Virtual endoscopy of the tracheo-bronchial system: sub-millimeter collimation with the 16-row multidetector scanner.

The purpose of this study was to evaluate the scope of sub-millimeter collimation reconstruction parameters using 16-row computer tomography and ECG triggering on image quality in virtual bronchoscopy. Thirty-two patients (5 women, 27 men, mean age 66.6+/-1.4) who had been admitted for coronary artery bypass graft surgery underwent CT examination of the thorax (Sensation 16, Siemens, Inc., Forchheim, Germany). All patients were examined with 16x0.75-mm collimation. Image reconstruction was performed for two groups. In group A ( n=32), slice thickness of 1.5 mm and an overlap of 0.75 mm were used. In group B ( n=32), slice thickness of 0.75 mm and an overlap of 0.4 mm were applied. Retrospective ECG triggering was performed in all patients. The maximum order of recognizable bronchi was determined in each data set. In addition to assessing the maximum order of bronchial bifurcation, bronchial diameter was determined in truly perpendicular sections in each patient. For every segment proximal to a bifurcation, image quality was subjectively graded as poor (grade 1), moderate (grade 2) or good (grade 3). The observers were asked to identify the minimum cardiac movement ECG-triggered image sets assuming that they would be of better quality than the maximum cardiac movement ECG-triggered image sets. The Mann-Whitney U-test and the Fisher's Exact Test were used for statistical evaluation. In group A, a mean of 4.8+/-0.2 bifurcations was ascertained vs. 6.5+/-0.3 bifurcations in group B [ P<0.0003]. For bronchial diameters in group A, a mean of 7.5+/-0.4 mm was determined vs. 4.6+/-0.4 mm in group B [ P<0.0001]. In group B, two independent radiologists observed a significant shift to better image quality in all segments evaluated [ P<0.006 to P<0.000001]. Motion artifacts were judged as being significantly reduced by minimum cardiac movement ECG-triggering in group B [observer 1: P=0.0007 (20/32); observer 2: P=0.008 (18/32)], but not in group A [observer 1: P=0.286 (13/32); observer 2: P=0.123 (16/32)]. Sub-millimeter collimation and minimum cardiac movement ECG-triggered data acquisition allow deeper penetration into the tracheo-bronchial system allowing visualization of the bronchial surface down to diameters below 5 mm in certain cases up to the eighth bifurcation. Along with an enhanced visualization as such, better image quality is acquired in all segments evaluated. Trade off between better image quality, of doubtful diagnostic consequence, and much higher irradiation dose must be made.