Current approaches to assessing the degree of airway narrowing in central airway obstruction.

Abstract

Accurate assessment of central airway obstruction (CAO) of various morphologic and mechanistic forms is an invaluable component of bronchoscopy practice. Reduction in airway lumen cross-sectional area (CSA), anatomic location, extent, morphologic features, and degree of dynamic narrowing affect decision-making and should be accurately described in bronchoscopy reports (1). Several techniques have been proposed for objective assessment of actual airway caliber or degree of airway narrowing (reported as percentage of normal). To date, however, no single “gold standard” technique is currently used in routine clinical practice. Current methods for objective quantification of CAO include several techniques that have various degrees of validity and reproducibility. Spirometry is not a reliable for quantifying the degree of CAO, and it does not identify the exact location, extent, or morphology of the airway narrowing. It also has low sensitivity for detecting mild to moderate reductions in airway caliber (2). Image-based quantification, such as multidetector computer tomography, with or without three-dimensional reconstruction, are dependent on respiratory cycle, anatomy, presence of secretions, and intraand interobserver variability. It may also be difficult to perform in critically ill patients because of poor cooperation, especially with end-inspiration breath-holding maneuvers (3). Transcutaneous acoustic analysis techniques have been shown to have clinical use in follow-up of laryngotracheal stenosis; however, this technique underestimates the absolute lumen size when used for measurement of transverse cervical tracheal diameter (4). Radial probe-balloon–based endobronchial ultrasound and optical coherence tomography have also been used, but they require bronchoscopy and additional technology that may not be readily available (5–7). Morphometric analysis of bronchoscopy images using cost-free image processing software has also been applied for assessing the percentage reduction in the airway CSA (8, 9). Indeed, from a flow dynamic standpoint, what matters most is the degree of narrowing based on percentage reduction in luminal CSA, and not the absolute airway diameter (10). Despite the availability of these objective techniques, practicing bronchoscopists continue to subjectively estimate the degree of airway narrowing. Subjective assessments, however, are inaccurate when they are based on visual estimation during bronchoscopy or simple still image analysis (9). Using the previously described stenosis and collapsibility indices is a practical and user-friendly alternative to measuring the absolute airway diameter (9). In fact, attempts to calculate the airway CSA are impractical for irregular airway morphologies, even for a known airway diameter. The study by Begnaud and colleagues in this month’s issue of AnnalsATS (pp. 85–90) analyzed current practices of quantifying CAO among members of American Association of Bronchology and Interventional Pulmonology (11). The study was performed by a survey-based method and hence suffers from potential recall and selection bias. One hundred eighteen survey responses were eventually analyzed. Each responder was asked to evaluate still bronchoscopic images of CAO and graded them on a numerical scale of 0 to 100. Baseline calculations of the stenosis index were performed using Fuji software analysis, and CSA area was calculated in pixels by software measurement tools in the region of polygon selection. Previously described bronchoscopic image acquisition techniques were not used, and hence accuracy of initial measurement may have been compromised and potentially inadvertently resulted in a false baseline reference standard (8). Indeed, distance image distortion invariably affects the estimation of degree of CAO. This has been previously demonstrated and thoroughly described in published reviews of morphometric bronchoscopy. The responses of numerical analysis of CAO were compared with previously calculated baseline. Statistical analysis of the means was calculated on the basis of lesion type and previous experience. Opinions and practices from all responders were solicited in free text form and tabulated. Bronchoscopists either underor overestimate the degree of stenosis when using still image analysis (9). Assessments of bronchoscopy videos may actually result in more accurate estimation of reduction in CSA, but this hypothesis has to be further studied. The study by Begnaud and colleagues supports the results from a previous study, in that there is no statistical correlation between the number of procedures performed and accurate subjective assessment of CAO (9). Most participants in this study agreed that an accurate and precise method should be used for quantification of CAO. Paradoxically, none of the participants surveyed reported use of quantitative