Combining small airway parameters with conventional parameters obtained during spirometry to diagnose airflow obstruction: A cross-sectional study

BACKGROUND AND OBJECTIVE

The use of small airway parameters generated by spirometry, namely forced expiratory flow between 25% and 75% of forced vital capacity (FVC) (FEF25%-75%) and forced expiratory flow at 50% and 75% of FVC (FEF50% and FEF75%, respectively), is widely discussed. We evaluated the importance of these spirometric parameters in a large Chinese population.

METHODS

We conducted a cross-sectional observational study in which spirometry and bronchodilator responsiveness (BDR) data were collected in a healthcare centre from May 2021 to August 2022 and in a tertiary hospital from January 2017 to March 2022. Discordance was assessed between the classification of test results by the large airway parameters of forced expiratory volume in 1 second (FEV1) and FEV1/FVC ratio and by the small airway parameters of FEF25%-75%, FEF75% and FEF50%. The predictive power of Z-scores of spirometric parameters for airflow limitation and BDR was assessed using receiver operating characteristic curves.

RESULTS

Our study included 26,658 people. Among people with a normal FVC (n = 14,688), 3.7%, 4.5% and 3.6% of cases exhibited normal FEV1/FVC ratio but impaired FEF25%-75%, FEF75% and FEF50%, respectively, while 6.8%-7.0% of people exhibited normal FEV1 but impaired FEF25%-75%, FEF75% and FEF50%. Using the Z-scores of combining both large and small airway parameters in spirometry showed the best area under the curve for predicting airflow limitation (0.90; 95% CI 0.87-0.94) and predicting BDR (0.72; 95% CI 0.71-0.73).

CONCLUSION

It is important to consider both large and small airway parameters in spirometry to avoid missing a diagnosis of airflow obstruction.

Supervised training models with or without manual lesion delineation outperform clinicians in distinguishing pulmonary cryptococcosis from lung adenocarcinoma on chest CT

BACKGROUND

The role of artificial intelligence (AI) in the discrimination between pulmonary cryptococcosis (PC) and lung adenocarcinoma (LA) warrants further research.

OBJECTIVES

To compare the performances of AI models with clinicians in distinguishing PC from LA on chest CT.

METHODS

Patients diagnosed with confirmed PC or LA were retrospectively recruited from three tertiary hospitals in Guangzhou. A deep learning framework was employed to develop two models: an undelineated supervised training (UST) model utilising original CT images, and a delineated supervised training (DST) model utilising CT images with manual lesion annotations provided by physicians. A subset of 20 cases was randomly selected from the entire dataset and reviewed by clinicians through a network questionnaire. The sensitivity, specificity and accuracy of the models and the clinicians were calculated.

RESULTS

A total of 395 PC cases and 249 LA cases were included in the final analysis. The internal validation results for the UST model showed a sensitivity of 85.3%, specificity of 81.0%, accuracy of 83.6% and an area under the curve (AUC) of 0.93. Similarly, the DST model exhibited a sensitivity of 88.2%, specificity of 88.1%, accuracy of 88.2% and an AUC of 0.94. The external validation of the two models yielded AUC values of 0.74 and 0.77, respectively. The average sensitivity, specificity and accuracy of 102 clinicians were determined to be 63.1%, 53.7% and 59.3%, respectively.

CONCLUSIONS

Both models outperformed the clinicians in distinguishing between PC and LA on chest CT, with the UST model exhibiting comparable performance to the DST model.