(No, there is no typo in the title)
Multiparametric prostate MRI (mpMRI) is the new “gold standard” for prostate imaging. Detection of significant disease is enhanced when biopsy is guided by mpMRI. Although mpMRI is increasingly being used to assess risk, uncertainties about its accuracy confuse treatment decisions. The response to diagnostic uncertainty is often therapeutic overcompensation -- many men undergo aggressive treatment for disease that is ultimately found to be insignificant, and suffer otherwise avoidable pain, stress, expenses, and complications including impotence and incontinence.
The DWI component of the mpMRI scan has stronger correlations with both cancer grade and volume than the T2 and dynamic contrast (DCE) scans,(1-3) and there is growing evidence that the DCE component may be redundant (eg.(4)). Nevertheless the DWI component of recommended mpMRI protocols is performed in the most unsophisticated way possible -- by calculation of a single apparent diffusion coefficient (ADC) using a tissue diffusion model that is now well-known to be wrong (5,6).
This presentation summarizes the recent prostate DWI model wars, new information-based methods of comparing models, and promising new methods of inferring diagnostic tissue microstructure changes from advance DWI acquisition and data analysis methods.