Comparison of Different Scoring Systems Based on Both Donor and Recipient Characteristics for Predicting Outcome after Living Donor Liver Transplantation

Future of U.S. living donor liver transplant: Donor and recipient criteria, transplant indications, transplant oncology, liver paired exchange, and non-directed donor graft allocation

In the United States, living donor liver transplant (LDLT), from both directed and nondirected living donors, has expanded over the past several years. LDLT is viewed as an important opportunity to expand the overall donor pool for liver transplantation (LT), shorten waiting times for a life-prolonging LT surgery, and reduce LT waitlist mortality. The LT community's focus on LDLT expansion in the United States is fostering discussions around future opportunities, which include the safe expansion of donor and recipient candidate eligibility criteria, broadening indications for LDLT including applications in transplant oncology, developing national initiatives around liver paired exchange, and maintaining vigilance to living donor and recipient candidate risk/benefit equipoise. Potential opportunities for expanding living liver donor and recipient candidate criteria include using donors with more than minimal hepatic steatosis, evaluating older donors, performing LDLT in older recipients to facilitate timely transplantation, and providing candidates who would benefit from an LT, but may otherwise have limited access (ie, lower MELD scores), an avenue to receive a life-prolonging organ. Expansion opportunities for LDLT are particularly robust in the transplant oncology realm, including leveraging LDLT for patients with advanced HCC beyond Milan, intrahepatic cholangiocarcinoma, and nonresectable colorectal cancer liver metastases. With ongoing investment in the deliberate growth of LDLT surgical expertise, experience, and technical advances in the United States, the LT community's future vision to increase transplant access to more patients with end-stage liver disease and selected oncology patients may be successfully realized.

Use of machine learning models for the prognostication of liver transplantation: A systematic review

BACKGROUND

Liver transplantation (LT) is a life-saving intervention for patients with end-stage liver disease. However, the equitable allocation of scarce donor organs remains a formidable challenge. Prognostic tools are pivotal in identifying the most suitable transplant candidates. Traditionally, scoring systems like the model for end-stage liver disease have been instrumental in this process. Nevertheless, the landscape of prognostication is undergoing a transformation with the integration of machine learning (ML) and artificial intelligence models.

AIM

To assess the utility of ML models in prognostication for LT, comparing their per formance and reliability to established traditional scoring systems.

METHODS

Following the Preferred Reporting Items for Systematic Reviews and Meta-Analysis guidelines, we conducted a thorough and standardized literature search using the PubMed/MEDLINE database. Our search imposed no restrictions on publication year, age, or gender. Exclusion criteria encompassed non-English stu dies, review articles, case reports, conference papers, studies with missing data, or those exhibiting evident methodological flaws.

RESULTS

Our search yielded a total of 64 articles, with 23 meeting the inclusion criteria. Among the selected studies, 60.8% originated from the United States and China combined. Only one pediatric study met the criteria. Notably, 91% of the studies were published within the past five years. ML models consistently demonstrated satisfactory to excellent area under the receiver operating characteristic curve values (ranging from 0.6 to 1) across all studies, surpassing the performance of traditional scoring systems. Random forest exhibited superior predictive capa bilities for 90-d mortality following LT, sepsis, and acute kidney injury (AKI). In contrast, gradient boosting excelled in predicting the risk of graft-versus-host disease, pneumonia, and AKI.

CONCLUSION

This study underscores the potential of ML models in guiding decisions related to allograft allocation and LT, marking a significant evolution in the field of prognostication.

Maximizing utility of nondirected living liver donor grafts using machine learning

Objective

There is an unmet need for optimizing hepatic allograft allocation from nondirected living liver donors (ND-LLD).

Materials and method

Using OPTN living donor liver transplant (LDLT) data (1/1/2000-12/31/2019), we identified 6328 LDLTs (4621 right, 644 left, 1063 left-lateral grafts). Random forest survival models were constructed to predict 10-year graft survival for each of the 3 graft types.

