Mammography Plus Ultrasound Better Than Mammography With AI for Dense Breasts
A mammogram with a supplemental ultrasound about a month later was better at detecting cancer in dense breasts than a mammogram read with the help of artificial intelligence (AI), according to a Korean study.
The research was published in the July 26, 2023, issue of the American Journal of Roentgenology. Read the abstract of “Screening in Patients With Dense Breasts: Comparison of Mammography, Artificial Intelligence, and Supplementary Ultrasound.”
Screening for cancer in dense breasts
If you have dense breasts, it means your breasts have a large amount of fibrous and glandular tissue and not a lot of fatty tissue.
Dense breasts make it hard for doctors to see cancer on a mammogram. Research suggests that mammograms can miss about half of the cancers in women with dense breasts. Besides making it harder to find cancer, dense breast tissue itself is linked to a higher risk of breast cancer. Doctors aren’t sure why.
Studies show that 3D mammograms are more accurate than traditional 2D mammograms for dense breasts. Doctors also may recommend supplemental breast cancer screening for women with dense breasts, often with ultrasound or MRI.
Using AI to read mammograms
To teach AI technology to read mammograms, technicians input information from millions of mammograms. The AI software creates a mathematical representation — an algorithm — of what a normal mammogram looks like and what a mammogram with cancer looks like. The AI system can see more detail in each mammogram than the human eye can, and checks each image against the standards to find any abnormalities.
AI-based mammogram support systems are considered by many to be a second set of eyes for radiologists and are considered one of the most promising applications for AI in radiology. Early studies suggest that AI-supported mammograms may find more cancers than a standard double reading by two radiologists.
About the study
Although mammograms are less accurate in women with dense breasts, there is no standard for supplemental screening for dense breasts. This is because research on different types of supplemental breast cancer screening has offered mixed results. It’s also not clear if using AI to help read mammograms would improve cancer detection in dense breasts.
In this study, the researchers compared the accuracy and recall rate of three screening methods:
mammography alone
AI-supported mammography
ultrasound
The study included 1,325 Korean women with dense breasts. The women’s average age was 53 and they all had a screening mammogram and supplemental ultrasound within one month of the mammogram between January and December 2017.
The researchers had earlier mammogram and ultrasound results for comparison for about 91% of the women.
During the study period, doctors diagnosed 12 breast cancers:
six were invasive ductal carcinoma (IDC)
six were DCIS
Of the 12 cancers:
mammography found eight: four IDC and four DCIS
AI-supported mammography found nine: five IDC and four DCIS
ultrasound found eight: six IDC and two DCIS
AI-supported mammography found one case of IDC and one of DCIS that were not found by mammography and three cases of DCIS that were not found by ultrasound. AI-supported mammography did not find any cancers that were missed by both mammography and ultrasound.
Ultrasound found two cases of IDC and two of DCIS that were not found by mammography and one case of IDC and one of DCIS that were not found by AI-supported mammography.
Mammography found one case of DCIS that was not found by AI-supported mammography and four cases of DCIS that were not found by ultrasound.
Cancer detection rates (cancers found per 1,000 women) for each screening method were:
6 for mammography
6.8 for AI-supported mammography
6 for ultrasound
The recall rate — the number of women called back for more screening because a screening test showed a suspicious area — was:
4.4% for mammography
11.9% for AI-supported mammography
9.2% for ultrasound
When a breast cancer screening test shows an abnormal area that looks like a cancer but turns out to be normal, it’s called a false positive. Ultimately the news is good: no breast cancer. But the suspicious area usually requires follow-up with more than one doctor, extra tests, and extra procedures, including a possible biopsy.
False positive rates were:
3.8% for mammography
8.6% for ultrasound
11.2% for AI-supported mammography
The researchers then looked at the cancer detection and recall rates when the tests were combined.
Cancer detection rates were:
7.5 for mammography plus AI-supported mammography
9.1 for mammography plus ultrasound
9.1 for mammography plus AI-supported mammography and ultrasound
Recall rates were:
14.9 for mammography plus AI-supported mammography
11.7 for mammography plus ultrasound
21.4 for mammography plus AI-supported mammography and ultrasound
The accuracy rates (the percentage of cancers found) were:
85.7% for mammography plus AI-supported mammography
89.2% for mammography plus ultrasound
79.5% for mammography plus AI-supported mammography and ultrasound
“Mammography with supplementary ultrasound showed higher accuracy, higher specificity, and lower recall rate in comparison to mammography with AI as well as in comparison to mammography with both ultrasound and AI,” the researchers concluded. “The findings fail to show benefit of AI with respect to screening mammography performed with supplementary breast ultrasound in patients with dense breasts.”
What this means for you
If you have dense breasts, then it’s important to know you may need a breast cancer screening plan that includes more than just mammograms.
To develop a screening plan that’s best for your unique situation, it makes sense to follow the American College of Radiology breast cancer screening guidelines and have a breast cancer risk assessment by age 25.
You and your doctor are likely to discuss the following points:
family history of breast or other related cancers (ovarian, melanoma)
any test results for gene mutations linked to higher breast cancer risk
results of past breast biopsies, even if they were benign
personal history of radiation treatment to the face, chest, or both before age 30
breast density
weight
level of physical activity
any use of post-menopausal combined hormone replacement therapy (HRT)
alcohol consumption, if you regularly drink more than three alcoholic beverages a week
the amount of processed food and trans fats you eat
your smoking history
whether or not you have carried a pregnancy to term or have breastfed
If you decide that you have a higher-than-average risk of developing breast cancer, ask your doctor when you should start annual mammograms and whether you need supplemental screening with ultrasound.
— Last updated on October 5, 2023 at 3:15 PM