A new study led by investigators at the UCLA Health Jonsson Comprehensive Cancer Center suggests that artificial intelligence (AI) could significantly enhance the detection of interval breast cancers—those that develop between routine screenings—before they become difficult to treat. This development promises earlier diagnosis, more effective treatments, and better patient outcomes.
Published in the Journal of the National Cancer Institute, the research found that AI could identify mammographically-visible types of interval cancers earlier by flagging them at screening. These include tumors visible on mammograms but missed by radiologists or those with very subtle signs that are easily overlooked due to faint or indiscernible features.
Experts estimate that integrating AI into screenings could help decrease the incidence of interval breast cancers by up to 30%.
Dr. Tiffany Yu, assistant professor of Radiology at UCLA, emphasized the significance: ‘This finding is important because these interval cancer types could be caught earlier when the cancer is easier to treat. For patients, early detection can make all the difference, leading to less aggressive treatments and better outcomes.’
The study analyzed nearly 185,000 past mammograms from 2010 to 2019, covering both digital mammography (DM) and digital breast tomosynthesis (DBT or 3D mammography). Researchers categorized 148 interval cancer cases and evaluated why they were missed initially, using a European classification system. They applied a commercially available AI software called Transpara to the mammograms taken before diagnosis to assess its ability to detect subtle signs of cancer missed by radiologists.
Results showed that AI flagged 76% of mammograms that had been initially read as normal but were linked to later interval cancers. It also identified 90% of cases where cancer was visible but overlooked by radiologists, and about 89% of cancers showing minimal actionable signs. Even for occult cancers, invisible on mammograms, AI flagged 69% of cases. However, its effectiveness dropped to about 50% for true interval cancers that developed later.
While these findings are promising, researchers note the need for larger prospective studies to understand how radiologists will use AI in real-world settings and to resolve issues related to false positives and pinpointing exact lesion locations.
Dr. Hannah Milch, senior author, stated, ‘AI isn’t perfect and should not be used in isolation. However, it could serve as a valuable second set of eyes, especially in detecting the hardest-to-find cancers early.’
Light on the future of AI in healthcare suggests a dynamic shift toward earlier detection and personalized treatment strategies, potentially transforming breast cancer management.
Concurrently, AI’s rapid adoption extends well beyond healthcare, significantly impacting industries and economies. Recent reports from AWS reveal that UK businesses are integrating AI every 60 seconds, with adoption rates climbing to 52% from 39% last year. Notably, startups are leading in strategic planning, with twice the likelihood of having comprehensive AI strategies compared to large enterprises.
Despite this progress, a notable skills shortage persists, with 38% of businesses citing lack of talent as a barrier, and all industries facing an average recruitment time of 5.5 months to acquire skilled digital professionals. To address this, AWS announced a UK initiative to train 100,000 people in AI skills by the end of the decade, aiming to unlock an estimated £45 billion in public sector savings and productivity benefits.
Nevertheless, geopolitical tensions and reliance on Taiwanese microchips—over 90% of which are produced by TSMC—pose challenges to the industry’s stability. These microchips are crucial for AI hardware, and disruptions, as seen during the COVID pandemic, can halt critical supply chains, affecting everything from car manufacturing to advanced computing.
With ongoing trade wars and shifting global alliances, some experts speculate a move toward reshoring manufacturing to the US, facilitated by new investments like TSMC’s factories in Arizona. Such changes could influence the future landscape of technological innovation and economic resilience.
Overall, while the AI industry faces hurdles, its potential for revolutionizing healthcare, boosting economic growth, and enhancing global technological development remains substantial. As Dr. Yu remarked, ‘AI could help shift interval breast cancers toward mostly true interval cancers,’ representing a significant step forward in early detection and treatment.’
