bridging the micro-macro gap with diffusion MRI


Diffusion MRI (dMRI) is a biomarker that derives its contrast based on micrometer level restrictions to water diffusion. Due to this property, it has gained considerable attention as a prostate cancer biomarker and plays a significant role in the standardized Prostate Imaging-Reporting and Data System (PI-RADS). While dMRI is highly sensitive, it is not specific. Gray-scale contrast may be derived from different sources of microscopic architecture. Due to this lack of specificity, dMRI results are typically confirmed through biopsy, which has its own limitations that may divert diagnostics. Moreover, the clinical dMRI protocol features the bare minimum of sampling within the tunable diffusion parameters: direction, b-value, and diffusion time. Given the complexity of the tissue, it is clear that the Diffusion MRI protocol is not exploring its full potential.

Our group has been interested in applying in expanding dMRI by varying diffusion time. As the diffusion time is varied, the average distance that water molecules may diffuse will also change. By modeling the diffusion coefficient as a function of time, it is possible to derive microscopic length scales that pertain to specific tissue structures. As prostate cancer progresses, in general, the acinar glands of the prostate shrink and the surrounding stroma will become increasingly dense. Since dMRI is sensitive to these microscopic changes, modeling time-dependent diffusion in the prostate can give us a specific quantitative metric for gland size, which can then be used as a biomarker for cancer grade. This research will allow future routine MRI visits to provide extended analysis on the cellular scale (in microns) rather than tissue scale (in millimeters), and serve as a non-invasive complement to biopsy. 

Research Topic Image: 
Research Category: 


D. S. Novikov, Jensen, J. H., Helpern, J. A., and Fieremans, E., Revealing mesoscopic structural universality with diffusion, Proceedings of the National Academy of Sciences of the United States of America (PNAS), vol. 111, pp. 5088-5093, 2014.

Scholarly Lite is a free theme, contributed to the Drupal Community by More than Themes.