AI Helps Rule Out Cancer in Women With Dense Breasts

Written by Kara Morris
HealthDay reporter

WEDNESDAY, Oct. 6, 2021 (HealthDay News) — While mammograms have reduced deaths by detecting breast cancer when they are small and easier to treat, they are less effective for women with denser breasts.

However, a new study finds that a complementary MRI scan could make a difference for these women, who are more likely to develop breast cancer. A new technology is being used to speed up the process.

Dutch researchers report that artificial intelligence can quickly and accurately screen MRI scans to rule out breast cancer in the majority of those without breast cancer — allowing radiologists to work on more complex cases.

In the Examination of Dense Tissue and Early Tumors of the Breast Experiment (DENSE), researchers trained an AI technique to distinguish between breasts with and without lesions.

“The DENSE trial showed that additional MRI screening for women with very dense breasts was beneficial,” said lead author Eric Verborg of the University Medical Center in Utrecht in the Netherlands. “On the other hand, the DENSE trial confirmed that the vast majority of women examined had no suspicious findings on MRI.”

Mammography is less sensitive in women with very dense breasts than in women with fatty tissue. Women with very dense breasts have up to six times the risk of developing breast cancer compared to women with oily breasts. Their risk is twice that of women on average.

The study analyzed MRI scans of nearly 9,200 extremely dense breasts. Of these, more than 8,300 had no outgrowths and 838 had at least one outgrowth. Of these, 77 had cancer.

The model identified 91% of the MRI with lesions for radiologist review. According to the study, about 40% of lesion-free MRIs were rejected without missing any cancers.

The results were published October 5 in the journal rays.

“We have shown that it is possible to safely use AI to reject breast MRI scans without missing any malignancy,” Verborg said in a press release. “The results were better than expected. Forty percent is a good start. However, we still have 60 percent to improve.”

Verburg said this AI-based system has the potential to significantly reduce radiologists’ workload.

“This approach could first be used to help radiologists to reduce overall reading time,” Verburg said. “So more time can be allowed to focus on really complex breast MRI scans.”

more information

The American Cancer Society has more on dense breasts.

Source: Radiological Society of North America, press release, October 5, 2021


Leave a Comment