34: An Artificial Intelligence Approach to Determining Optimal Margin Width in Patients Undergoing Hepatectomy for Intrahepatic Cholangiocarcinoma

Friday, March 24, 2023

2:46 PM – 2:58 PM

Location: Veterans Memorial Auditorium

Author(s)

LA

Laura Alaimo, MD

Research Scholar
Department of Surgery, The Ohio State University Wexner Medical Center and James Comprehensive Cancer Center, United States

Disclosure information not submitted.
ZM

Zorays Moazzam, MD

Research Scholar
Department of Surgery, The Ohio State University Wexner Medical Center and James Comprehensive Cancer Center, United States

Disclosure information not submitted.
HA

Henrique Araujo Lima, MD

Research Scholar
Department of Surgery, The Ohio State University Wexner Medical Center and James Comprehensive Cancer Center, United States

Disclosure information not submitted.
Yutaka Endo, MD, PhD

Research Scholar
Department of Surgery, The Ohio State University Wexner Medical Center and James Comprehensive Cancer Center, United States

Disclosure(s): No financial relationships to disclose
AR

Andrea Ruzzenente, MD, PhD

Professor
Department of Surgery, Dentistry, Gynecology, and Pediatrics, Division of General and Hepato-Biliary Surgery, University of Verona, University Hospital G.B. Rossi, Verona, Italy
Verona, Veneto, Italy

Disclosure information not submitted.
AG

Alfredo Guglielmi, MD

Professor
Department of Surgery, Dentistry, Gynecology, and Pediatrics, Division of General and Hepato-Biliary Surgery, University of Verona, University Hospital G.B. Rossi, Verona, Italy
Verona, Veneto, Italy

Disclosure information not submitted.
Aslam Ejaz, MD, MPH

Assistant Professor of Surgery
The Ohio State University

Disclosure information not submitted.
Jordan M. Cloyd, MD

Assistant Professor
The Ohio State University
Columbus, OH, United States

Disclosure information not submitted.
Joal D. Beane, MD

Assistant Professor
The Ohio State University Wexner Medical Center- Division of Surgical Oncology
Columbus, Ohio, United States

Disclosure information not submitted.
Alex C. Kim, MD, PhD

Assistant Professor of Surgery
The Ohio State University Wexner Medical Center, The James Cancer Hospital and Solove Research Institute
Columbus, Ohio, United States

Disclosure information not submitted.
Timothy M. Pawlik, MD, PhD

Chair of Surgery
The Ohio State University Wexner Medical Center
Columbus, Ohio, United States

Disclosure(s): No financial relationships to disclose

Abstract Presenter(s)

Yutaka Endo, MD, PhD

Research Scholar
Department of Surgery, The Ohio State University Wexner Medical Center and James Comprehensive Cancer Center, United States

Disclosure(s): No financial relationships to disclose

Disclosure(s):

Yutaka Endo, MD, PhD: No financial relationships to disclose

Timothy M. Pawlik, MD, PhD: No financial relationships to disclose

Introduction: Intrahepatic cholangiocarcinoma (ICC) is generally associated with high rates of recurrence. While the recommended curative-intent treatment of ICC is complete liver resection with loco-regional lymphadenectomy, there is poor evidence to establish the optimal resection margin width. We sought to identify using artificial intelligence the optimal margin width at the time of hepatectomy for ICC to maximize recurrence-free survival (RFS).

Methods: Patients who underwent curative-intent resection for ICC from 2000-2020 were identified from an international multi-institutional database. A Random Forest (RF) regression model was used to estimate the counterfactual probabilities associated with the maximal decrease in risk of recurrence relative to each margin width. These estimations were used to train a globally optimal policy tree (OPT) that allowed assigning different treatment recommendations to subgroups of patients. Patient, tumor, and treatment-related characteristics were included in the model. The goodness of fit of the models was assessed by calculating Harrell’s concordance index (c-index).

Results: Among 926 patients who underwent curative-intent resection of ICC, median tumor size was 6.0 cm (IQR 4.0-8.5) and 158 (17.1%) patients had multifocal disease. A total of 528 (57.0%) patients underwent major hepatectomy, while lymphadenectomy was performed in 453 (48.9%) patients. Median resection margin was 4.0 mm (IQR 1.0-10.0). Of note, 556 (60.0%) patients experienced recurrence with a 3-year RFS of 31.9%. Overall, 651 (70.3%) patients were randomly assigned to the training cohort, while 275 (29.7%) were included in the testing cohort. The RF model demonstrated good performance to predict counterfactual probabilities of death for each value of margin width from 0 to 20 mm (c-index: training 0.65; testing 0.60). Specifically, the OPT suggested 4 different acceptable optimal margin widths (i.e., 1.0, 10.0, 16.0, and 19.0 mm) associated with the highest 3-year RFS probability among different subgroups of patients based on age, tumor size, and treatment characteristics (i.e., bile duct resection, lymphadenectomy, extent of resection) (Figure).

Conclusions: An artificial intelligence methodology identified different acceptable optimal resection margin widths for various subgroups of patients with ICC. Rather than one standard cut-off, variable margin widths based on individualized patient and disease characteristics may optimize long-term outcomes for patients with ICC.

Learning Objectives:

Assess the differences in long-term outcomes of patients who undergo hepatectomy for Intrahepatic Cholangiocarcinoma relative to the resection margin width achieved.
Prescribe the individual optimal resection margin width to maximize the probability of 3-year recurrence-free survival for Intrahepatic Cholangiocarcinoma.
Recognize that resection margin-width should be individualized based on underlying patient and disease characteristics.