New MR technique quickly builds 3-D images of knees
faster magnetic resonance imaging (MRI) data-acquisition
technique will cut the time many patients spend in a cramped MR scanner,
yet deliver more precise 3-D images of their bodies.
Developed by a biomedical engineer at the University
of Wisconsin, the faster technique will enable clinics to image more
patients—particularly the burgeoning group of older adults with
osteoarthritis-related knee problems—and can help researchers
more rapidly assess new treatments for such conditions.
MR has long been touted as the ideal method for capturing
3-D images of the human body. “But unfortunately, it’s kind
of a slow technique,” says Associate Professor Walter
Block, also of medical physics. “At one time, you can only
sample a few pieces of information needed to build the image.”
Consequently, most MR technicians acquire images as
a series of 2-D slices, which yield high resolution in a single plane
and poor resolution in the remaining direction, says Block.
To capture an image, an MR scanner commonly conducts
hundreds to thousands of little "experiments," or encodings,
that help to make up the big picture.
Block’s data-acquisition technique capitalizes on recent MR hardware
advances that, coupled with a clever way of maintaining a high-level
MR signal throughout the scan, will speed an MRI session. “But
to maintain the high-level signal,” he says,” you need to
be able to complete each of these smaller encodings within a couple
of milliseconds.”
Rather than using the conventional Cartesian raster
method, which sweeps horizontally to gather MR data, Block’s technique
acquires the body’s signals radially, in a way that looks somewhat
like a toy Koosh ball. “The Koosh ball is much more efficient
in terms of doing these encodings very rapidly,” he says. “We
can essentially acquire data during the whole experiment, where in the
Cartesian case, a lot of time was spent either prepping for the experiment
or returning it to the steady state so that you could do the next experiment.
What we’re doing now is capable of a study that you can visualize
in any plane in about the same time as people are doing one plane.”
For example, when imaging a joint like the knee—Block’s
inspiration for developing the new technique—suppressing the fat
signal in bone provides image contrast between bone and the cartilage
surface. The conventional data-acquisition method would spend half its
scan time suppressing the signal from fat, instead of imaging cartilage.
However, Block’s technique exploits the difference in resonant
frequencies between fat and water. During the scan time, then, the technique
maximizes each component of the image, so that a technician can view
any aspect.
High-resolution 3-D images are important not only
from diagnostic and clinical standpoints, but also to help patients
better understand their health conditions, says Block. “If you
could actually look at a 3-D model from different perspectives, you’d
have a much better chance to make sense of the pain you’re feeling,
your doctor’s diagnosis, and your treatment options,” he
says.
The technique, which Block patented through the Wisconsin
Alumni Research Foundation, also will make it easier to image parts
of the body, such as the heart or abdomen, in which motion is a factor.
In related research, Block also has developed an
algorithm that, within less than a second, can calibrate an MR system
to use non-conventional methods of data acquisition, yet produce clearer
images.
Grants from the National Institutes of Health, the
W.H. Coulter Foundation, and the UW-Madison Graduate School fund Block’s
research.
