Revolutionising liver transplantation: The LEOPARD Project

7th August 2024

The new EU-funded LEOPARD project aims to transform liver transplantation by leveraging artificial intelligence to better stratify patients waiting for life-saving surgery. Here, Professor Christophe Duvoux, LEOPARD project coordinator, discusses the current state of liver transplantation in Europe, how the project aims to overcome the challenges and its anticipated impact on clinical care and patient outcomes.

Liver Electronic Offering Platform with ARtificial Intelligence-based Devices (LEOPARD) is a pioneering, EU-funded initiative in liver transplantation, uniting stakeholders across Europe to revolutionise organ allocation strategies for individuals with decompensated cirrhosis (DC) and hepatocellular carcinoma (HCC).¹

Liver transplantation is the only lifesaving option for patients with end-stage cirrhosis and early, moderately advanced HCC,² accounting, on average, for 50% and 35% of liver transplantation indications, respectively. However, its efficacy is hampered by significant challenges related to organ shortages and increasingly limited allocation models.^3,4

With only 5,000 liver transplantations performed across Europe annually,⁵ compared to 12,000 patients on waitlists, there is fewer than one liver graft available for every two patients. The consequence? Among patients currently listed for liver transplantation in Europe, waitlist mortality and dropout averages 15-20%, with large disparities across European countries.¹

Outdated methods hindering progress

Continued reliance on the Model for End-Stage Liver Disease (MELD) score to prioritise transplantations is hindering progress within the liver transplantation landscape. Originally devised over two decades ago for a different patient demographic, the MELD score was designed to limit waitlist mortality risk in patients with DC but now falls short in addressing the complexities of today’s patient population.

This score, which hinges on three biological predictors – creatinine, bilirubin and International Normalised Ratio – fails to adequately capture the 21^st century DC and HCC patient, many of whom are older, living with multiple comorbidities, or are in intensive care settings.

Approximately 25-30% of DC patients receive organ transplants based on extra MELD points, while there is currently no applicable predictive model for waitlist dropout for HCC patients.^3,4,6 HCC patients are often prioritised through MELD exception rules, as their lab MELD scores do not encapsulate their transplant urgency.

The reality is stark. Overall, only 30-40% of transplant candidates are allocated organs based solely on their true lab MELD scores. This system needs an urgent overhaul to ensure fair and accurate prioritisation, enhanced patient outcomes, and reduced waitlist mortality.

Addressing liver transplantation challenges using AI

Aligned with the ‘Horizon Europe HLTH 2022 tender, Tool 12 01: Computational models for new patient stratification strategies’ – a funding call under the Horizon Europe programme aimed at advancing healthcare through the development of innovative computational models to improve patient categorisation based on risks and needs – the LEOPARD project seeks to address these challenges and to unlock the potential of artificial intelligence (AI).

The project focuses on developing and validating an AI-based predictive algorithm that surpasses current prioritisation models and accurately stratifies both DC and HCC patients by their risks of waitlist mortality or dropout. This will be achieved by:¹

• Enhanced predictive models: Integrating new predictors of mortality in DC and dropout in HCC patients to create superior predictive models
• Development of calculators: Creating tools to aid in patient prioritisation for timely transplants and reduced waitlist mortality
• Integration of OMICs and radiomics: Incorporating predictive signatures from OMICs (e.g. genomics, proteomics, etc.) and radiomics (quantitative imaging data) to improve risk assessment accuracy.

LEOPARD’s backbone is a robust consortium of major European Organ Sharing Organisations (OSOs), clinical research experts, research labs, small or midsize enterprises, patient organisations and scientific societies including the European Society for Organ Transplantation (ESOT).

Some 50 liver transplant centres across seven European countries will contribute to a dataset of 3,000 patients with DC and HCC – data that will inform the development of advanced predictive models using machine learning techniques.

Validation will occur through real-life prospective cohorts of 1,500 patients, supplemented by a cohort of 600 patients with bio- and imaging banking. Simulations will assess the impact of LEOPARD predictive algorithms on graft allocation and patient outcomes.

Impacting transplantation clinical practice

Critically, LEOPARD will produce outputs that can be used in clinical practice, including:¹

• A validated algorithmic predictive solution outperforming the MELD system, ready for adoption by OSOs to refine graft allocation in Europe and beyond
• LEOPARD calculators to assist healthcare professionals in clinical decision-making
• Validated OMICs and radiomics signatures for improved patient assessment
• A unique multimodal database available for external investigations, fostering further research and clinical applications.

LEOPARD presents an unprecedented opportunity to unite the liver transplantation community around a robust European initiative. By leveraging cutting-edge AI and collaborative expertise, LEOPARD has the potential to significantly improve patient outcomes, harmonise prioritisation schemes across Europe and advocate for equitable access to lifesaving liver transplantations. This project promises immediate clinical applications and opens new research perspectives, envisioning a future with improved waitlist mortality or dropout to compensate current organ allocation inefficiencies.

