Clinical Research Directory

LEOPARD Training and Validation Data Collection Study

Intro: The present clinical research protocol is part of the LEOPARD European project (Grant n° 101080964 Horizon Europe) which aims to design and validate new predictive models of mortality among liver transplantation (LT) candidates. MELD based-liver graft allocation systems have become increasingly inaccurate over the last decade to predict mortality/dropout of liver transplantation (LT) candidates on the waitlist (WL). Wide disparities in mortality/dropout on the WL also exist across European countries, ranging from 5 to 30% according to transplantation indications and countries. In this setting, the European Commission- Horizon Europe funded-LEOPARD project intends to design new, 2nd generation, AI-machine learning-based predictive models of delisting in LT candidates, to better serve on time patients with the highest risk of dropout on the WL and to improve equity of access to LT across Europe. Hypothesis/Objective: The scientific justification of the LEOPARD TVDCS is therefore to collect a large set of data in liver transplantation candidates listed in Europe a) to design and b) to validate LEOPARD 2nd generation AI-based predictive models of mortality/dropout The primary objective is to develop new predictive models of mortality/drop out on the waitlist in patients with decompensated cirrhosis, or other end-stage chronic liver diseases, and in patients listed for Hepato-cellular carcinoma (HCC). Method: Longitudinal multicenter prospective health care data collection cohort study in 2 sets : Training/development set : Prospective health care data collection in 3,000 patients listed in 50 centres across 7 countries and Validation set: Prospective health care data collection in 1,500 subsequent patients listed in the same 50 centres.

LEOPARD Prospective Validation Cohort 1

Intro: The present clinical research protocol is part of the LEOPARD European project (Grant n° 101080964 Horizon Europe) which aims to design and validate new predictive models of mortality among liver transplantation (LT) candidates. MELD based-liver graft allocation systems have become increasingly inaccurate over the last decade to predict mortality/dropout of liver transplantation (LT) candidates on the waitlist (WL). Wide disparities in mortality/dropout on the WL also exist across European countries, ranging from 5 to 30% according to transplantation indications. In this setting, the European Commission- Horizon Europe funded-LEOPARD project intends to design new, 2nd generation, AI-machine learning-based predictive models of delisting in LT candidates, to better serve on time patients with the highest risk of dropout on the WL and to improve equity of access to LT across Europe. Hypothesis/Objective The scientific justification of the LEOPARD PVC1 is therefore 1. to build an external cohort of LT candidates to test and validate the LEOPARD models, therefore providing robust evidence for adoption of LEOPARD models by Organ Sharing Organizations (OSOs). 2. to collect granular data, bio- and tissues sampes and images to test last-generation OMICs predictors and radiomics, therefore opening the door to design of 3rd generation, precision medicine-based predictive models. The primary objective of the LEOPARD longitudinal study is to test and validate AI-based 2nd generation LEOPARD predictive models of mortality/drop out on the waitlist in patients with decompensated cirrhosis, or other end-stage chronic liver diseases, and in patients listed for HCC. Method Multicenter Prospective longitudinal study in up to 630 enrolments (in case of replacing participants after inclusion) to obtain 600 patients meeting selection criteria, in 30 hospitals in 5 European countries including France, Italy, The Netherlands, Belgium and Germany.