Doctors can get ‘second opinion’
Motivated by some high-profile cases of misdiagnosed mammograms, a research team at the University of Regina has developed a computer-aided diagnosis that can provide a second opinion for radiologists.
After Wei Peng, Esam Hussein and Rene Mayorga from the Faculty of Engineering and Applied Science at the U of R read about an investigation into how a lab in St. John’s, N.L., misinterpreted test results of nine breast cancer patients in 2014, they started work on an “Automated Confirmatory System for Analysis of Mammograms.”
While breast cancer is the most common cancer among women, the researchers note in their paper: “The accuracy and diagnosis depends on both the quality of the mammographic images and the ability of the radiologist to interpret those images.”
The U of R research team created a computer system that analyzes the same patient mammograms available to radiologists and provides a diagnosis that can serve as “a second opinion” to imaging specialists.
“We want to maximize the benefits of doing a mammogram by giving as much information to the physician and patient as possible,” Hussein said.
The analysis can confirm a radiologist’s diagnosis or “If the diagnosis is different from that of the doctor, then the doctor can take a second look,” Mayorga said.
The researchers stress their aim is not to replace the expert judgment of radiologists, but to provide a confirmatory tool.
The team worked with two large independent data sets from which large numbers of random mammogram images were extracted to train, test and validate the system. The team reports a diagnostic accuracy of at least 92 per cent.
“Our system provides clear filtered-segmented and contrast-enhanced image to radiologists for further physical diagnosis,” Mayorga said.
The system can be implemented in a small computer, such as a laptop.
“It’s very easy to use it,” Mayorga said. “One of the things that we’d like to do is test our system with some radiologists here in Regina or in Saskatchewan.”
Typically, mammograms involve two views, Hussein said. Down the road, he hopes the computer system will provide a three-dimensional view of a mammogram.
“That will give the radiologist some idea of where the tumour is, how big it is and maybe even how dense it is,” he said.
The researchers are also interested in working on other computer aided diagnosis to assist radiologists in detecting other cancers.