Clinicians can utilize chest X-rays to determine if a patient has tuberculosis, cancer, or other conditions, but they cannot assess lung function through these images alone. In a recent study, Japanese scientists developed a high-accuracy artificial intelligence (AI) model capable of evaluating lung function based on chest X-rays. This groundbreaking research was published in The Lancet Digital Health.
Typically, doctors assess lung function using spirometry, which relies on patient cooperation. This can be challenging for infants or individuals with dementia who may struggle to follow instructions. Additionally, patients with physical limitations may find it difficult to adopt standard testing positions, resulting in inaccurate measurements.
The research team from Osaka City University trained, validated, and tested their AI model using 140,000 chest X-rays collected over the past 20 years. They compared actual spirometry measurements with the AI's estimates to refine the model's accuracy. The results demonstrated a remarkable level of consistency, indicating the AI model's promising potential for real-world applications.
By providing an effective lung function assessment for those who struggle with traditional spirometry tests, this AI model could not only ease the burden on patients but also reduce healthcare costs significantly.