Predicting Hepatocellular Carcinoma Graft Survival Rate in Post Liver Transplantation Using DeepHit

• Published in: Procedia computer science Vol. 233; pp. 307 - 316
• Main Authors: Rajeev, Devi, Dr. Remya, S, Nayyar, Dr. Anand, Nair, Dr. Krishnanunni
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 2024
• Subjects: DeepHit; Graft survival analysis; Hepatocellular carcinoma; Post-liver transplantation; Graft survival analysis; Hepatocellular carcinoma; DeepHit; Post-liver transplantation
• ISSN: 1877-0509; 1877-0509
• DOI: 10.1016/j.procs.2024.03.220

Hepatocellular carcinoma (HCC) is still a severe worldwide health issue, and liver transplantation is frequently the primary curative option for people who meet the criteria for it. It is challenging to predict the long-term graft survival of HCC patients after transplantation. However, it is an essential clinical decision-making task. The DeepHit algorithm is used in this study to predict the long-term survival of grafts using a deep learning-based approach. In this study, we used the dataset collected from the Amrita Institute of Medical Science & Research Centre, Kochi, with patient-specific information on age, gender, tumor size, AFP levels, cirrhosis status, treatment approach, recurrence, comorbidity, and time to transplant failure. The temporal correlations in patient data are modeled, and censoring is considered using the DeepHit technique. It blends survival analysis with the benefits of deep neural networks. DeepHit predicts long-term graft survival in liver cancer patients after transplantation with 94 % accuracy, outperforming conventional methods and improving prognosis and quality of life.