An Operator Independent Method for Bronchial Tree Analysis from Trachea to the Small Airways Using Volumetric Multi-Detector Computed Tomography

Measurement of bronchial lumen and wall thickness of large and small airways from CT images is useful in clinical practice for the characterization of several diseases. Automatic methods are high desirable to save reporting time and to improve measurement reproducibility. In this study airways segmentation was performed by the application of a multi-step algorithm including airways segmentation, scheletonization and classification of the bronchial three, and calculation of the bronchial diameter and the airway wall thickness. The proposed method was validated on images from ten patients with suspect airways disease, comparing measurements obtained by the proposed algorithm with manual measurements performed by two expert users. Wall thickness and lumen diameter automatic measurements performed by the proposed approach were interchangeable with manual measurements, in the sense that the difference in measurements between the automatic method and the manual analysis is equivalent to the inter-observer variability. The proposed approach thanks to their simplicity showed computational efficiency, reliability, and was demonstrated to be totally operator-independent.