Tim Willert
GLENDALE -- Call it "spell-check" for radiologists.
Except these doctors aren't correcting spelling errors, they're trying
to detect early signs of breast cancer.
Since January, Glendale Adventist Hospital radiologists have used a
new computer-assisted detection system called an ImageChecker. The device
recognizes minute calcium deposits or very subtle masses that can escape
the radiologist's eye during a standard mammogram.
"It uses artificial intelligence to review the scanned image of the
mammogram by computer in order to detect suspicious masses or
calcification," said Dr. Alan Miyamoto, a radiologist at Glendale
Adventist. "The Image Checker helps find the subtle features that might
make a benign lesion become suspicious."
The system, which was developed by R2 Technology, a Los Altos-based
company, works with the regular X-ray image taken in a mammogram.
The film is run through a computer processor that creates a digital
image. The computer, which is trained to recognize certain subtle
patterns, scans the image and marks suspicious areas, indicating places
that deserve a second look.
"It's like a double check or a spell-checker on a word processor,"
said Jimmy Roehrig, chief science officer for R2 Technology. "This
machine is something that doesn't get tired; it keeps pointing out the
same sorts of things over and over again."
Glendale Adventist paid $200,000 for the ImageChecker -- approved in
1998 by the Food and Drug Administration -- and is the only hospital in
L.A. County using the system.
During FDA testing, the ImageChecker caught up to 85% of cancerous
lesions that were missed during standard mammograms a year earlier,
Roehrig said.
"It means a very high degree of performance on our part, and a high
degree of confidence on the patient's part," Miyamoto said.
