PURPOSE: To establish a practical method that uses concurrent b0 images to standardize the display conditions for diffusion-weighted images (DWI) that vary among institutions and interpreters. METHOD: Using identical parameters, we obtained DWI for 12 healthy volunteers at 4 institutions using 4 MRI scanners from 3 vendors. Three operators manually set the window width for the images equal to the signal intensity of the normal-appearing thalamus on b0 images and set the window level at half and then exported the images to 8-bit gray-scale images. We calculated the mean pixel values of the brain objects in the images and examined the variation among scanners, operators, and subjects. RESULT: Following our method, the DWI of the 12 subjects obtained using the 4 different scanners had nearly identical contrast and brightness. The mean pixel values of the brain on the exported images among the operators and subjects were not significantly different, but we found a slight, significant difference among the scanners. CONCLUSION: Determining DWI display conditions by using b0 images is a simple and practical method to standardize window width and level for evaluating diffusion abnormalities and decreasing variation among institutions and operators.
PURPOSE: To establish a practical method that uses concurrent b0 images to standardize the display conditions for diffusion-weighted images (DWI) that vary among institutions and interpreters. METHOD: Using identical parameters, we obtained DWI for 12 healthy volunteers at 4 institutions using 4 MRI scanners from 3 vendors. Three operators manually set the window width for the images equal to the signal intensity of the normal-appearing thalamus on b0 images and set the window level at half and then exported the images to 8-bit gray-scale images. We calculated the mean pixel values of the brain objects in the images and examined the variation among scanners, operators, and subjects. RESULT: Following our method, the DWI of the 12 subjects obtained using the 4 different scanners had nearly identical contrast and brightness. The mean pixel values of the brain on the exported images among the operators and subjects were not significantly different, but we found a slight, significant difference among the scanners. CONCLUSION: Determining DWI display conditions by using b0 images is a simple and practical method to standardize window width and level for evaluating diffusion abnormalities and decreasing variation among institutions and operators.