Validation of microstructure models of diffusion signal using histology. In Eleftheria Panagiotaki’s and David Hawkes’s labs, I used mathematical models to describe diffusion MRI signal in terms of vascular, cellular and extracellular components. The aim was to characterize tumours with more biologically relevant parameters and make diffusion-based MRI biomarkers more specific. This work has been used to characterize prostate and breast cancers as well as bone metastases. It includes histological validation using patient-specific 3-D-printed molds and image registration.
Diffusion anisotropy in stromal regions of breast and prostate. Diffusion MRI measurements in cancer have focused on associations between the signal attenuation and cell density. However, changes in anisotropy (directional differences) have also been observed in breast and prostate tissue, but their origin and importance remains largely unknown. I scanned breast and prostate samples ex vivo, demonstrating that anisotropy existed in the absence of ductal structures and may be related to patterns of collagen in the stroma that have been implicated in cancer invasion. These directional diffusion patterns are difficult to detect with conventional in vivo methods and I am therefore developing new methods to explore anisotropy in the clinical setting.
MRI markers of apoptotic cell death. MRI shows anatomical features, but the signal intensity has a complex relationship with microscopic and cellular characteristics of the tissue. I developed a method using conventional contrast agents, T1- and T2-sensitive MRI sequences and a mathematical model to estimate the water exchange across the cell membrane, a parameter that increases during apoptotic cell death. Previous attempts to detect apoptosis with relaxation-based sequences had been unsuccessful or shown changes only at very late stages of cell death. The addition of contrast agents allows early detection and also explains the competing effects that make detection challenging in the absence of contrast. This work was developed in vitro and then translated to clinic, where it predicted the eventual tumour volume decrease following treatment.