A. Altaf1, M. Khalil1, Z. Rashid1, J. Kawashima1, S. Zindani1, A. Guglielmi2, L. Aldrighetti3, T.W. Bauer4, H.P. Marques5, G. Martel6, S. Alexandrescu7, M.J. Weiss8, M. Kitago9, G. Poultsides10, S.K. Maithel11, C. Pulitano12, F. Shen13, F. Cauchy14, B.G. Koerkamp15, I. Endo16, T.M. Pawlik1 1Ohio State University, Columbus, OH, USA 2University of Verona, Verona, PROVINCE OF VERONA, Italy 3San Raffaele Hospital, Milan, MILANO, Italy 4University Of Virginia, Charlottesville, VA, Virgin Islands, U.S. 5Curry Cabral Hospital, Lisbon, LISBON, Portugal 6University Of Ottawa, Ottawa, Ontario, Canada 7Fundeni Clinical Institute, Bucharest, BUCHAREST, Romania 8Johns Hopkins University School Of Medicine, Baltimore, MD, USA 9Keio University, Tokyo, Japan 10Stanford University, Palo Alto, CA, USA 11Emory University School Of Medicine, Atlanta, GA, USA 12University Of Sydney, Sydney, NSW, Australia 13Eastern Hepatobiliary Surgery Hospital, Shanghai, China 14AP-HP, Beaujon Hospital, Clichy, France 15Erasmus University Medical Centre, Rotterdam, Netherlands 16Yokohama City University School of Medicine, Yokohama, Japan
Introduction: While the primary curative-intent treatment option for intrahepatic cholangiocarcinoma (ICC), liver resection (LR) can be associated with high postoperative morbidity and mortality. As such, a high-risk resection may not be warranted when the oncological benefits are minimal. In the current study, we sought to develop two preoperative models to predict 90-day mortality and overall survival (OS) after LR for ICC.
Methods: Patients who underwent curative-intent LR for ICC between 1990 and 2020 were identified from an international multi-institutional database. Two prognostic models were developed with preoperative factors using multivariable regression analysis for 90-day mortality and OS. Patients were categorized into three risk groups: favorable profile for LR (90-day mortality <10% and predicted OS >3 years), unfavorable profile (90-day mortality >25% and/or predicted OS <1.5 years), and an intermediate group.
Results: Among 889 patients included in the cohort, 90-day mortality was 8.5% (n=76) and median OS was 24.1 months. On multivariable regression analyses, independent risk factors for both 90-day mortality and OS were American Society of Anesthesiologists (ASA) grade >2, high tumor burden score, and receipt of neoadjuvant therapy. Additional risk factors for 90-day mortality included CA 19-9 >37 U/ml; relative to OS, advanced age, male sex, and major hepatectomy were associated with worse long-term survival (all p-values <0.05). In total, 168 patients (18.9%) had a favorable surgical risk profile with a 90-day mortality of 4.3% and a median OS of 47.0 months, 156 patients (17.5%) had an unfavorable profile with a 90-day mortality of 33.3% and a median OS of 11.9 months, and 565 patients (63.6%) had an intermediate profile with a 90-day mortality of 15.1% and a median OS of 29.9 months (Figure). The models were made available online for ease-of-use and clinical applicability (https://altaf-pawlik-icc-mortality-calculator.streamlit.app/).
Conclusion: Risk models for 90-day mortality and long-term OS can be used to identify patients preoperatively with a high-risk surgical profile who are unlikely to benefit from LR. Information on surgical risk profile should be used to tailor shared decision-making within a multidisciplinary setting.