Professors Kim Jung-hoon

Seoul National University Bundang Hospital (SNUBH) said Thursday its researchers have discovered ways to diagnose sleep apnea accurately by analyzing a patient's breathing sound while sleeping.

Obstructive sleep apnea refers to a disorder, in which airflow to the respiratory tract stops during sleep and causes severe snoring and temporarily breathlessness. Because it is a common symptom, it does not initially seem to be dangerous. However, if left untreated for a long time, it can lead to cerebrovascular disease and possibly death, making an early diagnosis of the disease very important.

Despite the importance of an early diagnosis it is hard to obtain an accurate result as it requires a nighttime sleep polygraphy. However, the test needs the patient to stay at a hospital overnight resulting in high costs.

Also, given that sleep apnea is closely related to a patient’s lifestyles, such as diet, exercise, drinking and smoking, it is essential to monitor sleep apnea on a regular basis, while continually correcting the wrong habits. However, as there are no pre-screening tests that meet such criteria, it is hard for patients to control their lifestyle systematically.

The research team, led by Professors Kim Jung-hoon at the hospital and Lee Kyo-gu at Seoul National University, decided to focus on finding a method that was both easier and cheaper for the patient.

The team focused on the fact that patients with sleep apnea tended to snore more than other people do. Patients’ breathing sound were also more rough and irregular. The researchers began to develop an algorithm after collecting the data from 120 patients, diagnosed with sleep apnea at SNUBH and had undergone a nighttime sleep polygraphy from November 2015.

The algorithm allowed the team to obtain various data such as respiration interval changes and the changes in the volume of the respiratory sound.

The team also applied mechanic and deep learning techniques to improve the accuracy of diagnostic algorithms, and used it to analyze the sleep stages of each patient to predict the severity of a patient’s sleep apnea accurately.

The developed algorithm diagnosed 88.3 percent of patients in the fourth stage in severity classification and 92.5 percent in the second stage.

“As a result of this study, medical professionals can now determine sleep apnea through sound,” Professor Kim said. “If the pre-screening diagnostic algorithm for sleep apnea is introduced into the actual medical field, it will increase convenience for the patient as it can diagnose the disease only by breathing sound recordings.”

Biomedical Engineering has published the results of the study.

corea022@docdocdoc.co.kr

<© Korea Biomedical Review, All rights reserved.>