Radiomics Pipeline for Staging Chronic Liver Disease

Summary

Chronic liver disease (CLD), driven by factors such as alcohol abuse, viral hepatitis, and fatty liver disease, leads to progressive liver damage, fibrosis, and potentially cirrhosis or liver cancer. With CLD accounting for around two million deaths annually worldwide, timely diagnosis and staging are essential. Although liver biopsy remains the clinical gold standard for assessing fibrosis, its invasiveness and limitations have prompted a shift toward non-invasive imaging alternatives like MRI and CT.

This project proposes the development of a radiomics-based pipeline to stage liver fibrosis using MRI. Radiomics, a developing field in medical imaging, enables the extraction of high-dimensional quantitative features—such as texture, shape, and intensity—from standard imaging data. These features can reveal patterns not discernible to the human eye and support disease classification, prognosis, and treatment planning.

Project aims

The core aim is to establish a robust and reproducible radiomics workflow to characterize liver tissue and predict fibrosis stage. The pipeline will include MRI image pre-processing, feature extraction, robustness testing of extracted features, feature selection based on biological relevance, and machine learning-based classification of fibrosis stage (significant fibrosis, advanced fibrosis, cirrhosis).

A novel element of this project is the integration of motion-robust quantitative relaxation time imaging to enhance biomarker discovery. Ultimately, this work seeks to standardize and validate non-invasive image-based tools, enabling earlier diagnosis and more effective clinical management of CLD through routine MRI data, contributing to the shift toward AI-assisted, precision hepatology.

Contact details

Martin Graves - mjg40@cam.ac.uk

Opportunities

This project is open to applicants who want to do a:

• PhD