Results

Donor-to-recipient body surface area ratio was an important predictor in all 3 models. Other predictors in all 3 models were: malignant diagnosis, medical location at LDLT (inpatient/ICU), and moderate ascites. Biliary atresia was important in left and left-lateral graft models. Re-transplant was important in right graft models. C-index for 10-year graft survival predictions for the 3 models were: 0.70 (left-lateral); 0.63 (left); 0.61 (right). Similar C-indices were found for 1-, 3-, and 5-year graft survivals. Comparison of model predictions to actual 10-year graft survivals demonstrated that the predicted upper quartile survival group in each model had significantly better actual 10-year graft survival compared to the lower quartiles (p<0.005).

Conclusion

When applied in clinical context, our models assist with the identification and stratification of potential recipients for hepatic grafts from ND-LLD based on predicted graft survivals, while accounting for complex donor-recipient interactions. These analyses highlight the unmet need for granular data collection and machine learning modeling to identify potential recipients who have the best predicted transplant outcomes with ND-LLD grafts.

BAR Score Performance in Predicting Survival after Living Donor Liver Transplantation: A Single-Center Retrospective Study

METHODS

146 adult liver transplant recipients were included. Univariate and multivariate analyses were used to determine the independent predictors of survival at 3 months, 1 year, and 5 years. The receiver operating characteristic (ROC) curve for the BAR score was plotted, and the area under the ROC curve (AUROC) was calculated. Kaplan-Meier curve and log-rank test were used to compare survival above and below the best cutoff values.

RESULTS

The mean age was 52.45 ± 8.54 years, and 59.6% were males. The survival rates were 89, 78.8, and 72% at 3 months, 1 year, and 5 years, respectively. The BAR score demonstrated a clinically significant value in the prediction of 3-month (AUROC = 0.89), 1-year (AUROC = 0.76), and 5-year survival (AUROC = 0.71). Among the investigated factors associated with survival, BAR score <10 points was the only independent predictor of 3-month (OR 7.34, p < 0.0001), 1-year (OR 3.37, p=0.001), and 5-year survival (OR 2.83, p=0.044).

CONCLUSIONS

BAR is a simple and easily applicable scoring system that could significantly predict short- and long-term survival after LDLT. A large multicenter study is warranted to validate our results in the Egyptian population.

Combining Donor and Recipient Age With Preoperative MELD and UKELD Scores for Predicting Survival After Liver Transplantation

OBJECTIVES

The end-stage liver disease scoring systems MELD, UKELD, and D-MELD (donor age × MELD) have had mediocre results for survival assessment after orthotopic liver transplant. Here, we introduced new indices based on preoperative MELD and UKELDscores and assessed their predictive ability on survival posttransplant.

MATERIALS AND METHODS

We included 1017 deceased donor orthotopic liver transplants that were performed between 2008 (the year UKELD was introduced) and 2019. Donor and recipient characteristics, liver disease scores, transplant characteristics, and outcomes were collected for analyses. D-MELD, D-UKELD (donor age × UKELD),DR-MELD[(donor age + recipient age) × MELD], and DR-UKELD [(donor age + recipient age) × UKELD] were calculated.

RESULTS

No score had predictive value for graft survival. For patient survival,DR-MELD and DR-UKELD provided the best results but with low accuracy. The highest accuracy was observed at 1 year posttransplant (areas under the curve of 0.598 [95% CI, 0.529-0.667] and 0.609 [95% CI, 0.549-0.67]forDR-MELDandDR-UKELD). Addition of donor and recipient age significantly improved the predictive abilities of MELD and UKELD for patient survival, but addition of donor age alone did not. For 1-year mortality (using receiver operating characteristic curves), optimal cut-off points were DR-MELD>2345 and DR-UKELD>5908. Recipients with DR-MELD >2345 (P < .001) and DR-UKELD >5908 (P = .002) had worse patient survival within the first year, but only DR-MELD >2345 remained significant after multivariable analysis (P = .007).

CONCLUSIONS

DR-MELD and DR-UKELD scores provided the best, albeit mediocre, predictive ability among the 6 tested models, especially at 1 year after posttransplant, although only for patient but not for graft survival. A DR-MELD >2345 was considered to be an additional independent risk factor for worse recipient survival within the first postoperative year.

