outcomes included postoperative morbidity. Result Between January 2024 and April 2025, 78 patients were enrolled: 16 underwent parenchyma-sparing hepatectomy and 62 underwent major hepatectomy. In 34 cases (44%), AI-assisted 3D modeling prompted a change in the operative plan, leading to an increased FLR. Postoperative complications classified as Clavien-Dindo grade ≥ IIIb occurred in 10% of patients. Two patients developed grade A post-hepatectomy liver failure, both of which resolved without further intervention. Discussion Patient specific AI-generated 3D liver models are a valuable tool in the planning of complex hepatic resections, enabling safer surgeries through enhanced visualization and improved preservation of liver parenchyma.
Source: British journal of surgery
Authors: Srinivas Sanjeevi, Farima Brandt, Jozef Urdzik
Institution: Uppsala University Hospital
Language: English
Subfield: Biomedical Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: bronze
OpenAlex ID: W4413185969