A recent study published in JMIR AI suggests that machine learning can accurately determine a patient's biological sex and estimate their physiological lung age from pulmonary function tests (PFTs). This lung age prediction approach provides a more intuitive understanding of respiratory health compared to traditional metrics. Consequently, clinicians may soon use these artificial intelligence models to enhance patient communication and monitoring.

Researchers from the Mayo Clinic analyzed retrospective data from 6,392 healthy adults across three regions. They developed gradient-boosted machine models to analyze various PFT parameters, including time-series features. The goal was to determine if artificial intelligence could go beyond standard interpretation to reveal underlying biological patterns related to aging and sex.

Accuracy of Lung Age Prediction Models

The most advanced model achieved a mean absolute error of only 5.55 years in predicting chronological age. Moreover, the sex classification model demonstrated exceptional performance with an area under the curve (AUC) of 0.981. Key predictors for these models included peak expiratory flow, height, and residual volume as a percentage of total lung capacity. Therefore, these findings suggest that AI can extract highly specific physiological markers from routine lung tests.

Although the models tended to overestimate age in younger individuals and underestimate it in older adults, they consistently followed chronological trends. This suggests that physiological lung age could become a valuable biomarker for monitoring lung aging and overall wellness. Furthermore, the ability to predict biological sex with 91.7% sensitivity highlights the profound anatomical differences captured by PFT data.

In conclusion, the integration of AI into pulmonary diagnostics offers a promising pathway for personalized medicine. However, further validation in diverse patient populations is necessary before widespread clinical adoption. Ultimately, this technology could revolutionize how we interpret lung function and manage chronic respiratory conditions.

Frequently Asked Questions

How does AI estimate lung age?

AI models analyze complex patterns within PFT data, such as airflow volume and capacity. They compare these metrics against large datasets to determine how closely a person's lung function matches specific age groups, resulting in a lung age prediction.

What are the benefits of using lung age in clinical practice?

Lung age offers an easy-to-understand metric for patients. This can potentially increase motivation for smoking cessation or adherence to respiratory treatments by making the impact of lung health more tangible.

Can these models replace traditional PFT interpretation?

Currently, these AI models are designed to supplement traditional interpretation. They provide additional context regarding physiological aging rather than replacing standard clinical diagnosis.

Disclaimer: This content is for informational and educational purposes only. It is not intended as a substitute for professional medical advice, diagnosis, or treatment. Always seek the advice of your physician or other qualified health provider with any questions you may have regarding a medical condition. Refer to the latest local and national guidelines for clinical practice.

References

Johnson PW et al. Estimating a Physiological Lung Function Score and Biological Sex Using Pulmonary Function Tests and Machine Learning: Retrospective Study. JMIR AI. 2026 Jun 01. doi: 10.2196/89060. PMID: 42224015.

Topalovic M, et al. Artificial intelligence for interpreting pulmonary function tests. Eur Respir J. 2019;53(3):1801703. doi: 10.1183/13993003.01703-2018.

American Thoracic Society. Standardizing PFT Interpretation. 2022.