Statistical methods versus machine learning techniques for donor-recipient matching in liver transplantation

Donor-Recipient (D-R) matching is one of the main challenges to be fulfilled nowadays. Due to the increasing number of recipients and the small amount of donors in liver transplantation, the allocation method is crucial. In this paper, to establish a fair comparison, the United Network for Organ Sharing database was used with 4 different end-points (3 months, and 1, 2 and 5 years), with a total of 39, 189 D-R pairs and 28 donor and recipient variables. Modelling techniques were divided into two groups: 1) classical statistical methods, including Logistic Regression (LR) and Naïve Bayes (NB), and 2) standard machine learning techniques, including Multilayer Perceptron (MLP), Random Forest (RF), Gradient Boosting (GB) or Support Vector Machines (SVM), among others. The methods were compared with standard scores, MELD, SOFT and BAR. For the 5-years end-point, LR (AUC = 0.654) outperformed several machine learning techniques, such as MLP (AUC = 0.599), GB (AUC = 0.600), SVM (AUC = 0.624) or RF (AUC = 0.644), among others. Moreover, LR also outperformed standard scores. The same pattern was reproduced for the others 3 end-points. Complex machine learning methods were not able to improve the performance of liver allocation, probably due to the implicit limitations associated to the collection process of the database.

Prognostic role of the Donor Risk Index, the Eurotransplant Donor Risk Index, and the Balance of Risk score on graft loss after liver transplantation

This study aimed to identify cutoff values for donor risk index (DRI), Eurotransplant (ET)-DRI, and balance of risk (BAR) scores that predict the risk of liver graft loss. MEDLINE and Web of Science databases were searched systematically and unrestrictedly. Graft loss odds ratios and 95% confidence intervals were assessed by meta-analyses using Mantel-Haenszel tests with a random-effects model. Cutoff values for predicting graft loss at 3 months, 1 year, and 3 years were analyzed for each of the scores. Measures of calibration and discrimination used in studies validating the DRI and the ET-DRI were summarized. DRI ≥ 1.4 (six studies, n = 35 580 patients) and ET-DRI ≥ 1.4 (four studies, n = 11 666 patients) were associated with the highest risk of graft loss at all time points. BAR > 18 was associated with the highest risk of 3-month and 1-year graft loss (n = 6499 patients). A DRI cutoff of 1.8 and an ET-DRI cutoff of 1.7 were estimated using a summary receiver operator characteristic curve, but the sensitivity and specificity of these cutoff values were low. A DRI and ET-DRI score ≥ 1.4 and a BAR score > 18 have a negative influence on graft survival, but these cutoff values are not well suited for predicting graft loss.

The role of the comprehensive complication index for the prediction of survival after liver transplantation

In the last years, several scoring systems based on pre- and post-transplant parameters have been developed to predict early post-LT graft function. However, some of them showed poor diagnostic abilities. This study aims to evaluate the role of the comprehensive complication index (CCI) as a useful scoring system for accurately predicting 90-day and 1-year graft loss after liver transplantation. A training set (n = 1262) and a validation set (n = 520) were obtained. The study was registered at https://www.ClinicalTrials.gov (ID: NCT03723317). CCI exhibited the best diagnostic performance for 90 days in the training (AUC = 0.94; p < 0.001) and Validation Sets (AUC = 0.77; p < 0.001) when compared to the BAR, D-MELD, MELD, and EAD scores. The cut-off value of 47.3 (third quartile) showed a diagnostic odds ratio of 48.3 and 7.0 in the two sets, respectively. As for 1-year graft loss, CCI showed good performances in the training (AUC = 0.88; p < 0.001) and validation sets (AUC = 0.75; p < 0.001). The threshold of 47.3 showed a diagnostic odds ratio of 21.0 and 5.4 in the two sets, respectively. All the other tested scores always showed AUCs < 0.70 in both the sets. CCI showed a good stratification ability in terms of graft loss rates in both the sets (log-rank p < 0.001). In the patients exceeding the CCI ninth decile, 1-year graft survival rates were only 0.7% and 23.1% in training and validation sets, respectively. CCI shows a very good diagnostic power for 90-day and 1-year graft loss in different sets of patients, indicating better accuracy with respect to other pre- and post-LT scores.Clinical Trial Notification: NCT03723317.

