.avif)
Devlopment and Evalution of New Liver Donor Risk Index
.avif)
The Challenge
Despite significant advancements in immunosuppressive therapies, long-term survival rates post-liver transplantation have not seen considerable improvements in recent decades. The client (Scripps Health) sought to introduce innovative solutions to enhance the accuracy of liver graft longevity predictions and optimize donor-recipient compatibility. The Liver Donor Risk Index (LDRI), established in 2006, doesn’t fully reflect recent advancements in donor-recipient dynamics and transplant factors.
Key limitations of the LDRI include:
- Exclusion of key predictors such as liver steatosis, fibrosis, and advancements in machine perfusion that directly affect graft survival.
- Outdated immunosuppression protocols and the development of hepatitis C treatments, which have changed survival outcomes, rendering the original LDRI model insufficient.
- While newer models have emerged, there remains a need for further refinement to accurately reflect the evolving complexities of liver transplantation.
As a result, the client needed a solution that could integrate these modern variables to provide more accurate, personalized predictions for graft longevity.
Our Solution
Agilisium developed an advanced analytics solution leveraging AI/ML models designed to enhance donor-recipient mapping and improve graft survival rates. This solution focuses on the following key aspects:
Exploratory Data Analysis (EDA)
Performed EDA on liver transplant data to identify the clinical variables that influence post-transplant liver graft survival.
Enhanced LDRI Model
Integrated both donor characteristics and recipient factors to provide a more accurate prediction model of liver graft outcomes.
Cytokine Profiling
Analyzed the effect of cytokine profiles in the Early Allograft Dysfunction (EAD) presentation, providing insights into the immune response post-transplant.
Ensemble Classifier Development
Developed an ensemble classifier to predict the relative risk of graft failure based on various clinical and molecular factors.


The Outcomes
"We are grateful for the support and partnership of Agilisium Labs in advancing our work on predictive biomarkers and transplant outcomes. We are hopeful that this work will be clinically translatable." - Sunil Kurian, Scientific Director - Scripps Clinic Biorepository and Bioinformatics Core.

Improved Predictive Accuracy
Enhances the accuracy of predicting liver graft survival, allowing for better donor-recipient matching.

Better Patient Outcomes
Enables the identification of the right course of treatment, leading to better patient outcomes and optimized care.

Comprehensive Insights
Delivers insights on model evaluation, validation, and cytokine impact, providing actionable insights for clinical decision-making.

Accelerated Research and Development
Aids in development of new transplant methodologies and improving patient care.