Histology-informed liver diffusion MRI: biophysical model design and demonstration in cancer immunotherapy

Innovative diffusion Magnetic Resonance Imaging (dMRI) models enable in vivo mapping of biologically meaningful properties such as cell size, potential biomarkers in cancer. However, while cancers frequently spread to the liver, models tailored for liver applications and easy to deploy in the clinic are still sought. We tackle this unmet need by delivering a practical and clinically viable liver dMRI modelling framework. Through direct comparison of candidate dMRI approaches in mouse models and cancer patients’ data, we select a model of intra-cellular diffusion fitted to highly diffusion-weighted images, as it provides the strongest radiological-histological correlates. We demonstrate the potential application of the proposed model in cancer immunotherapy, stratifying the risk of progression based on baseline tumour cell size/density from dMRI. This result, heretofore unreported and not achievable with standard dMRI indices (e.g., apparent diffusion coefficient), suggests that our approach may become a useful tool for precision imaging in oncology.