| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Institute of Diagnostic Radiology, Inselspital, University of Berne,
Freiburgstr. 20, CH-3010 Berne, Switzerland.
2 Department of Pneumonology, Inselspital, University of Berne, CH-3010 Berne,
Switzerland.
Received July 19, 2001;
accepted after revision October 30, 2001.
Address correspondence to H.-P. Dinkel.
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. Multidetector CT and flexible bronchoscopy were used to examine 200 bronchial sections obtained from 20 patients (15 patients with bronchial carcinoma and five without central airways disease). Multidetector CT was performed using the following parameters: collimation, 4 x 2 mm, pitch, 1.375; and reconstruction intervals, 2 mm. Postprocessing was performed using surface rendering and multiplanar reformatted images. CT images were independently interpreted by two radiologists. The tracheobronchial stenoses revealed on flexible bronchoscopy were graded by a pulmonologist.
RESULTS. Virtual bronchoscopic findings, axial CT scans, and multiplanar reformatted images were highly accurate (98% accuracy for virtual bronchoscopic images, 96% for axial slices and coronal reformatted images, and 96.5% for sagittal reformatted images) in revealing tracheobronchial stenosis. In allowing accurate grading of tracheobronchial stenosis, images from virtual bronchoscopy correlated closely (r = 0.91) with those of flexible bronchoscopy. Because use of virtual bronchoscopic images reduced the overestimation of stenosis, these images allowed better assessment of stenosis than did axial CT slices (r = 0.84) or multiplanar reformatted images (r = 0.84) alone.
CONCLUSION. Multidetector CT virtual bronchoscopy is a reliable noninvasive method that allows accurate grading of tracheobronchial stenosis. However, it should be combined with the interpretation of axial CT images and multiplanar reformatted images for evaluation of surrounding structures and optimal spatial orientation.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The capacity of virtual bronchoscopy to depict tracheobronchial stenosis has been proven with single-detector helical CT [5, 6], but the z-axis resolution has been limited by scan collimation that could not be reduced below a certain level (3-5 mm) if complete scan acquisition of the thorax was to be achieved during a single breath-hold. In our study, we compared various multidetector CT display modes (virtual bronchoscopy, axial CT, coronal reformatting, and sagittal reformatting) separately to the gold standard of flexible bronchoscopy in revealing and providing the detail required for grading tracheobronchial stenosis. To our knowledge, ours is the first report on the use of multidetector CT virtual bronchoscopy for the evaluation of central airways stenosis.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Multidetector CT Technique
CT examinations were performed on a multidetector CT scanner (Asteion;
Toshiba, Tokyo, Japan) with the following parameters: collimation, 4 x 2
mm; pitch, 1.375 (corresponding to manufacturer's pitch of 5.5); rotation
time, 0.75 sec; 120 kVp; and 100-180 mAs. Acquisition time was roughly 30 sec
to allow completion of the acquisition during a single breath-hold. The thorax
was scanned during inspiration in a caudocranial direction after power
injection of 80 mL (flow rate, 2 mL/sec; scan delay, 30 sec) of iopromide
(Ultravist; Schering, Berlin, Germany) IV contrast medium containing 300 mg
I/mL. The reconstruction intervals and slice thickness were 2 mm.
Image Reconstruction
Axial CT images were transferred to a workstation (Advantage Windows 4.0;
General Electric Medical Systems, Milwaukee, WI) running on hardware (Ultra
Sparc 60; Sun Microsystems, Mountain View, CA) that featured two 450-MHz
central processing units (Ultra Sparc II; Sun Microsystems) and 512-MB of
random access memory. Reconstruction software (Navigator, version 2.03;
General Electric Medical Systems) was used for virtual bronchoscopic images.
The image display used a surface-rendering algorithm and produced perspective
gray-scale images with a matrix of 512 x 512. Image segmentation was
based on thresholding. All voxels with a density under the threshold level
were considered to be within the bronchial lumen. An upper threshold of -500 H
was applied to reconstruct the airways (black-on-white mode)
[7]. Each bronchoscopic image
simulated a coned-down view, with a cone angle adjusted to 45°. Navigation
through the tracheobronchial tree was performed in the fly-through mode
beginning in the trachea. In addition, reformatted coronal and sagittal images
of the tracheobronchial tree were produced.
