Cardiff University Scientists Develop AI System to Enhance Breast Cancer Detection
Scientists at Cardiff University have made a breakthrough in breast cancer detection with the development of an artificial intelligence (AI) system that mimics the gaze of radiologists when analyzing medical images like mammograms. This revolutionary system, the first of its kind, improves the speed, accuracy, and sensitivity of medical diagnostics, potentially leading to early detection of breast cancer.
The team, from Cardiff University’s School of Computer Science and Informatics, designed the AI system using a sophisticated algorithm called a convolutional neural network. This algorithm, inspired by the human visual cortex, is adept at assigning importance to different objects or aspects within an image.
To create the AI system, the researchers collaborated with radiologists from three National Health Service (NHS) hospitals in the UK: Breast Test Wales, University Hospital of Wales (UHW), and Great Ormond Street Hospital. They worked together to develop an algorithm that accurately predicts the areas within an image where radiologists are most likely to focus their attention when making a diagnosis.
Dr. Hantao Liu, a Reader at Cardiff University’s School of Computer Science and Informatics, emphasized that the goal of this research is not to replace humans with robots but to demonstrate how AI and machine learning can support and enhance the work of medical professionals. The developed AI system acts as a critical friend or colleague, assisting radiologists in their decision-making processes during medical diagnostics.
The shortage of radiologists in the UK is a significant concern, and the Cardiff University team hopes that their AI system can help address this issue. By training and educating radiologists through this technology, it is possible to alleviate the burden on healthcare professionals and expedite the diagnostic process.
Dr. Richard White, a Consultant Radiologist at UHW who participated in the study, highlighted the advantages of incorporating AI into radiology practice. He acknowledged that even the most experienced radiologists can miss subtle details in medical images, and the AI system can help spot abnormalities that might otherwise go unnoticed. Consequently, the technology enhances the precision and accuracy of radiologists’ work, reducing the risk of errors.
Moreover, the AI system can aid in prioritizing patient referrals. By identifying abnormal scans and images that require immediate attention, healthcare professionals can streamline the reporting process, ensuring timely diagnoses for patients in need.
While the focus of this study was on gaze prediction, there is potential for further advancements in AI systems that aid decision-making in clinical applications. The researchers from Cardiff University’s Multimedia Computing Research Group foresee a future where AI helps revolutionize deep learning and its role in medical diagnostics.
Dr. Zelei Yang, another Radiologist at UHW involved in the study, emphasized that this AI system brings a greater humanization to AI technology, making it a more realistic representation of the work radiologists perform. As a result, the system can effectively support and enhance the process of deep learning in clinical settings.
With its potential to enhance breast cancer detection and support healthcare professionals, the AI system developed by Cardiff University scientists marks a significant step forward in the field of medical diagnostics. By embracing data science and AI, it becomes possible to address the challenges faced by the healthcare system and maximize the benefits of technological advancements in patient care.