Predictive Capacity of Risk Models in Liver Transplantation

Supplemental Digital Content is available in the text.

Several risk models to predict outcome after liver transplantation (LT) have been developed in the last decade. This study compares the predictive performance of 7 risk models.

Potential value and limitations of different clinical scoring systems in the assessment of short- and long-term outcome following orthotopic liver transplantation

BACKGROUND

In an attempt to further improve liver allograft utilization and outcome in orthotopic liver transplantation (OLT), a variety of clinical scoring systems have been developed. Here we aimed to comparatively investigate the association of the Balance-of-Risk (BAR), Survival-Outcomes-Following-Liver-Transplant (SOFT), Preallocation-Survival-Outcomes-Following-Liver-Transplant (pSOFT), Donor-Risk-Index (DRI), and the Eurotransplant-Donor-Risk-Index (ET-DRI) scores with short- and long-term outcome following OLT.

METHODS

We included 338 consecutive patients, who underwent OLT in our institution between May 2010 and November 2017. For each prognostic model, the optimal cutoff values were determined with the help of the Youden-index and their diagnostic accuracy for 90-day post OLT-mortality and major postoperative complications was measured by the area under the receiver operating characteristic curve (AUROC). Patient- and graft survival were analyzed using the Kaplan-Meier method and the log-rank test. Morbidity was assessed using the Clavien-Dindo classification and the Comprehensive-Complication-Index.

RESULTS

BAR, SOFT, and pSOFT performed well above the conventional AUROC-threshold of 0.70 with good prediction of early mortality. Only BAR showed AUC>0.70 for both mortality and major morbidity. With the cutoffs of 14, 31, and 22 respectively for BAR, SOFT, and pSOFT, subgroup analysis showed significant differences (p<0.001) in morbidity and mortality, length of intensive care- and hospital-stay and early allograft dysfunction rates. Five-years patient survival was inferior in the high BAR, pSOFT, and SOFT groups.

CONCLUSIONS

Out of all scores tested, the BAR-score had the best value in predicting both 90-day morbidity and mortality after OLT showing the highest AUCs. The pSOFT and SOFT scores demonstrated an acceptable accuracy in predicting 90-day morbidity and mortality. The used BAR, SOFT, and pSOFT cutoffs allowed the identification of patients at risk in terms of five-year patient survival. The DRI and ET-DRI scores have failed to predict recipient outcomes in the present setting.

Is Muscle MELD a More Promising Predictor for Mortality After Living Donor Liver Transplantation?

BACKGROUND

To improve the outcome of living donor liver transplantation (LDLT), a scoring system that could predict accurately the patient and graft survival posttransplant is necessary. The aim of this study is to evaluate our previously proposed Muscle-model for end-stage liver disease (M-MELD) score and to compare it with the other available scores to find the best system that correlates with postoperative outcome after liver transplant.

METHODS

We retrospectively reviewed the data of 199 patients who underwent LDLT from January 2010 to July 2016 and calculated the preoperative MELD, MELD Na, the product of donor age and MELD (D-MELD), M-MELD, integrated MELD, and the balance of risk (BAR) score in all patients. The area under the receiver operating characteristics curves (AUCs) of each score was computed and compared at 3-, 6-months, and 1-year after LDLT.

RESULTS

The M-MELD, D-MELD, and integrated MELD had a good discriminative performance in predicting 3-month mortality after LDLT with AUCs > 0.7, while the M-MELD was the only score that showed a good discriminative performance in predicting 6-month and 1-year mortality after LDLT with AUCs > 0.7.

CONCLUSION

Muscle-MELD score is a simple and useful predictor of patient survival after LDLT which showed a better predictive performance than other available scores.