Canadian liver journal · 2026
Investigating the role of artificial intelligence in predicting risk factors after liver transplantation: A systematic review.
СтатьяСистематический обзорНазвание и аннотация: Перевод пока не подготовлен
Обработка статьи · этапы, ошибки и управление
Загрузка состояния…
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Esmailzadeh Aynaz, Norouzkhani Narges, Heidari Amir Ehsan, Habibi Mohammad Reza Mazaheri
Аннотация
BACKGROUND: Liver transplantation (LT) is the treatment of choice for patients with end-stage liver disease, yet its success is often compromised by serious complications such as acute kidney injury (AKI), graft failure, recurrence of hepatocellular carcinoma (HCC), and infections. Recent advances in AI-especially machine learning (ML) and deep learning (DL)-offer promising tools for predicting these outcomes and supporting early clinical interventions. This systematic review aimed to evaluate the performance and limitations of AI-based models in predicting key post-LT complications.
METHODS: PubMed, Scopus, Web of Science, and Google Scholar were comprehensively searched from January 2000 to December 2025. Studies applying AI algorithms to predict post-LT outcomes in humans were included. Non-English-language articles, non-original studies, and articles without predictive modelling were excluded. Two reviewers independently screened and extracted data, including study design, AI type, input data, validation method, and performance metrics. Due to high heterogeneity, no meta-analysis was performed.
RESULTS: Of 6,744 articles identified, 14 met the inclusion criteria. ML models (eg, random forest, XGBoost, support vector machines, and neural networks) were used in most studies; DL architectures (eg, CNNs, DeepSurv, and CapsNets) were used in a few, and reported AUROC ranges were 0.61-0.96. Only four studies used external validation. Commonly predicted outcomes included AKI, HCC recurrence, sepsis, graft-versus-host disease, and malignancy.
CONCLUSIONS: AI algorithms show strong potential for predicting post-LT complications, particularly when ensemble or imaging-based DL models are used. However, limited validation and interpretability remain challenges. Future research should focus on multicentre validation, transparent reporting, and clinical integration.
METHODS: PubMed, Scopus, Web of Science, and Google Scholar were comprehensively searched from January 2000 to December 2025. Studies applying AI algorithms to predict post-LT outcomes in humans were included. Non-English-language articles, non-original studies, and articles without predictive modelling were excluded. Two reviewers independently screened and extracted data, including study design, AI type, input data, validation method, and performance metrics. Due to high heterogeneity, no meta-analysis was performed.
RESULTS: Of 6,744 articles identified, 14 met the inclusion criteria. ML models (eg, random forest, XGBoost, support vector machines, and neural networks) were used in most studies; DL architectures (eg, CNNs, DeepSurv, and CapsNets) were used in a few, and reported AUROC ranges were 0.61-0.96. Only four studies used external validation. Commonly predicted outcomes included AKI, HCC recurrence, sepsis, graft-versus-host disease, and malignancy.
CONCLUSIONS: AI algorithms show strong potential for predicting post-LT complications, particularly when ensemble or imaging-based DL models are used. However, limited validation and interpretability remain challenges. Future research should focus on multicentre validation, transparent reporting, and clinical integration.