Introduction:
Apparent Diffusion Coefficient (ADC). This measure takes advantage of MRI's ability to detect subtle fluid changes associated with various disease states. For example, it can identify brain tissue in the early stages of stroke that is ischemic yet potentially recoverable, thus differentiating it from either normal or permanently damaged surrounding tissue. We cannot see the ischemia directly: there is no pooling of blood in this early stage as the blood-brain barrier has not yet been broken. But we can see it indirectly: the ischemia causes cytotoxic edema, in which some cells die and lyse their contents into the surrounding interstitial fluid, altering the diffusion rate of water molecules through it. While we see neither the ischemia nor the subsequent subtle edema with MRI, we can see their effects on diffusion. How is this done? A given type of tissue has a characteristic rate (coefficient) of diffusion during the relaxation time T1 or T2. The greater the coefficient of diffusion, the brighter the signal intensity on the ADC image, for a given magnetic field strength. As the intensity at a specific brain location is dependent in a log-linear relationship on the magnetic strength at the corresponding location in the magnetic gradient field, changing the gradient and regressing this relationship produces the diffusion coefficient (See Note). Altering b by changing the field gradient G, and then regressing these b values with the corresponding changes in the dependent variable S, yields the coefficient D. If this calculated coefficent is different from the known normal one for that particular tissue type, cytotoxic edema is likely.
Methods:
MR imaging was performed on a 1.5 Tesla Magnetom Vision (Siemens, Erlangen, Germany) with 25 mT/M gradients and a 600 msec rising time. A quadrature cp head coil was used. A spin-echo EPI sequence with a diffusion-sensitive gradient applied along one axis acquired the data. Five to seven diffusion-sensitive gradients (chosen from the set: 0, 2.5, 5, 7.5, 10, 12.5, 15.0 mT/M) generated the whole series of images. Acquisition time was 100 msec per image. Five different gradient amplitudes were implemented in five separate sequences. Scanning time for 10 slices at each diffusion gradient took a total of about two seconds. A 10-second interval between sequences was needed for T1 relaxation. Total scanning time for the study was about one minute. Each slice represented a 5 mm thick transverse cut, with a matrix of 128X128 interpolated to 256X256, and a TR/TE ratio equal to infinity/84 msec. Raw images were registered before analysis, with a Georgetown software program written in IDL, to correct for phase distortion caused by eddy current inteference.
Results:
Non-registered ADC image
Registered ADC image.
Footnotes:
For those who are so inclined: I = e^(-Db) + Io, where I is the MR signal intensity, D is the coefficient of diffusion, and the constant b is proportionate to the absolute value of G^2, where G is the magnetic field gradient strength. Taking the natural log of both sides gives us:
ln(I) = -Db + ln(Io).
The full equation for b was derived by Jeff O`Neill of Washinton University and can be found in "JMRI Nov/Dec 1994, V4 #6, p. 787"