Comparative Evaluation of Super High-Resolution CT Scan and Virtual Bronchoscopy for the Detection of Tracheobronchial Malignancies: Results
Image acquisition and anatomic simulation of the tracheobronchial tree were successful in all patients, No adverse events were noted in the study group, There were no respiratory or cardiac motion artifacts that precluded image reconstruction,
SHR-CT and VB
Of the 32 patients who underwent correlative FBs, 9 (28%) had normal examination findings. In all patients with normal anatomy, the results of SHR-CT scans and VB accurately correlated with those of FBs (Web movie 1).
In the remaining 23 patients, a total of 35 abnormal FB findings were detected (Figs 1 and 2). With respect to the location of FB findings, there were 20 lesions (57%) in the right lung, 12 lesions (34%) in the left lung, and 3 lesions (9%) in the trachea/ carina. There appeared to be some clustering of lesions (13 of 35 lesions [37%]) at the right mainstem and right upper lobe. Of these 13 lesions, 9 were obstructive, 2 were endoluminal, and 2 were mucosal. Lesions beyond the proximal segmental bronchi were difficult to visualize by FB due to the size of the endoscope (Fig 3). Reading here
VBs and SHR-CT scans exhibited an identical pattern of detection for these lesions (Fig 1) [Web movies 2 and 3], These modalities detected 29 of these abnormalities (83%) [19 of 19 obstructive lesions; 9 of 10 endolu-minal masses; and 1 of 6 mucosal lesions]. The sensitivities of SHR-CT scans and VBs were 100% for obstructive lesions, 90% for endoluminal masses, 17% for mucosal lesions, and 83% for all abnormalities. The specificities of SHR-CT scanning and VB were 100% (Table 1).
Six FN lesions were identified by FBs and were not detected by SHR-CT scans or VBs (Fig 3, top left, A, and top middle, B). These included a small peripheral endoluminal mass in one patient that was located in the left lower lobe, mucosal inflammation in two patients, and the presence of blood with no evidence of distal endobronchial lesions in three patients. With respect to these three cases of hemoptysis, it is possible that bleeding was intermittent and thus not present on the day of imaging, but was present on the day of FB.
Figure 1. Endoluminal lesion obstructing the superior segment of the left lower lobe in a 30-year-old gentleman in whom metastatic melanoma has been diagnosed. The visualization of this lesion (white arrow) took place by FB (top left, A), VB (top middle, B), and SHR-CT scan (bottom left, D: axial section; bottom middle, E: coronal section; and bottom right, F: sagittal section). However, this lesion was not seen by consecutive conventional CT sections (top right, C).
Figure 2. Location of lesions by FB. Each circle represents the anatomic location of a lesion, as follows: the black circle (19 lesions) represents an obstructive lesion (ie, a bronchial narrowing of > 50%); the red circle (10 lesions) represents an endolu-minal mass (ie, a mass protruding into the lumen with < 50% obstruction); and the blue circle (n = 6) represents a mucosal lesion (ie, hemorrhage, erythema, abnormal color, or tissue friability).
Figure 3. Location of FN lesions determined by SHR-CT scan (top left, A), VB (top middle, B), and CT scan (top right, C). Location of additional distal lesions not reachable by FB but nevertheless visualized by SHR-CT scan (bottom left, D), VB (bottom middle, E), and CT scan (bottom right, F). A black circle represents an obstructive lesion, a red circle represents an endoluminal lesion, and a blue circle represents a mucosal lesion. Lesions highlighted in yellow were confirmed to be positive by pathology.
Table 1—Comparison of Imaging Modalities for Detection of Tracheobronchial Malignancies
|Sensitivity||Specificity (Patient TN)|
|CT scan||50% (4/8)||72% (13/18)||0% (0/3)||59% (17/29)||85% (6/7)|
|SHR-CT scan||90% (9/10)||100% (19/19)||16% (1/6)||83% (29/35)||100% (9/9)|
|VB||90% (9/10)||100% (19/19)||16% (1/6)||83% (29/35)||100% (9/9)|