Continuous Risk Score Predicts Waitlist and Post-transplant Outcomes in Hepatocellular Carcinoma Despite Exception Changes.

BACKGROUND & AIMS: Continuous risk-stratification of candidates and urgency-based prioritization have been utilized for liver transplantation (LT) in patients with non-hepatocellular carcinoma (HCC) in the United States. Instead, for patients with HCC, a dichotomous criterion with exception points is still used. This study evaluated the utility of the hazard associated with LT for HCC (HALT-HCC), an oncological continuous risk score, to stratify waitlist dropout and post-LT outcomes. METHODS: A competing risk model was developed and validated using the UNOS database (2012-2021) through multiple policy changes. The primary outcome was to assess the discrimination ability of waitlist dropouts and LT outcomes. The study focused on the HALT-HCC score, compared with other HCC risk scores. RESULTS: Among 23,858 candidates, 14,646 (59.9%) underwent LT and 5196 (21.8%) dropped out of the waitlist. Higher HALT-HCC scores correlated with increased dropout incidence and lower predicted 5-year overall survival after LT. HALT-HCC demonstrated the highest area under the curve (AUC) values for predicting dropout at various intervals post-listing (0.68 at 6 months, 0.66 at 1 year), with excellent calibration (R2 = 0.95 at 6 months, 0.88 at 1 year). Its accuracy remained stable across policy periods and locoregional therapy applications. CONCLUSIONS: This study highlights the predictive capability of the continuous oncological risk score to forecast waitlist dropout and post-LT outcomes in patients with HCC, independent of policy changes. The study advocates integrating continuous scoring systems like HALT-HCC in liver allocation decisions, balancing urgency, organ utility, and survival benefit.

Enhancing the usability of older DCD donors through strategic approaches in liver transplantation in the United States.

The use of older donors after circulatory death (DCD) for liver transplantation (LT) has increased over the past decade. This study examined whether outcomes of LT using older DCD (≥50 y) have improved with advancements in surgical/perioperative care and normothermic machine perfusion (NMP) technology. A total of 7602 DCD LT cases from the United Network for Organ Sharing database (2003-2022) were reviewed. The impact of older DCD donors on graft survival was assessed using the Kaplan-Meier and HR analyses. In all, 1447 LT cases (19.0%) involved older DCD donors. Although there was a decrease in their use from 2003 to 2014, a resurgence was noted after 2015 and reached 21.9% of all LTs in the last 4 years (2019-2022). Initially, 90-day and 1-year graft survivals for older DCDs were worse than younger DCDs, but this difference decreased over time and there was no statistical difference after 2015. Similarly, HRs for graft loss in older DCD have recently become insignificant. In older DCD LT, NMP usage has increased recently, especially in cases with extended donor-recipient distances, while the median time from asystole to aortic cross-clamp has decreased. Multivariable Cox regression analyses revealed that in the early phase, asystole to cross-clamp time had the highest HR for graft loss in older DCD LT without NMP, while in the later phases, the cold ischemic time (>5.5 h) was a significant predictor. LT outcomes using older DCD donors have become comparable to those from young DCD donors, with recent HRs for graft loss becoming insignificant. The strategic approach in the recent period could mitigate risks, including managing cold ischemic time (≤5.5 h), reducing asystole to cross-clamp time, and adopting NMP for longer distances. Optimal use of older DCD donors may alleviate the donor shortage.

LightGBM outperforms other machine learning techniques in predicting graft failure after liver transplantation: Creation of a predictive model through large-scale analysis.

BACKGROUND: The incidence of graft failure following liver transplantation (LTx) is consistent. While traditional risk scores for LTx have limited accuracy, the potential of machine learning (ML) in this area remains uncertain, despite its promise in other transplant domains. This study aims to determine ML's predictive limitations in LTx by replicating methods used in previous heart transplant research. METHODS: This study utilized the UNOS STAR database, selecting 64,384 adult patients who underwent LTx between 2010 and 2020. Gradient boosting models (XGBoost and LightGBM) were used to predict 14, 30, and 90-day graft failure compared to conventional logistic regression model. Models were evaluated using both shuffled and rolling cross-validation (CV) methodologies. Model performance was assessed using the AUC across validation iterations. RESULTS: In a study comparing predictive models for 14-day, 30-day and 90-day graft survival, LightGBM consistently outperformed other models, achieving the highest AUC of.740,.722, and.700 in shuffled CV methods. However, in rolling CV the accuracy of the model declined across every ML algorithm. The analysis revealed influential factors for graft survival prediction across all models, including total bilirubin, medical condition, recipient age, and donor AST, among others. Several features like donor age and recipient diabetes history were important in two out of three models. CONCLUSIONS: LightGBM enhances short-term graft survival predictions post-LTx. However, due to changing medical practices and selection criteria, continuous model evaluation is essential. Future studies should focus on temporal variations, clinical implications, and ensure model transparency for broader medical utility.

Development of a portable device to quantify hepatic steatosis in potential donor livers.

An accurate estimation of liver fat content is necessary to predict how a donated liver will function after transplantation. Currently, a pathologist needs to be available at all hours of the day, even at remote hospitals, when an organ donor is procured. Even among expert pathologists, the estimation of liver fat content is operator-dependent. Here we describe the development of a low-cost, end-to-end artificial intelligence platform to evaluate liver fat content on a donor liver biopsy slide in real-time. The hardware includes a high-resolution camera, display, and GPU to acquire and process donor liver biopsy slides. A deep learning model was trained to label and quantify fat globules in liver tissue. The algorithm was deployed on the device to enable real-time quantification and characterization of fat content for transplant decision-making. This information is displayed on the device and can also be sent to a cloud platform for further analysis.