Eight of the Top 10 causes of death in Taiwan in 2020 are related to obesity. The National Health Research Institutes is working with a hospital and a corporation on a smart prediction system to analyze data for future trends with an emphasis on prevention.
This patient has a normal electrocardiogram, but AI analysis put his heart attack risk at 66 percent based on his genetic metabolites and clinical data. Further tests revealed he did indeed have a blocked coronary artery. This obesity smart prediction system was jointly developed by the National Health Research Institutes, Chang Gung Memorial Hospital, and Acer. As eight of the Top 10 causes of death in Taiwan are related to obesity, the system is aimed at using national health and clinical data to prevent disease.
Sytwu Huey-kang, Vice President, National Health Research Institutes: “We want to use genetic information to accurately predict the molecular pathogenic mechanisms of obesity-related diseases for accurate diagnoses and accurate predictions in the future.”
The NHRI says a polysomic disease database can monitor national health long-term and be used to develop a smart prediction system for the prevention of metabolic disorders. The team has found young people that wouldn't generally be considered at-risk with high risk.
Yeh Chi-hsiao, Cardiac Surgeon: “We identified many completely normal patients with coronary stenosis. They had completely normal electrocardiograms. We had one patient that underwent leg surgery last year. He had a completely normal EKG before his surgery. This year, he had a heart attack. We found that he had three blocked coronary arteries.”
Using AI and big data, the team has identified a set of biomarkers related to diabetes and renal disease. The system has an accuracy rate of at least 75% and will be used in the future for early diagnosis and early treatment.
專科醫師診斷，看起來無明顯異常的這張心電圖，但由AI分析病人基因體代謝體以及臨床疾病資料，判斷出罹患心肌梗塞機率有6成6，果真經過進一步檢查，冠狀動脈已經有阻塞狀況。這是國衛院、長庚醫院和宏碁三方聯手打造出的「肥胖智慧預測系統」，考量國人10大死因中8項與肥胖相關，希望透過建置國人健康資料庫，將醫療院所臨床收案導入產業大數據分析，達到預防醫學的效果。
國衛院副院長司徒惠康表示：「我們希望藉由對遺傳訊息流動的精準掌握，對肥胖這樣一個重要的代謝疾病，它的整個分子的致病機制以及後續，我們可以延伸的包括一個精準的診斷，精準的一個預測。」
國衛院指出，藉由建立疾病多體學資料庫，可以長期追蹤國人健康，進而開發與肥胖息息相關的「代謝疾病智慧預測系統」，降低多項代謝疾病對健康的影響。研究團隊在收案過程中，透過預測系統，也找出了原本不在高風險族群中的相對年輕病例。
心臟外科主治醫師葉集孝表示：「我們找到相當多完全正常的病人，他就是有冠狀動脈狹窄，心電圖檢查完全是正常的。我們曾經看過病人他是去年來開腳，那時候開刀前做心電圖是完全正常，今年就發現，原來他就是3條冠狀動脈阻塞，然後突然心肌梗塞就來醫院了。」
研究團隊利用AI大數據還發現一組生物標記與糖尿病腎病變有關，預測系統有7成5以上準確率，未來可望透過早期精準發現提早治療。
