Estimating a Physiological Lung Function Score and Biological Sex Using Pulmonary Function Tests and Machine Learning: Retrospective Study
JMIR AI · 01.06.2026 · 5 · e89060-e89060
DOI: 10.2196/89060 ↗Kurz gefasst
Fachartikel in der Zeitschrift JMIR AI, online erschienen am 01.06.2026. Verfasst von Patrick W Johnson und 7 weiteren Autor:innen. Beteiligte Einrichtung: Department of Quantitative Health Sciences, Mayo Clinic Hospital, 4500 San Pablo Road S, Jacksonville, FL, 32224, United States, 1 9049536728. Die Arbeit nennt 45 Literaturangaben. Der Volltext ist frei in PubMed Central verfügbar.
Abstract (Originalsprache)
Background
Sex and age have long been known to affect lung function. Several biological variables and anatomical factors may contribute to sex- and age-related differences in pulmonary metrics.
Objective
We hypothesized that a machine learning model could be trained to predict a person’s lung age and self-reported sex using pulmonary function test data.
Methods
We retrospectively analyzed complete pulmonary function tests from 6392 healthy adults across 3 Mayo Clinic regions. Four models of increasing complexity were trained using gradient-boosted machines to predict chronological age and biological sex. Model interpretability was assessed using Shapley additive explanation values and partial dependence plots. Quantile regression was used to estimate reference percentiles for predicted lung age.
Results
The best-performing age model (model 4, inclusive of time-series features) achieved a root mean square error of 7.01 years (95% CI 6.73‐7.30) and a mean absolute error of 5.55 years (95% CI 5.32‐5.80). The best-performing sex classification model (model 4) achieved an area under the curve of 0.981 (95% CI 0.975‐0.988), sensitivity of 91.7% (95% CI 89.0%‐93.9%), and specificity of 95.6% (95% CI 93.9%‐97%). Key predictors for lung age included residual volume as a percentage of total lung capacity (TLC), forced expiratory volume in 1 second, and alveolar volume. For sex classification, peak expiratory flow, height, and age were among the most influential features. Age-stratified evaluation showed the overestimation of lung age in younger adults and underestimation in older adults. Predicted lung age increased broadly with chronological age, and quantile regression provided normative reference ranges.
Conclusions
Applying artificial intelligence to pulmonary function data allows the prediction of a patient’s sex and estimation of lung age. The ability of an artificial intelligence algorithm to determine physiological lung age, with further validation, may serve as a measure of overall respiratory health.
Schlagwörter
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Department of Quantitative Health Sciences, Mayo Clinic Hospital, 4500 San Pablo Road S, Jacksonville, FL, 32224, United States, 1 9049536728.
- Department of Medicine, Division of Pulmonary Medicine, Mayo Clinic Hospital, Rochester, MN, United States.
- Department of Medicine, Division of Pulmonary Medicine, Mayo Clinic Hospital, Phoenix, AZ, United States.
Zitieren
APA
Johnson, P. W., Quicksall, Z. S., Lee, J., Lee, A. S., Lim, K. G., Ortega, V. E., Arunachalam, S. P., & Helgeson, S. A. (2026). Estimating a Physiological Lung Function Score and Biological Sex Using Pulmonary Function Tests and Machine Learning: Retrospective Study. JMIR AI, 5, e89060-e89060. https://doi.org/10.2196/89060
Vancouver
Johnson PW, Quicksall ZS, Lee J, Lee AS, Lim KG, Ortega VE, et al. Estimating a Physiological Lung Function Score and Biological Sex Using Pulmonary Function Tests and Machine Learning: Retrospective Study. JMIR AI. 2026;5:e89060-e89060. doi:10.2196/89060
BibTeX
@article{Johnson2026,
author = {Johnson, Patrick W and Quicksall, Zachary S and Lee, Jieun and Lee, Augustine S and Lim, Kaiser G and Ortega, Victor E and Arunachalam, Shivaram Poigai and Helgeson, Scott A},
title = {Estimating a Physiological Lung Function Score and Biological Sex Using Pulmonary Function Tests and Machine Learning: Retrospective Study},
journal = {JMIR AI},
year = {2026},
volume = {5},
pages = {e89060--e89060},
doi = {10.2196/89060}
}
EUVIMED