National Taiwan University Hospital has recently launched an innovative AI-powered diagnostic imaging service aimed at detecting pancreatic cancer. This groundbreaking service, known as PANCREASaver, is available as a self-pay option at the NTUH Department of Medical Imaging and is considered the first of its kind globally.
PANCREASaver was developed in collaboration with the Institute of Applied Mathematical Sciences at National Taiwan University. The system utilizes deep learning algorithms and a comprehensive multi-image training dataset to automatically analyze CT scans, identify the pancreas, and highlight suspicious lesions. It seamlessly integrates with NTUH’s Picture Archiving and Communication System (PACS) and is accessible across various departments, including gastroenterology, surgery, and oncology.
In rigorous national clinical trials, PANCREASaver demonstrated an impressive 80% sensitivity in detecting early-stage pancreatic tumors measuring less than 2 centimeters, with an overall diagnostic accuracy exceeding 90%. This cutting-edge technology has received regulatory approval from the Taiwan Food and Drug Administration and has been designated as a breakthrough device by the United States Food and Drug Administration.
Pancreatic cancer poses significant challenges in terms of diagnosis and treatment, with Taiwan experiencing a gradual increase in mortality rates and a five-year survival rate of approximately 10%. Early-stage pancreatic tumors, particularly those under 2 cm that are undetectable to the naked eye, are notoriously difficult to detect, leading to a significant number of missed lesions.
The research team behind PANCREASaver developed this AI system to address the diagnostic challenges associated with pancreatic cancer. They are currently exploring opportunities to expand the system’s capabilities to other pancreatic-related conditions, such as pancreatitis and pancreatic cystic lesions. By incorporating additional multi-modal imaging and clinical data, the team aims to enhance the diagnostic system’s coverage and accuracy.
In a broader context, Fujitsu and the Southern Tohoku General Hospital in Japan have also developed an AI-driven solution for early pancreatic cancer detection. Leveraging Fujitsu’s AI technology trained with anonymized pancreas CT images, they are working towards improving diagnostic capabilities in this critical area.
Furthermore, NTUH is actively engaged in large language model development, having recently acquired two new supercomputers to facilitate this initiative. By incorporating a diverse range of data types into their models, NTUH aims to optimize operations and enhance patient services through the development of multimodal large language models.
Overall, the introduction of PANCREASaver represents a significant advancement in the field of pancreatic cancer diagnosis and underscores the potential of AI-powered technologies to revolutionize healthcare outcomes. As NTUH continues to innovate and expand the capabilities of their diagnostic systems, patients stand to benefit from more accurate and timely detection of life-threatening conditions.
