Purpose: To investigate tissue changes using intravoxel incoherent motion (IVIM) magnetic resonance imaging (MRI), and MR texture analysis of diffusion weighted (DW) images for patients with cancerous lesions of the liver undergoing stereotactic body radiotherapy. Two cohorts were investigated, the first consisting of five healthy volunteers, and a second with two patients.
Methods: Cohorts were imaged on a 3T Philips Ingenia MRI, acquiring anatomical, CINE, and diffusion weighted images with eight unique b-values ranging from 0 to 1000s/mmÂ² (b = 0, 20, 40, 80, 100, 300, 600, and 1000). Subjects were imaged while maintaining breath hold, using an inhouse respiratory motion management system, which also provided visual biofeedback for breath holds. Healthy volunteers were imaged twice, with one week between scans. Two patients were imaged prior to receiving treatment, one week post (A), and 1-2 months post treatment (B). The DW images were denoised using non-local means filtering, and analyzed using a biexponential model, fitting for diffusion, perfusion, and perfusion fraction. ROIs were created in healthy and diseased tissue, for all three patient time points. First and second order texture features were extracted from the ROIs using histogram and gray level co-occurrence matrices (GLCM), and included entropy, range, standard deviation (STD), contrast, correlation, energy, and homogeneity. Correlation between texture features and IVIM parameters was assessed using Pearson correlation coefficient (PCC).
Results: IVIM analysis resulted in mean Dfast of 55.13x10â?»Â³mmÂ²/s, 54.258x10â?»Â³mmÂ²/s, and 8.50x10â?»Â³mmÂ²/s and mean Dslow of 0.66x10â?»Â³mmÂ²/s, 0.93x10â?»Â³mmÂ²/s, and 1.03x10â?»Â³mmÂ²/s for diseased tissue pre- and post-treatment time points A and B respectively. PCC was greater than 0.9 for Dslow and range, STD, and contrast for post-treatment A images.
Conclusion: Initial results showed a difference in pre- and post-treatment diseased tissue for patients receiving SBRT for IVIM and Texture analysis parameters.
Funding Support, Disclosures, and Conflict of Interest: This research has been supported by Institutional Research Grant IRG-14-192-40 from the American Cancer Society.