Estimation of cardiopulmonary exercise testing (CPX) values by machine learning
[Japanese|English]
Abstract
This research aims to estimate CPX values without exercise load using machine learning.
- Continued exercise at an appropriate intensity is important for preventing cardiovascular disease and preventing relapse or re-hospitalization after treatment.
- To prescribe and measure the effects of exercise, doctors now use ergometers or treadmills to stress people to their limits (⇒ CPX).
- In this study, test values are estimated by machine learning without exercise load.
*This is not a medical device under the Pharmaceuticals and Medical Devices Act.
Existing method: CPX
Our method: Estimation of CPX values using machine learning
Contributions
- Our method estimates CPX values required for exercise prescription without applying exercise load.
- Through joint research with a specialized hospital with the largest number of cardiac rehabilitation cases in Japan, we will build an estimation model based on a wide variety of cases.
- It is possible to estimate various CPX values such as heart rate at AT and load 1 minute before AT.
⇒ If exercise prescription becomes available not only at large medical institutions with well-equipped facilities and systems, but also at clinics, it is expected to lead to the spread of cardiac rehabilitation.
Contact
Hiroki Sakuma
Biomedical Informatics Research Group, Media Information Laboratory, NTT Communication Science Laboratories