The AI-powered solution demonstrates 95.8% accuracy for defining severe obstructive sleep apnea.


Summary: Researchers at China Medical University Hospital’s Sleep Medicine Center in Taiwan are developing the iDREAM (Intelligent Detection of Respiratory Events through Automated Monitoring) home sleep test. Over 100 patients have participated in clinical trials, with AI-powered ECG analysis enhancing obstructive sleep apnea diagnosis accuracy and reducing diagnostic and treatment time. Dr Liang-Wen Hang highlights iDREAM’s potential to address limited sleep center capacity and personnel issues, offering comprehensive and convenient sleep apnea tracking. Clinical testing continues, with plans for certification and integration into clinical practice.


Researchers at China Medical University Hospital’s Sleep Medicine Center in Taiwan are developing the iDREAM (Intelligent Detection of Respiratory Events through Automated Monitoring) home sleep apnea test. 

More than 100 patients have been in clinical trials of iDREAM. With this artificial intelligence (AI)-powered ECG analysis, China Medical University Hospital’s physicians are able to identify obstructive sleep apnea more accurately and to reduce the time spent for diagnosis and treatment.

Dr Liang-Wen Hang, chief of China Medical University Hospital’s Sleep Medicine Center, says that the capacity of the Sleep Medicine Center is always limited while the need of the patients staying overnight are increasing. More, the Sleep Medicine Center has its own manpower issues, as sleep technicians are required to keep watch of recorded parameters and interpret data overnight. Thanks to iDREAM, the issues are resolved and create a more comprehensive and convenient tracking system for sleep apnea patients.

iDREAM detects ECG change from obstructive sleep apnea episodes and determines severity using its deep learning methodology. It demonstrates 92.7% and 93.2% accuracy for sleep apnea and wake-up events (interruption of sleep from apnea) and 95.8% accuracy for defining severe obstructive sleep apnea (30 episodes per hour). 

iDREAM is still undergoing clinical testing, but plans are in place to file for Taiwan Food and Drug Administration smart medical device certification and then formally introduce it into clinical practice. 

The AI Center at China Medical University Hospital trains artificial neural networks using the hospital’s medical data. Such models can provide suggestions in clinical diagnostics and may be combined with clinical decision-making systems to lessen the burden on healthcare personnel. China Medical University Hospital’s AI Center has already obtained more than a dozen certifications for intelligent medical materials from Taiwan Food and Drug Administration. 

Photo caption: China Medical University Hospital researchers show the iDREAM home sleep test.

Photo credit: China Medical University Hospital