Mono-exponential, diffusion kurtosis and stretched exponential diffusion MR imaging response to chemoradiation in newly diagnosed glioblastoma
To quantify changes and prognostic value of diffusion MRI measurements obtained using mono-exponential, diffusion kurtosis imaging (DKI) and stretched exponential (SE) models prior and after chemoradiation in newly diagnosed glioblastoma (GBM).
Diffusion-weighted images (DWIs) were acquired in twenty-three patients following surgery, prior chemoradiation and within 7 days following completion of treatment, using b-values ranging from 0 to 5000s/mm2. Mono-exponential diffusion (apparent diffusion coefficient: ADC), isotropic (non-directional) DKI model with apparent diffusivity (Dapp) and kurtosis (Kapp) estimates as well as SE model with distributed-diffusion coefficient (DDC) and mean intra-voxel heterogeneity (α) were computed for all patients prior and after chemoradiation. Median values were calculated for normal appearing white matter (NAWM) and contrast-enhancing tumor (CET). The magnitudes of diffusion change prior and after chemoradiation were used to predict overall survival (OS).
Diffusivity in NAWM was consistent for all diffusion measures during chemoradiation, while diffusivity measurements (ADC, Dapp and DDC) within CET changed significantly. A strong positive correlation existed between ADC, Dapp, and DDC measurements prior to chemoradiation; however, this association was weak following chemoradiation, suggesting a more complex microstructural environment after cytotoxic therapy. When combined with baseline tumor volume and MGMT status, age and ADC changes added significant prognostic values, whereas more complex diffusion models did not show significant value in predicting OS.
Despite increased tissue complexity following chemoradiation, advanced diffusion models have longer acquisition times, provide largely comparable measures of diffusivity, and do not appear to provide additional prognostic value compared to mono-exponential ADC maps.
KeywordsGlioblastoma MRI Diffusion models Biomarkers
Magnetic resonance imaging
Diffusion weighted imaging
Apparent diffusion coefficient
Diffusion corrected apparent diffusion coefficient
Apparent kurtosis coefficient
Distributed diffusion coefficient
- α (alpha)
Intra-voxel diffusion heterogeneity
Normal appearing white matter
National Brain Tumor Society (NBTS) Research Grant (Ellingson, Cloughesy); American Cancer Society (ACS) Research Scholar Grant (RSG-15-003-01-CCE) (Ellingson); Art of the Brain (Cloughesy); UCLA SPORE in Brain Cancer (NIH/NCI 1P50CA211015-01A1) (Ellingson, Cloughesy); NIH/NCI 1R21CA223757-01 (Ellingson); NIH/NCI 1R21CA167354 (Ellingson).
- 20.Falk DA, Nilsson M et al (2017) Glioma grade discrimination with MR diffusion kurtosis imaging: A meta-analysis of diagnostic accuracy. Radiology 4:171315Google Scholar