A wearable device for clinical diagnosis of lung diseases with integrated artificial intelligence algorithms

Lung disease is one of the most widely spread diseases that endangers the lives and health of the Chinese people. The main ones include pneumonia, emphysema, interstitial lung disease, chronic obstructive pulmonary disease, etc.

Different lung diseases have different treatments and prognoses, so early diagnosis is critical. Currently, passive diagnosis of lung diseases mainly relies on mature medical instruments in hospitals, while wearable clinical diagnostic technologies that are expected to achieve active health are relatively scarce.

Recently, Professor Liu Hong and Associate Professor Zhao Chao of Southeast University led a research team to develop a wearable device for clinical diagnosis of lung diseases that integrates artificial intelligence algorithms.

Figure 1 Wearable devices that can be used for clinical diagnosis of lung diseases: (a) schematic diagram; (b) photo; (c) actually worn by the patient.

This work achieved high signal-to-noise ratio acquisition of lung sounds in clinical environments through the design and preparation of wearable flexible circuits; by collecting lung sound signals from multiple patients with different lung diseases and performing preprocessing/feature extraction/machine learning algorithm optimization on the signals, it ultimately achieved a lung disease recognition accuracy rate of more than 95%.

Figure 2 (a) Performance of the four optimal algorithm models; (b) Confusion matrix results of the ETC model

Aiming at people's life and health, this device can provide key core technology for the early diagnosis of lung diseases. The achievement was published in the journal Sensors & Diagnostics of the Royal Society of Chemistry under the title "Wearable stethoscope for lung disease diagnosis".