Flexible Bronchoscopy
Flexible bronchoscopy was performed by an experienced pulmonologist using a
videobronchoscope (BF-1T200; Olympus Optical, Tokyo, Japan) with the patients
under local anesthesia. The mean interval between virtual bronchoscopy and
flexible bronchoscopy was 9.4 days (range, 1-17 days). To provide a standard
of reference, an experienced pulmonologist reviewed the flexible bronchoscopy
reports in random order without knowledge of the CT findings or clinical
histories (derived from written bronchoscopy reports). The grade of
tracheobronchial narrowing was categorized semiquantitatively as grade 1
(luminal narrowing < one third), grade 2 (luminal narrowing 
one third
but < two thirds), and grade 3 (luminal narrowing two thirds).
CT Image Analysis
Two radiologists who were unaware of flexible bronchoscopic findings and
clinical histories independently performed CT image analysis. The axial CT
images were interpreted first, followed by the coronal and sagittal
reformatted images. Finally, virtual bronchoscopy in the multiview mode was
performed. In cases of disagreement, a consensus was achieved in a final
common interpretation session. In the multiview mode, the screen is divided
into four quadrants that display virtual bronchoscopic images, axial CT
images, sagittal reformatted images, and coronal reformatted images
simultaneously on one screen (Fig.
1A,1B,1C,1D).
The axial CT images and multiplanar reformatted images were viewed with
standard lung window settings (level, -450 H; width, 1850 H) and standard
soft-tissue window settings (level, 50 H; width, 450 H). For image analysis,
the central airways of each patient were divided into 10 separate sections:
trachea, right main bronchus, intermediate bronchus, right upper lobe
bronchus, middle lobe bronchus, right lower lobe bronchus, left main bronchus,
left upper lobe bronchus, lingula bronchus, and left lower lobe bronchus. A
total of 200 sections were evaluated. For each bronchial section, the imaging
findings for the degree of luminal narrowing of virtual bronchoscopy, axial
CT, coronal reformatting, and sagittal reformatting were separately correlated
with those of flexible bronchoscopy. The semiquantitative grade of
tracheobronchial narrowing was categorized separately for virtual
bronchoscopic images, axial CT images, and multiplanar reformatted images.
Each case of bronchial stenosis was rated as grade 1 (luminal narrowing <
one third), grade 2 (luminal narrowing one third but < two thirds),
and grade 3 (luminal narrowing two thirds).
,,,
|
|
|
|
|
|
|
|
Statistical Analysis
Qualitative results regarding the presence of stenosis were classified as
true-positive, true-negative, false-positive, or false-negative findings. A
true-positive finding was defined as stenosis noted on both flexible
bronchoscopy and CT. A true-negative finding was an airway section described
as normal based on both modalities. A false-positive finding was a bronchial
section noted as abnormal on CT but normal on flexible bronchoscopy. A
false-negative finding was a stenosis verified by flexible bronchoscopy that
appeared to be normal on CT. The sensitivity, specificity, and accuracy of the
respective modes were calculated from 2 x 2 contingency tables, with
confidence intervals being derived from binomial distribution. Fisher's exact
test was used to test for significance when the chi-square test was not
applicable in cases with five or fewer field entries.
Assessments by the different methods were defined as overestimations or underestimations according to whether the particular stenosis had been assigned a higher or lower grade at the review of the findings of flexible bronchoscopy. Spearman's rank order correlation (r) was calculated to measure the strength of correlation between the results of CT (virtual bronchoscopy, axial slices, sagittal reformatted images, and coronal reformatted images) and flexible bronchoscopy. A p value of less than 0.05 was considered significant.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Virtual bronchoscopy depicted 22 stenoses. Among these, findings for two grade 1 stenoses were false-positive, one located in the left main bronchus and one in the right upper lobe bronchus. Findings for two stenoses were false-negative, one located in the left main bronchus and one in the right lower lobe bronchus.
Accuracy for diagnosis of stenosis was highest using virtual bronchoscopy (98%), followed by sagittal reformatted images (96.5%), coronal reformatted images (96%), and axial slices (96%), as shown in Table 1. However, sensitivity was higher for axial slices (95.5%) than for virtual bronchoscopic images (90.9%). Sagittal reformatted images had the lowest sensitivity (81%) in depicting stenosis.
