The pros and cons of current breast screening modalities and the role of AI

CEM, like MRI, uses contrast agents, but comes at a lower cost. However, it also has drawbacks, such as the requirement for contrast injection.  Monticciolo expressed hope that ongoing research into noncontrast MRI or alternative contrast agents could help simplify MRI screening in the future.

Breast ultrasound faces challenges

Breast ultrasound, particularly automated breast ultrasound, has also been explored as a supplemental screening tool. While it offers the advantage of no radiation, it has not seen widespread adoption. Monticciolo pointed out that the introduction of digital breast tomosynthesis, which detects about half of the cancers that ultrasound finds after a negative mammogram, has made ultrasound less favorable as a supplemental screening tool. Additionally, ultrasound has been plagued by a high number of false positives, leading to unnecessary biopsies and patient anxiety.

"People didn't like having all those benign biopsies, and it hurt ultrasound's wider adoption," she said.

ABUS was designed to reduce labor, make exams more reproducible, and to reduce time spent on handheld breast imaging. But, she said it has not fully solved these issues either. Many ABUS scans still require follow-up with handheld ultrasound, limiting the time savings, she explained.

The role of AI in breast imaging

With the rise of digital breast tomosynthesis, which generates many more images per exam compared to standard mammography, radiologists are overwhelmed with data. Here, AI has the potential to make a significant impact. Monticciolo highlighted how AI is being used to triage large volumes of images, allowing radiologists to focus on high-risk cases or specific image slices.

"The hope is that AI can help us manage the volume," she said, noting that AI could eventually classify low-risk cases, reducing the workload for human readers.

However, AI in breast imaging remains in its early stages. While promising, the technology still requires continuous validation on large and diverse datasets to ensure accuracy.

"We’re not at a point where AI can read alone," Monticciolo cautioned, referencing a study in Sweden where AI and human readers combined yielded the best results, but AI alone missed several tumors.

Another area in breast imaging where AI might help is for risk assessment. Algorithms are already being developed to  determine which women might benefit from supplemental screening, such as MRI, by identifying those with a higher likelihood of developing breast cancer.

"It could help make MRI more useful by focusing on those who are at higher risk," Monticciolo explained.

The future of breast imaging

While current modalities each have their strengths and weaknesses, AI could be the key to improving efficiency and accuracy in breast cancer screening. By assisting in image analysis, helping triage patients, and guiding decisions on supplemental screening, AI is poised to become an integral part of breast imaging workflows. However, as Monticciolo emphasized, the technology is still evolving and, for now, human expertise remains essential in ensuring accurate diagnoses.

As AI advances, its integration into breast imaging may help address the growing shortage of breast imaging specialists, easing the burden on radiologists while enhancing cancer detection rates.

"Computers are getting better and better. If they can drive a car, they’ll eventually be able to help us with mammography," she said.

Watch another interview with Monticciolo on when to start breast screening.