Virtual laryngoscopy and multiplanar reformats with multirow detector CT for detection and grading of upper airway stenosis

Abstract

A controlled trial was performed to compare non-invasive multislice CT (virtual laryngoscopy, axial CT slices, coronal and sagittal reformats) in the detection and grading of upper airway stenosis with fiberoptic laryngoscopy. Multislice CT and fiberoptic laryngoscopy were used to examine 111 upper airway sections (supraglottis, glottis, subglottis, trachea) in 29 patients. CT data were acquired on a multirow detector CT (collimation 4 x 1 mm, reconstruction interval 1 mm, IV contrast) and postprocessing was performed using multiplanar reformatted images (MPR) and virtual laryngoscopy. All CT methods accurately detected upper airway stenosis (accuracy was 96 % for virtual laryngoscopy and MPR and 94 % for axial CT-slices). Correlation of fiberoptic and virtual laryngoscopy (r = 0.94) for grading of stenosis was closer than with sagittal reformats (r = 0.80), coronal reformats (r = 0.72), and axial CT slices (r = 0.57). Even high grade stenosis could be passed with virtual laryngoscopy that was impassable for fiberoptic laryngoscopy. Virtual laryngoscopy enabled better assessment of stenosis as compared to reading of axial CT slices or MPR. Virtual laryngoscopy is complementary to fiberoptic laryngoscopy and should be combined with axial CT slices and MPR readings for evaluation of the surrounding structures.