Based on an NHIA research project on information integration and application in 2021 conducted by Dr. Wei-Zi Liao's team of National Taiwan University Hospital, their thesis on AI deep learning for pancreatic cancer was published in well-known medical journal Radiology (JCR 2022 impact factor of 29.146) on September 13, 2022. The journal ranks first in the field of radiology. This research uses the pancreatic cancer detection tool developed by the team and CT imaging data collected by the NHIA from different hospitals to conduct evaluations, providing feasible suggestions for improving image interpretation.
The NHIA data integration application services started in 2015, allowing external academic and research institutions to apply for data for researches. In April 2022, medical images were added, with more than 3 billion images available. The diversity of the data for verification purposes has been confirmed by the research as valuable and recognized by academia. It can provide substantial contributions to the public in clinical applications. The NHIA will continue to provide diverse and practical information for medical research in the future.
Pancreatic cancer, also known as the “king of cancers,” is notorious for its difficulty to detect in the early stages and a high recurrence rate. Currently, computed tomography is the main diagnostic tool, but nearly 40% of tumors smaller than 2cm are not discovered. Therefore, in the clinical setting, most of confirmed patients have already progressed to the terminal stage and the survival rate is low. The clinical suggestions of the study do not completely rely on AI tools for detection. It still prioritizes the insights of professional radiologists, who are then assisted by AI tools to achieve early detection of pancreatic cancer and prevention before the fact.
Director General Lee stated that the NHIA continues to uphold a humble attitude to communicate with the public about the personal data. The NHIA is dedicated to completing the laws and regulations related to the secondary use of NHI data, thereby finding a balance between medical improvements and personal privacy. The NHIA hopes to inspire more diverse AI researches and implementation projects in the future, such as disease interpretation, health prediction, and other AI models. The high value and extensive medical database collected by the NHI has received medical and professional recognition. The NHIA hopes to use value-added data applications to create more socially beneficial results for all citizens and improve the health and welfare of the public.