Author

Professor Christophe Duvoux
LEOPARD project coordinator and full professor of hepatology and head of the Medical Liver Transplantation Program at Mondor Hospital, Paris, France

References

1. LEOPARD (2024) Liver Electronic Offering Platform with ARtificial Intelligence-based Devices: A European commission-funded project (last accessed August 2024).
2. Yankol, Y., Aguirre, O., & Fernandez, L. A. (2024). Impact of Living Donor Liver Transplantation on the Improvement of Hepatocellular Carcinoma Treatment. Sisli Etfal Hastanesi tip bulteni, 58(1), 1–9.
3. Burro, P., Samual, D., Sundaram, V., Duvoux, C., Petrowsky, H., Terrault, N., Jalan, R. (2021) Limitations of current liver donor allocation systems and the impact of newer indications for liver transplantation.
4. Godfrey, E. L., Malik, T. H., Lai, J. C., Mindikoglu, A. L., Galván, N. T. N., Cotton, R. T., O’Mahony, C. A., Goss, J. A., & Rana, A. (2019). The decreasing predictive power of MELD in an era of changing etiology of liver disease. American journal of transplantation: official journal of the American Society of Transplantation and the American Society of Transplant Surgeons, 19(12), 3299–3307.
5. Samuel, D., & Coilly, A. (2018). Management of patients with liver diseases on the waiting list for transplantation: a major impact to the success of liver transplantation. BMC medicine, 16(1), 113.
6. Agence de la biomedicine (2022) [Annual Report: Organs – Organ transplantation: general data and methods] – report in French (last accessed August 2024).

Machine-learning MRI improves prediction of liver cancer recurrence

24th August 2022

A machine-learning MRI model better predicts liver cancer recurrence compared to a clinical-based model but is similar to a combined model

A machine-learning MRI model is better able to predict the recurrence of hepatocellular carcinoma (HCC) after a liver transplant than a model based on clinical and laboratory data but is equally effective to a model which uses a combination of clinical/laboratory and MRI-derived data according to a study by US and German researchers.

Liver cancer, of which HCC accounts for about 90% of all cases, remains a global health challenge and it is estimated to have an incidence of over a million cases by 2025. Potentially curative treatment options for hepatocellular carcinoma include liver transplantation, liver resection and thermal ablation, with transplantation offering the lowest rate of cancer recurrence and highest chance of long-term survival.

However, despite this, estimated post-transplantation recurrence rates are between 15% and 20%. Methods to estimate the risk of recurrence are therefore needed and hepatobiliary magnetic resonance imaging (MRI) preoperative findings have been found to be associated with a higher tumour recurrence rate in transplanted patients.

Machine-learning MRI models have the potential ability to extract information from unstructured medical imaging data and might be of predictive value for cancer recurrence but whether this approach is of value for HCC is uncertain and was the purpose of the present study.

Researchers retrospectively analysed data from a cohort of patients with HCC treated by liver transplant, surgical resection or thermal ablation and who had undergone pre-and post-treatment MRI scans. The US and German team trained a machine-learning MRI system to extract imaging features and developed three predictive models.

The first used imaging-derived data only, the second clinical and laboratory individual patient data and a final model, combined the imaging and clinical/laboratory data. The risk of HCC recurrence was predicted over a 6 year period after a patient’s first-line treatment. The predictive value of the different models were assessed based on the area under the receiver operating characteristic curve (AUC).

Machine-learning MRI model and prediction of HCC recurrence

The study included 120 patients with a mean age of 60 years (26.7% male) of whom, 36.7% experienced tumour recurrence during follow-up, with the mean time to recurrence being 26.8 months.

The highest AUC for each of the three models was achieved for the periods 4 and 6 years after treatment. After 6 years, the AUC for the clinical model was 0.69 (95% CI 0.54 – 0.84), 0.85 (95% CI 0.75 – 0.95) for the imaging model and 0.86 (95% CI 0.76 – 0.96) for the combined model.

Over the 6-year period the mean AUC for the imaging model was 0.76, 0.68 for the clinical model and this difference was statistically significant (p = 0.03) although the AUC for the combined model was the same as the imaging model (0.76).

Turning to the individual patient data, the clinical model correctly predicted 25% of recurrences, whereas the imaging model and combined models, both corrected predicted 87.5% of recurrences.

The authors concluded that a machine-learning MRI model could successful predict recurrence of early-stage HCC and that this model was superior to the use of clinical data alone and called for prospective cohort studies to externally validate these algorithms prior to clinical use.

Citation
Iske S et al. Machine-Learning Models for Prediction of Posttreatment Recurrence in Early-Stage Hepatocellular Carcinoma Using Pretreatment Clinical and MRI Features: A Proof-of-Concept Study AJR Am J Roentgenol 2022