Breast imagers and artificial intelligence (AI) experts have shown that a new AI algorithm measures breast density with accuracy comparable to an experienced breast imager, according to new research published online Oct. 16 in Radiology.
"We're dependent on human qualitative assessment of breast density and that approach has significant flaws. We need a more accurate tool," lead author Constance D. Lehman, MD, PhD, from Massachusetts General Hospital (MGH) in Boston, said in a prepared statement.
To train and test the deep convolutional neural network (CNN)-based algorithm, Lehman and colleagues used more than 41,000 digital mammograms obtained in over 27,00 women between January 2009 to May 2011 from MGH. Eight radiologists then reviewed the accuracy of 10,763 mammograms the algorithm determined were either dense or non-dense tissue.
Overall, the AI algorithm evaluated the mammograms for dense breast tissue with 94 percent accuracy. Additionally, the researchers noted that reader variability may affect the remaining six percent disagreement rate between the radiologists and the algorithm.
"The study results show that the algorithm worked remarkably well, but what's more important is that it is being used every day to measure breast density in mammograms at a major hospital [MGH],” co-author Regina Barzilay, PhD, professor of computer science and electrical engineering at the Massachusetts Institute of Technology (MIT), in Cambridge, said in a prepared statement. She added that the system has been used at MGH since January and has processed roughly 16,000 mammograms.
The algorithm has the potential to standardize and automate routine breast density assessment, according to the researchers, and could outperform current predicative models that predict breast cancer risk in more vulnerable populations, such as African-American women.
"With AI, we now have the ability to leverage vast amounts of information into more personalized, more targeted care for our patients," Lehman said. "In the case of breast cancer, we can better predict how likely a woman will have cancer in her future and improve the chances that it will be treated successfully."