As Taiwan sets ambitious carbon reduction goals—targeting a 32% decrease by 2032—Kaohsiung Medical University (KMU) is spearheading a transformative initiative to integrate artificial intelligence (AI) into medical practice while promoting environmental sustainability. In collaboration with MacroInsight Innovation Ltd., KMU has leveraged edge AI and the latest Apple M4 chip to develop three groundbreaking AI-powered healthcare models: the Mammography Image Quality Assessment System, the Endotracheal Tube Abnormality Warning System, and the Cardiogenic Sudden Death Risk Prediction System. These innovations aim to establish a new paradigm for low-carbon healthcare by improving diagnostic accuracy, reducing resource consumption, and enhancing patient outcomes.
KMU Vice Superintendent Lin Tsung-Hsien emphasized that sudden cardiac death is a leading cause of mortality, often stemming from undiagnosed heart conditions. Traditionally, physicians rely on electrocardiograms (ECGs) to assess risk, but subjective interpretation can lead to variability. The Cardiogenic Sudden Death Risk Prediction System employs real-time ECG analysis to provide objective cardiovascular risk assessments. Its lightweight design makes it ideal for use in primary care and rural areas, reducing reliance on specialized equipment and operational costs. By enabling long-term data tracking, this AI model embodies the principle that prevention is better than cure.
According to KMU Secretary Tsai Ming-Ju, overdiagnosis and unnecessary medical procedures have been long-standing challenges in healthcare. Mammography, a critical tool for early breast cancer detection, requires at least four high-quality images for accurate analysis. The Mammography Image Quality Assessment System provides real-time feedback to radiographers, ensuring standardized image quality and improving breast cancer detection rates. This technology enhances diagnostic accuracy while streamlining the patient experience. In the future, the system could be integrated into mobile mammography units, extending healthcare accessibility to remote and indigenous communities.
ICU environments present unique challenges due to the complexity of critically ill patients. KMU Pulmonologist Dr. Cheng Chih-Hung introduced the Endotracheal Tube Abnormality Warning System, which employs AI-powered image recognition to verify tube placement and predict extubation success rates. This system enables healthcare professionals to promptly address potential risks, significantly reducing the incidence of tube dislodgement. With an AI prediction accuracy rate of 96%, this innovation lowers the risk of complications and mortality while alleviating the workload of medical personnel.
MacroInsight Innovation Ltd. CEO Tu Ming-Da highlighted that these AI models demonstrate Taiwan’s leading role in AI-driven healthcare innovation. The company aims to extend the adoption of these technologies across medical institutions and rural regions, ensuring equitable access to smart healthcare solutions. Through continued advancements, AI is poised to redefine modern medical practice by enhancing efficiency, accuracy, and sustainability.
Resource: 高醫三大醫療AI模型 打造「低碳醫療」
