Novel opportunistic imaging biomarkers with AI-assisted sarcopenia assessment associated with survival after liver transplantation for hepatocellular carcinoma.
Authors
Affiliations (7)
Affiliations (7)
- Radiology and Oncology, School of Medical Sciences. University of Campinas - UNICAMP, BRAZIL.
- Liver Transplant Unit. Surgery, School of Medical Sciences. University of Campinas - UNICAMP, Brasil.
- Radiology and Oncology, School of Medical Sciences. University of Campinas - UNICAMP, Brasil.
- Thoracic Imaging, The University of Texas MD Anderson Cancer Center, United States of America.
- Pathology, School of Medical Sciences. University of Campinas - UNICAMP, Brazil.
- Radiology, School of Medicine. Case Western Reserve University, United States of America.
- Radiology and Oncology, School of Medical Sciences. University of Campinas - UNICAMP, Brazil.
Abstract
The scarcity of donor organs necessitates the development of refined prognostic tools to optimize candidate selection and survival outcomes for patients with hepatocellular carcinoma (HCC) who receive liver transplantation. Standard computed tomography scans provide additional health indicators that reveal physiological vulnerabilities beyond tumor staging. In this observational cohort study with retrospective patient inclusion and prospective follow-up, we evaluated the effects of new computed tomography-derived biomarkers, with artificial intelligence (AI)-assisted sarcopenia assessment, on the survival of patients with hepatocellular carcinoma undergoing liver transplantation. We used preoperative chest and abdominal computed tomography scans to measure: 1. AI-assisted sarcopenia assessment; 2. myosteatosis; 3. osteoporosis, that although tomography diagnosis has been recently validated in other cohorts, it has not yet been assessed in survival of HCC patients undergoing transplantation; and 4. thymic fatty infiltration, a novel imaging biomarker of immunosenescence, that thus far has not been studied in the transplant setting. Survival analysis was performed using Kaplan-Meier curves and multivariable Cox regression models with stepwise selection. Sarcopenia, myosteatosis, osteoporosis, and thymus fatty replacement occurred in 35.4%, 51.3%, 27.3% and 82.1% of patients, respectively. The presence of microvascular invasion (HR= 1.943, p=0.0003), myosteatosis (HR=3.182, p=0.0001), osteoporosis (HR=2.133, p=0.0004), and thymic fatty replacement (HR=4.613, p=0.0001) independently were associated with a higher risk of mortality. While sarcopenia was significantly associated with survival in univariate (Kaplan-Meier) analysis (p=0.0002), it was not retained as an independent predictor in the final multivariable Cox model. These opportunistic imaging biomarkers provide significant noninvasive prognostic value, suggesting they may complement existing risk models and enhance the comprehensive assessment of patients undergoing liver transplantation. However, our findings require multicenter validation before integration into clinical workflows.