AI Improves Early Detection of Interval Breast Cancers

Interval breast cancers, which occur between routine screenings, are easier to treat when detected earlier. Early detection can reduce the need for aggressive treatments and improve the chances of better outcomes. A new study suggests that artificial intelligence (AI) could help identify these cancers before they progress to more advanced and harder-to-treat stages, potentially leading to better screening practices, earlier treatments, and improved patient outcomes.

The study, led by researchers at UCLA Health Jonsson Comprehensive Cancer Center (Los Angeles, CA, USA), demonstrated that AI was capable of detecting certain "mammographically-visible" types of interval cancers at earlier stages by flagging them during the screening process. These include tumors that are visible on mammograms but missed by radiologists, or those with subtle signs on the mammogram that are easy to overlook due to their faintness or because they fall below the level of detection by the human eye. Researchers estimate that integrating AI into screening could reduce the occurrence of interval breast cancers by as much as 30%. While similar research has been conducted in Europe, this study is one of the first to investigate the use of AI in detecting interval breast cancers in the United States. The researchers also note significant differences between U.S. and European screening practices. In the U.S., most mammograms are conducted using digital breast tomosynthesis (DBT), commonly known as 3D mammography, with patients typically screened annually. In contrast, European programs usually use digital mammography (DM), or 2D mammography, with screenings every two to three years.

The retrospective study, published in the Journal of the National Cancer Institute, analyzed data from nearly 185,000 mammograms taken between 2010 and 2019, which included both DM and DBT. The study focused on 148 cases where women were diagnosed with interval breast cancer. Radiologists reviewed these cases to understand why the cancer was not detected earlier. The researchers adapted a European classification system to categorize the interval cancers into six types: Missed reading error, minimal signs–actionable, minimal signs–non–actionable, true interval cancer, occult (truly invisible on mammogram), and missed due to a technical error. The researchers then applied AI software called Transpara, developed by ScreenPoint Medical (Nijmegen, Netherlands), to the initial screening mammograms taken before the cancer diagnosis to determine if it could detect subtle signs of cancer that had been missed by radiologists or at least flag them as suspicious. The tool rated each mammogram on a scale from 1 to 10 for cancer risk, with scores of 8 or higher flagged as potentially concerning.

The AI system flagged 76% of mammograms that were initially read as normal but later linked to interval breast cancer. It identified 90% of missed reading error cases, where the cancer was visible on the mammogram but missed or misinterpreted by the radiologist. The AI system detected about 89% of minimal-signs-actionable cancers, which showed subtle signs that could reasonably have been acted upon, and flagged 72% of those with minimal signs that were likely too subtle to prompt action. For cancers that were occult, or completely invisible on the mammogram, the AI flagged 69% of cases. It was somewhat less effective in identifying true interval cancers, those that were not present during the screening but developed later, flagging around 50% of those cases. Further large-scale prospective studies are needed to understand how radiologists would incorporate AI into practice and to address key questions, such as how to handle cases where AI flags areas that aren't visible to the human eye, especially when the AI is not always accurate in pinpointing the exact location of the cancer.

“While AI isn’t perfect and shouldn't be used on its own, these findings support the idea that AI could help shift interval breast cancers toward mostly true interval cancers,” said Dr. Tiffany Yu, assistant professor of Radiology at the David Geffen School of Medicine at UCLA and first author of the study. “It shows potential to serve as a valuable second set of eyes, especially for the types of cancers that are the hardest to catch early. This is about giving radiologists better tools and giving patients the best chance at catching cancer early, which could lead to more lives saved.”

Related Links:
UCLA Health Jonsson Comprehensive Cancer Center
ScreenPoint Medica