Women who chose to enroll in a self-pay, artificial intelligence (AI)-enhanced breast cancer screening program were more likely to have cancer detected, researchers reported.
Across 10 clinical practices, the overall cancer detection rate was on average 43% higher for those who enrolled in the AI program versus unenrolled women, and further analyses showed that 21% of that increase could be attributed to the AI component, reported Bryan Haslam, PhD, of DeepHealth in Somerville, Massachusetts.
The remaining increase in detection was attributed to the fact that higher-risk patients chose to enroll in the AI program more frequently, he noted during the Radiological Society of North America annual meeting in Chicago.
“These data indicate that many women are eager to utilize AI to enhance their screening mammogram, and when AI is coupled with a safeguard review, more cancers are found,” said co-author Gregory Sorensen, MD, also from DeepHealth.
The rate at which women were called back for additional imaging was 21% higher for enrolled versus unenrolled women, and the positive predictive value for cancer was 15% higher for the enrolled women, which indicated that each recall led to more cancer diagnoses in the enrolled population, the researchers noted.
“The AI-enhanced program leverages FDA-cleared software in a novel workflow to help detect many more breast cancers,” Haslam said. “An expert breast radiologist provides a safeguard review in cases where there is discordance between the first reviewer and the AI.”
He said that the number of women seeking the enhanced screening — even if they have to pay for it — continues to grow “and the rate of cancer detection continues to be substantially higher for those women.”
Jessica Leung, MD, of the University of Texas MD Anderson Cancer Center in Houston, told MedPage Today that, “anecdotally, AI is a term that is in the public space, and in general, patients are attracted toward new technology, so I would not be surprised if women at higher risk of breast cancer show more acceptance of AI than physicians.”
The researchers noted that even though AI has shown promise in mammography as a “second set of eyes” for radiologists providing decision support, risk prediction, and other benefits, AI is not yet routinely reimbursed by insurance, which likely is slowing its adoption in the clinic.
“Some practices have elected to offer AI at additional cost, much like what was done when digital breast tomosynthesis was originally deployed,” they noted. “While quantification of benefit will require prospective controlled trials, and it is difficult to separate enrollment bias from the effectiveness of AI, we seek to share data from experience with initial implementations from several different practices that implemented a self-pay AI program.”
Erik Thompson, PhD, of Queensland University of Technology in Brisbane, Australia, said that “this study’s findings echo the kind of improvements being reported with accelerating frequency in the literature of specific studies in distinct cohorts. The results are not surprising and provide a validation of the improved functionality of mammography with AI in terms of positive predictive value, and the broad acceptance in the community. Mammography is such a great example of how AI can see reproducible, meaningful patterns better than the human eye.”
“The developments have accrued rapidly and, of course, it takes time for the health systems to keep up,” he told MedPage Today. “In the interim, it is great to have an avenue for the public to pay out of pocket, despite the access inequities that brings. I hope the study will provide support for improved access — reimbursement for AI-assisted mammography readings for all.”
For this study, a self-pay AI-driven screening mammography program was deployed across 10 clinical practices ranging from a few sites up to 64 sites at the largest practice. The researchers collected data on 747,604 women who underwent screening mammography over a 12-month period and who were offered the option to pay for the AI-driven enhanced review.
Of these women, 23% chose to enroll, with the enrollment rate increasing over time, with a present enrollment of 36% and growing.
![author['full_name']](https://clf1.medpagetoday.com/media/images/author/ESusman_188.jpg)
Disclosures
Haslam and Sorensen are employees of DeepHealth.
Leung and Thompson disclosed no relevant relationships with industry.
Primary Source
Radiological Society of North America
Source Reference: Haslam B, et al “Deep Health – patient self-pay for AI-driven enhanced review program in screening mammography: initial experience” RSNA 2024.
Please enable JavaScript to view the