AI Boosts Radiologists' Accuracy in Breast Cancer Detection

Artificial intelligence (AI) is significantly enhancing the accuracy of breast cancer detection by radiologists, according to a study published today in Radiology, a journal of the Radiological Society of North America (RSNA). The research highlights how AI aids radiologists in focusing more effectively on suspicious areas within screening mammograms, ultimately improving diagnostic accuracy.

Previous studies have demonstrated that AI decision support systems can increase the sensitivity of cancer detection without lengthening the time required for reading mammograms. However, the impact of AI on the visual search patterns of radiologists has remained largely unexplored until now.

Study Methodology and Findings

To delve deeper into this aspect, researchers employed an eye-tracking system to compare radiologist performance and visual search patterns with and without AI assistance. This innovative system utilized a small camera-based device, equipped with infrared lights and a central camera, to capture the precise coordinates of the radiologists’ eye movements on the screen.

Jessie J. J. Gommers, M.Sc., a joint first author of the study from the Department of Medical Imaging at Radboud University Medical Center in Nijmegen, Netherlands, explained, “By analyzing this data, we can determine which parts of the mammograms the radiologist focuses on, and for how long, providing valuable insights into their reading patterns.”

The study involved 12 radiologists who reviewed mammography examinations from 150 women, comprising 75 cases with breast cancer and 75 without. The results showed that breast cancer detection accuracy was higher when radiologists used AI support compared to unaided reading. Importantly, there was no significant difference in mean sensitivity, specificity, or reading time.

“The results are encouraging,” Gommers said. “With the availability of the AI information, the radiologists performed significantly better.”

AI’s Influence on Radiologists’ Focus

Eye-tracking data revealed a notable shift in radiologists’ focus when AI support was available. Radiologists spent more time examining areas containing actual lesions, suggesting that AI effectively directed their attention to regions of interest. “Radiologists seemed to adjust their reading behavior based on the AI’s level of suspicion,” Gommers noted. “When the AI gave a low score, it likely reassured radiologists, helping them move more quickly through clearly normal cases. Conversely, high AI scores prompted radiologists to take a second, more careful look, particularly in more challenging or subtle cases.”

The AI’s region markings acted as visual cues, guiding radiologists’ attention to potentially suspicious areas. In essence, the AI functioned as an additional set of eyes, enhancing both the accuracy and efficiency of interpretation.

“Overall, AI not only helped radiologists focus on the right cases but also directed their attention to the most relevant regions within those cases, suggesting a meaningful role for AI in improving both performance and efficiency in breast cancer screening,” Gommers said.

Balancing AI Assistance and Human Judgment

While the study underscores the potential benefits of AI in breast cancer screening, Gommers cautioned against overreliance on AI suggestions. Incorrect AI recommendations could lead to missed cancers or unnecessary recalls for additional imaging. However, multiple studies have found that AI can perform as well as radiologists in mammography interpretation, suggesting that the risk of erroneous AI information is relatively low.

To mitigate the risks of errors, Gommers emphasized the importance of ensuring that AI systems are highly accurate and that radiologists remain accountable for their decisions. “Educating radiologists on how to critically interpret the AI information is key,” she said.

Future Directions and Research

The researchers are conducting further studies to explore optimal scenarios for AI information availability, such as whether it should be provided immediately upon opening a case or only upon request. Additionally, they are developing methods to predict AI uncertainty in its decisions, which could enable more selective use of AI support, applying it only when it is likely to provide meaningful benefit.

“This would enable more selective use of AI support, applying it only when it is likely to provide meaningful benefit,” Gommers said.

The findings from this study, titled “Influence of AI Decision Support on Radiologists’ Performance and Visual Search in Screening Mammography,” offer promising insights into the integration of AI in medical imaging, potentially paving the way for more accurate and efficient breast cancer screening processes.