[clinical research] the field of traditional lung sound diagnosis is facing a new future
1. What is the current routine auscultation of lung sounds by clinicians?
Lung sound is the change of air flow movement due to the structure of human respiratory tract during breathing, forming vortex or turbulence and generating sound waves. The sound wave propagates to the surface of the chest wall in the form of solid vibration through the lung tissue and chest wall, and is collected by the auscultation doctor, which is the lung sound. Lung sound consists of three parts: time domain (time process of lung sound change), frequency domain (distribution of frequency components) and amplitude (amplitude change). Clinicians divide abnormal lung sound into different types of pathological sound by combining the actual symptoms through the change combination of three dimensions.
So, where does lung sound auscultation come from? Lung sound, also known as sound absorption, is a natural physiological phenomenon discovered since 1500 BC. However, it was not until the gradual rise of modern medicine in the 19th century that lung sound auscultation attracted people's attention as an important way to diagnose respiratory diseases. In 1761, the German doctor Leopold auenbrugger put his ear on the patient's chest to listen to the patient's respiratory sound and identify the patient's lung pathological state. However, this method is not widely accepted because it is inconvenient in the examination process of diagnosing heterosexual patients. In 1816, the famous French doctor Ren ¨¦ - th ¨¦ ophile Hyacinthe Laennec invented the wooden stethoscope, and lung sound auscultation was gradually popularized in clinic.
2. How is lung sound signal detected?
With the continuous development of acoustics and computer technology in recent years, especially the widespread use of artificial intelligence and machine learning, the detection methods of lung sound signals are constantly updated. After the lung sound samples are collected, they are amplified and filtered. After eliminating the interference of air flow, volume, residual air volume, body position, secretion, heart sound and muscle contraction sound, they are sampled according to a certain frequency to analyze the time domain, frequency domain and amplitude of audio, so as to obtain relatively objective judgment results.
3. What breakthroughs can be made in the clinical research of lung sound collection?
At present, due to the interference of the above factors, there are still some problems in traditional lung sound auscultation, such as environmental noise and the need for clinicians' experience. It is difficult to realize the scientific concept of early detection, early diagnosis and early treatment advocated at this stage. With the rise of AI diagnosis of clinical diseases in recent years, there are more and more new detection and diagnosis methods of lung sound. In order to implement the relevant national policies on the application of medical and health technology and further promote the clinical application of lung sound AI diagnosis scientific research achievements, Guangdong society of clinical medicine - clinical medicine research center plans to carry out the scientific research project of "lung sound pathological phonetic book collection". The technology adopted in the project is to combine neural network model with frequency band analysis technology, The ds88b electronic stethoscope of Hanhong medical technology is used to collect and realize AI diagnosis of lung sound. The purpose is to verify the effectiveness of in vivo sound acquisition and AI diagnosis from the perspective of clinical application.
Author: Xia Jichao
Reviewed by: Liu Yanlu
Source: Guangdong society of clinical medicine
This article is reprinted from Guangdong society of clinical medicine. If you have any copyright objection to the content and pictures, please contact us in time for deletion (mobile phone and wechat: 18923210789)