Clinical Scorecard: Comparing Three Modeling Approaches to Predict Resection Zones for Parenchyma-Sparing Hepatectomy Planning
At a Glance
Category
Detail
Condition
Preoperative planning of liver resection for primary hepatic tumors.
Key Mechanisms
The study compares distance-based, perfusion-based, and deep-learning approaches to predict resection zones using varying anatomical and functional inputs.
Target Population
Retrospective analysis of 104 patients and 104 resection zones selected from a source dataset of 312 patients undergoing 338 non-laparoscopic liver tumor resections; procedures were predominantly for HCC or ICC.
Care Setting
Retrospective study using data from the University Medical Center Mainz’s Clinic for General, Visceral, and Transplant Surgery.
Key Highlights
The study compared distance-based, perfusion-based, and deep-learning approaches for predicting parenchyma-sparing resection zones.
The retrospective dataset included 104 patients and 104 reference resection zones.
The distance-based model used liver boundaries and tumor location, with an 8 mm dilation based on the cohort’s mean minimal resection margin.
The perfusion-based model incorporated liver, tumor, and vascular structures to estimate perfusion loss after vessel transection.
Predicted and reference zones were assessed using overlap and distance measures, including DSC, IoU, HD95, maximum Hausdorff distance, and ASD.
Guideline-Based Recommendations
Diagnosis
The article addresses surgical planning and model evaluation, not diagnostic criteria for liver tumors.
Management
The introduction describes parenchyma-sparing surgery as limiting the volume of liver removed while preserving oncological safety.
The introduction states that preoperative planning should account for tumor location, hepatic perfusion, biliary drainage, and future liver remnant.
Monitoring & Follow-up
The study compared predicted and retrospectively defined resection zones using DSC, IoU, HD95, maximum Hausdorff distance, and ASD.
Risks
The introduction states that an adequate future liver remnant is needed to avoid postoperative liver failure: at least 20–30% in healthy livers and at least 40% when hepatic reserve is impaired.
The source notes that vessel transections can cause downstream perfusion loss.
Patient & Prescribing Data
Retrospective cohort of 104 patients selected from a source dataset of 312 patients undergoing 338 non-laparoscopic liver tumor resections; procedures were predominantly for hepatocellular carcinoma or intrahepatic cholangiocarcinoma.
No medication or prescribing intervention was evaluated. The study examined computational prediction of resection zones for surgical planning.
Clinical Best Practices
The introduction states that preoperative planning should account for tumor location, hepatic perfusion, biliary drainage, and future liver remnant.
The introduction identifies preservation of an adequate future liver remnant as necessary to avoid postoperative liver failure.
For this study, semiautomatically generated masks covered the future liver remnant, tumor, portal and hepatic veins, inferior vena cava, and retrospectively defined resection zones.
Harold Burstein, MD, PhD, and Ana C. Garrido-Castro, MD discuss new antibody drug conjugate treatment options in first line triple negative breast cancer.
