Reviewed by 
It’s no secret that mammography misses many cancers. We've been talking about this problem for ages - how mammograms can be like trying to find a polar bear in a snowstorm, especially for women with dense breast tissue. It's been a real head-scratcher for doctors and a worry for patients.
Now, there may be a workaround. And it involves artificial intelligence or AI.
Key Takeaways
- The AISmartDensity tool can identify women at high risk of missed breast cancer after a negative mammogram.
- The AI tool is more effective than doctors at selecting a smaller, high-risk group of women for additional screening such as MRI.
- Early results from the ScreenTrustMRI trial indicate that using AI to guide supplemental MRI screenings significantly improves the detection of aggressive breast cancers, particularly in women with dense breast tissue.
When Mammograms Miss Breast Cancer
Many women need to undergo supplemental breast cancer screening with magnetic resonance imaging (MRI). This additional step can reduce the number of cancers missed by a mammogram. However, since qualified MRI staff are lacking, and the equipment is expensive to purchase and operate, the cost-effectiveness is unconvincing. So, the use of MRI for breast cancer screening is currently limited.
But doctors are excited about a new tool that can help…
A recently developed Artificial Intelligence (AI) tool to choose who gets supplemental MRI scans after negative mammography. The tool identifies women who are most at risk of undetected cancer following negative mammography.
The ScreenTrustMRI randomized clinical trial used this tool, called AISmartDensity, to score each mammogram. They offered trial participation to individuals with a negative mammogram and a high AI score (top 6.9%). When women agreed to participate, they were randomly assigned to one of two groups – those receiving supplemental MRI and those not receiving MRI.
Found Four Times More Cancers
According to early reports from a randomized trial slated to be completed in August of 2025, the AI tool led to the detection of many cancers that would otherwise have been missed.
The main goal (or "primary endpoint" in doctor-speak) is to catch the big, bad cancers—we're talking about the troublemakers that either pop up between regular screenings and grow larger than the size of a grape (that's about 15 mm) or have already started spreading to the lymph nodes. Doctors give these cancers about 27 months to potentially show themselves before they check again.
Now, here's where it gets really interesting. The researchers are also looking at how many cancers this method can spot right off the bat. And boy, did they find something exciting! Compared to the old-school way, this new AI whiz-kid is like a cancer-finding superhero. It's nearly four times better at spotting potential problems - finding about 64 cancers per 1,000 MRI exams, while the traditional method only catches about 16 or 17.
It's like upgrading from a magnifying glass to a super-powered microscope in this treasure hunt for health. Pretty impressive, right?
Found More Aggressive Cancers
Results showed that using an AI-based score to select a small group for supplemental MRI after negative mammography detected many missed cancers, bringing the cost more in line with traditional mammography.
Frederik Strand, MD, PhD, of Karolinska University Hospital in Stockholm, Sweden, and colleagues say using the AI tool resulted in a four-fold higher cancer detection rate than traditional breast density measures.
Moreover, “Most additional cancers detected were invasive and several were multifocal, suggesting that their detection was timely,” Dr. Strand and colleagues reported.
A multifocal cancer is a breast cancer in which multiple tumors show up in the same area of the breast. The multiple tumors are generally in the same section or quadrant of the breast.
Solves Numerous Diagnostic Problems
Dr. Strand and colleagues point out that although early cancer detection through mammography may have decreased breast cancer, many breast cancers in screened women are found as interval cancers – those that are detected before the next planned screening. And these are only usually detected because women show symptoms. This new AI tool could help locate these cancers earlier by identifying these women for screening earlier.
What’s more, the sensitivity of mammography is lower for women with extremely dense breasts, while the sensitivity of MRI is unaffected. Put simply, MRI is more accurate, especially for women with dense breasts.
Dr. Strand’s team hypothesizes that an AI-based imaging analysis could be more efficient in terms of the number of MRI exams that lead to a cancer diagnosis.
The ScreenTrustMRI enrolled 59,354 women whose mammograms were screened with AISmartDensite. Of those, 3,821 were eligible for MRI due to a “very high” AISmartDensity score (top 6.9%) and 1,315 were eventually randomized, with half (663) assigned to get MRI. Only 559 completed the MRI and formed the final cohort for the analysis. Median age was 56. Of that group, 22 women (4%) had a previous history of breast cancer and 104 (19%) had a family member with a history of breast cancer.
Our Takeaway
This new AI tool doesn’t resolve the problem of the compression of the breast during a mammogram, potentially releasing and spreading cancer cells in the breast. But it does help reduce total dependence on mammograms for cancer detection. Thermography may also be a good option to consider.
Thermography uses infrared technology to detect and record the skin's temperature on the breasts' surface. The camera detects heat patterns and blood flow. Areas with higher temperatures might indicate increased blood flow, which can be associated with cancerous tissue. By comparing thermograms taken over time, doctors can identify changes in temperature patterns that may indicate cancer development.
Summary
Mammography often misses cancers, especially in women with dense breast tissue. The use of supplemental MRI screening can reduce missed cancers, but it's costly and resource-intensive. A new AI tool called AISmartDensity has been developed to select women at high risk of undetected cancer for supplemental MRI screening after negative mammograms. This AI tool has proven to be nearly four times more efficient than traditional methods, leading to higher cancer detection rates. Early results from the ScreenTrustMRI randomized clinical trial show that using AISmartDensity significantly improves the detection of invasive cancers, making supplemental MRI more cost-effective.
The Expert:
