Diffusion MRI for Head and Neck Cancer

The proposed study is to investigate the feasibility of using quantitative diffusion MRI (dMRI) methods for accurate and comprehensive assessment of treatment response. dMRI is a powerful tool to probe treatment-induced change in tumors. It is a unique in vivo imaging technique sensitive to cellular microstructures at the scale of water diffusion length on the order of a few microns. Previous studies have shown that both diffusion coefficient D and diffusional kurtosis coefficient K are promising imaging markers of (i) cell viability which can be used for evaluation of early treatment response. However, it is often underappreciated that these dMRI metrics are not fixed constants, but rather functions of the diffusion time t, D(t) and K(t); their t-dependency is determined by tissue properties, such as cell size and membrane permeability of tissue. D(t) and K(t) of tumors can vary substantially depending on t in the range of diffusion times (30-100 ms) typically used in clinical scan.

This study will investigate the t-dependency of dMRI over a range of diffusion times (30-500 ms) to determine an optimal diffusion time for treatment response assessment when only one diffusion time needs to be used, particularly in routine clinical studies. Furthermore, the data with multiple diffusion times will also be used to measure the water exchange time of cancer cells. Exchange time has been studied using Dynamic Contrast Enhanced (DCE) MRI by multiple groups including ours, and has been suggested as a marker of (ii) cellular metabolism that regulates the ATP-dependent ion channels co-transporting water molecules. The study will measure with dMRI, without using a contrast agent. The investigators also demonstrated that Intra-Voxel Incoherent Motion (IVIM) MRI metrics (pseudo diffusivity, Dp; perfusion fraction, fp), from multiple b-values at a fixed diffusion time, can be used to assess the perfusion status of tumor and they are also associated with tumor interstitial fluid pressure. The IVIM effect has been observed in various cancer types (33-39) and animal tumor models. The product fp\*Dp - a quantity including both blood volume and velocity information - is considered as a parameter analogous to (iii) perfusion flow .