The National Health Insurance (NHI) has been in operation for 25 years. During this time, it has accumulated a total of 64.9 billion medical records and 2.2 billion medical images since 2018, becoming a valuable resource for the development of big data analysis and medical artificial intelligence (AI) in Taiwan. In line with the smart government action plan, the NHI has opened its data to maximize value-added applications. Under the premise of compliance with the Personal Data Protection Act, it has allowed for de-identified medical image applications through industry-academia cooperation since June 2019. On the 27th, the “25th Anniversary of NHI: Health Insurance Data and Artificial Intelligence Applications Seminar” was held, with Vice President Lai Ching-te, Deputy Director-General of the National Development Council Shien-Quey Kao, Deputy Minister of the Ministry of Science and Technology Dar-Bin Shieh, and Quanta Computer Chairman Barry Lam delivering speeches. Attendees had the opportunity to witness AI achievements demonstrated on-site by hospitals, and the event was met with great enthusiasm. Participants from industry, government, and academia expressed their admiration and increased confidence in Taiwan’s development of AI.

As Taiwan’s NHI continues its 25-year journey, the application of AI and big data is continuously improving healthcare quality. The NHI has been actively promoting and introducing various healthcare technologies. Over two years ago, when Lai Ching-te was serving as the Premier, he was invited to attend a press conference for the “Information Upload to the Cloud and Medical Data Sharing” initiative. He highly praised the NHI’s breakthrough in enabling hospitals to upload and share imaging data and reports in real-time. Today, in his capacity as Vice President, Lai emphasized that the era of smart healthcare is an inevitable global trend. Taiwan’s large and internationally renowned health insurance database, combined with the technological capabilities of the information and communication industry, has the potential to establish Taiwan as a leader in smart healthcare, aligning with President Tsai Ing-wen’s vision for the development of six core industries and enhancing Taiwan’s competitiveness in the smart healthcare industry.

As the wave of artificial intelligence (AI) sweeps globally, the NHI has been involved in the National Development Council’s “ Asia Silicon Valley Development Agency ” since 2018, focusing on the development of medical materials, document building, and AI applications. The NHI’s AI platform is now rapidly integrating data and de-identifying it for use in AI-powered diagnostic systems. In June 2019, the NHI began a pilot program that allows de-identified CT and MRI images to be used for AI algorithm training and model development. A total of 15 projects were submitted, involving 9 medical institutions and 6 academic institutions, with 10 cases featuring industry-academia collaboration.

At today’s seminar, twelve teams presented their AI research findings, each with unique highlights. Here is a brief overview of their projects:

Taipei Veterans General Hospital–AI-assisted diagnostic system for brain metastasis: The DeepMets®-Plus system was developed to accurately detect brain metastases, providing important information such as the number, maximum diameter, and volume of metastases. This system was integrated into the hospital’s neuroimaging and pulmonary departments to aid in diagnosing brain metastasis from lung cancer, improving medical workflow efficiency and quality.

National Taiwan University– BodyPart Medical Imaging Body Part AI and Cross-National Federated Learning by COVID-19 Models: The “BodyPart Intelligent Search System” has been developed to identify corresponding human body parts in medical images, enabling rapid recognition of organs in the head, neck, chest, and abdomen based on annotated images. In collaboration by worldly known corporations NVIDIA, the research utilizes AI models to predict the oxygen supplementation needs of COVID-19 patients .

National Taiwan University Hospital– AI-Assisted Pancreatic Diagnosis Tool: PANCREASaver: The world’s first AI automatic collection detection model for pancreatic cancer by computed tomography scan（CT scan） image acquisition has been developed. Validated on a large-scale national database, it achieves an accuracy of up to 91.1%. Integrated into the Picture Archiving and Communication System (PACS), it assists in the rapid diagnosis of pancreatic cancer.

National Taiwan University Hospital – Intelligent Brain Metastasis Tumor Auto-Contour and Evaluation System Analysis: A highly precise stereotactic radiosurgery (SRS) planning intelligent assistant has been developed, featuring an automated segmentation system for brain metastases and surrounding critical organs. It provides tumor segmentation results for evaluation, monitors brain MRI scans, and delivers real-time final result rates. Additionally, it offers distance parameters between tumors and adjacent critical organs, assisting physicians in surgical planning and decision-making.

National Taiwan University – Developing a Cardiovascular Disease Risk Prediction Model Using Nationwide Health Insurance CT Imaging Big Data: The AI model HeaortaNet® has been developed for coronary aortic calcification and fat vascular analysis. It is the world’s only AI model capable of automatically classifying and quantifying thoracic calcification and fat deposits. With a cardiac reconstruction accuracy of 94.2%, it significantly enhances the added value of CT imaging, advancing precision medicine.

Taipei Medical University Hospital – Validation of an AI-Assisted Pulmonary Diagnosis System Using National Health Insurance Big Data: The Deep-Lung: LungRads AI-assisted diagnostic system for lung nodules has been developed, integrating deep learning and radiomics algorithms. It can identify lung nodule locations within 20 seconds and assist in determining the benign status of lung nodules.

Far Eastern Memorial Hospital– Development of an AI Model for Automated Nasopharyngeal Cancer (T Stage) Staging: Nasopharyngeal cancer (NPC) is particularly prevalent in Southeast Asia, but its interpretation remains challenging due to the complex anatomy of the head and neck.

AI-Powered Pandemic Prevention for the Future

The National Health Insurance Administration (NHIA) stated that to enhance the development capacity of medical AI in Taiwan, it signed a collaboration agreement in March 2020 with Quanta Computer Chairman Barry Lam. Quanta Computer has integrated NHIA’s “AI Medical Cloud Computing Platform (QOCA® AIM)” system and customized it to create an environment that facilitates the added value of National Health Insurance data.

Additionally, the NHIA collaborated with Professor Jung-Hsien Chiang’s team at National Cheng Kung University to develop an AI model for COVID-19, which is currently undergoing a special medical device production application with the Taiwan FDA.

In June and October of this year, the NHIA revised the “Operational Guidelines for the Provision of Data by the National Health Insurance Administration” and the “Operational Guidelines for the Integrated Information Application Service Center of the National Health Insurance Administration.” After de-identification of personal data, the industry is now allowed to apply for access and utilization.

Director-General Po-Chang Lee of the National Health Insurance Administration (NHIA) stated that the NHIA has accumulated a high-value de-identified big data medical database. Today’s seminar marks the initial framework for opening up the NHIA database for application, fostering collaboration with the industry to leverage the synergy between the public sector, industry, and academia. This initiative aims to further drive AI application advancements.

Looking ahead, the NHIA will continue to uphold an open approach, encouraging teams willing to undertake projects to engage in value-added applications of the database, provided they comply with all relevant laws and regulations. This will contribute to national development, enhance public health and well-being, and ensure that the benefits are shared with all citizens.