Comparison of coronal minimum intensity projection CT reconstructions with flexible bronchoscopy for airway compression in children with Lymphobronchial TB

Abstract

INTRODUCTION Tuberculosis (TB) remains one of the most important causes of morbidity and mortality in children in Africa, as well as the rest of the world. Lymphobronchial TB (LBTB) occurs when tuberculous adenopathy affects the airways, either by direct involvement (inflammation and erosion) or by indirect involvement (compression and stenosis). Endobronchial TB (EBTB) is the inflammation of the tracheobronchial tree, which is caused by tuberculosis, and is secondary to the rupture of lymph nodes into the bronchi, or the extension thereof to the peribronchial region, by lymphatic drainage. Identification of airway compromise due to any of these processes can be performed invasively using flexible bronchoscopy or non-invasively with CT scanning, including post processing techniques such as minimum intensity projections (MinIP). AIM This retrospective study aimed to generate standardised coronal minimum intensity projection (MinIP) CT reconstructions, and compare these with fibreoptic bronchoscopy in children with LBTB. METHOD Standardised coronal MinIP reconstructions were performed from CT scans in children with LBTB and the findings of three readers were compared with flexible bronchoscopy (FB), regarding airway abnormalities. Intraluminal lesions, the site of the stenosis, and the degree of stenosis were evaluated. The length of stenosis was evaluated by CT MinIP only, and no comparison to FB has been made. RESULTS 65 children with LBTB met the inclusion criteria (38 males; 58.5% and 27 females; 41.5%), with ages ranging from 2.5 to 144 months. Coronal CT MinIP demonstrated a sensitivity of 96% and specificity of 89% against FB. The most common site of stenosis was the bronchus intermedius (91%), followed by the left main bronchus (85%), the right upper lobe bronchus RUL (66%), and

the trachea (60%). Agreement between coronal CT MinIP and FB ranged from 36.9% at the carina to 84.6% at the RLL in normal and abnormal airways. CONCLUSION This study has proven that a standardised coronal CT MinIP reconstruction is useful in demonstrating airway stenosis in children with lymphobronchial TB, with sensitivity of up to 96% and specificity up to 89%. The most common sites of stenosis found by the coronal MinIP CT reconstruction were the BI (91%), followed by the LMB (85%), the RUL (66%), and the trachea (60%). The coronal CT MinIP had additional advantages over FB in that it allowed objective measurement of the diameter of the stenosis, measurement of the length of the stenosis as well as visualisation of the post-stenotic segments of the airways. CT MinIP was also able to provide information about lung parenchymal abnormalities. Standardised coronal MinIP reconstructions are easily performed, as described in our paper, and should be provided with each set of cross sectional MDCT images in children with LBTB. This one single image can provide easily appreciable and useful airway information and additional information not available from FB.