|
Virtual bronchoscopy was better than all other CT methods in revealing detail needed for the accurate grading of tracheobronchial stenosis. Eighteen of the 22 stenoses were correctly graded using virtual bronchoscopy. The correlation between the findings of virtual bronchoscopy and flexible bronchoscopy was excellent (r = 0.91). Fewer stenoses were correctly graded using axial CT images (10/22, r = 0.84), coronal reformatted images (10/22, r = 0.84), or sagittal reformatted images (12/22, r = 0.84) (Table 2). In no case did the reviewers misinterpret the grade of stenosis by more than one degree. Using virtual bronchoscopy, reviewers overestimated the grade of stenosis once and underestimated the grade three times. Axial CT images and coronal reformatted images tended to lead interpreters to overestimate the grade of stenosis, whereas virtual bronchoscopy and sagittal reformatted images tended to lead interpreters to underestimate it. Thirteen stenoses had remaining airways lumina too narrow to be passed using virtual bronchoscopy compared with 11 stenoses that were impassable using flexible bronchoscopy.
|
The quality of CT data sets was excellent in all patients. Breathing artifacts were not observed. In all patients, moderate cardiac motion artifacts were observed, which, in the region of the bifurcation, appeared as normal ring structures; these artifacts were not observed in the peripheral bronchi. Cardiac artifacts could be easily identified and had no effect on the grading of tracheobronchial stenosis.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Virtual bronchoscopy produces a three-dimensional fly-through view of the tracheobronchial tree from CT data [2]. It takes advantage of the natural contrast between the air-containing lumen and the surrounding tissues. The specific advantage of using virtual bronchoscopy to grade stenosis is that the perspective of the endoluminal view, unlike that in orthoplanar images, stays within the axis of the airway, allowing reliable semiquantitative assessment of tracheobronchial stenosis. Accordingly, virtual bronchoscopic images were more accurate (98%) in depicting tracheobronchial stenosis than were axial CT slices or multiplanar reformatted images and showed an excellent correlation with flexible bronchoscopic findings in providing the detail necessary for accurate grading of tracheobronchial stenosis.
Virtual bronchoscopy proved to be reliable for semiquantitative assessment of tracheobronchial stenosis and was superior to other CT display modes because it produced more realistic images of the narrowed bronchial lumen. Conversely, reproducible quantitative measurements of the bronchial diameter are problematic with virtual bronchoscopy. For absolute measurement, calipers must be set exactly to the two points that represent the shortest distance within the bronchial lumen. Because of perspective and depth orientation, such exactness is difficult to achieve in the fly-through mode.
An intrinsic limitation of virtual bronchoscopy is its inability to depict mucosal color and texture [2]. Unlike virtual bronchoscopy, flexible bronchoscopy allows endoluminal visualization of the tracheobronchial tree and permits therapeutic and diagnostic maneuvers, such as obtaining specimens for culture and biopsy or removing foreign bodies [8]. However, flexible bronchoscopy may be uncomfortable for the patient, and sedation may be required. Complications related to the procedure itself and to anesthesia have been well described [9].
However, the complications of flexible bronchoscopy must be weighed against the considerable amount of diagnostic and therapeutic information gained from the procedure. Bronchoscopy and CT should not be pitted against each other but viewed as complementary techniques. In this sense, virtual bronchoscopy may provide important diagnostic and potentially therapeutic information before flexible bronchoscopy is undertaken. Furthermore, virtual bronchoscopy may be used to evaluate patients with known tracheobronchial stenosis after treatment and may reduce the frequency of repeated flexible bronchoscopy performed for that purpose.
Axial CT slices are indispensable tools for thoracic CT diagnosis that provide a wealth of anatomic and pathologic information beyond the view of a bronchoscope. In addition to the endoluminal airways, the surrounding mediastinal structures, lung parenchyma, breast wall, supraclavicular region, axilla, skeleton, and upper abdomen are shown. Axial slices permit differentiation between endoluminal tracheobronchial stenosis and extraluminal airways compression; axial slices can aid in detection of lymph node metastasis. In our study, axial CT had a sensitivity of 95.5% and a specificity of 96.1% in depicting tracheobronchial stenosis. However, misinterpretation of stenosis was more frequent with axial CT than with virtual bronchoscopy. This finding may be due to a combination of factors, including perceptual difficulties in the accurate assessment of luminal caliber on sequential axial images, especially in images in which longitudinal structures display a course parallel or oblique to the scanning plane.
Secondary reformatted images offer a means of reducing the flood of data. Besides virtual bronchoscopy, multiplanar reformatted images have proven to be advantageous in detecting central airways stenosis and particularly in evaluating the length of a stenosis [10]. Although coronal and sagittal reformatted images were accurate (96% for coronal reformatted images and 96.5% for sagittal reformatted images) in depicting tracheobronchial stenosis in our study, semiquantitative misinterpretation of the grade of stenosis was more frequent with multiplanar reformatted images than with virtual bronchoscopic images. However, quantitative measurement of a stenosis is more reliable with multiplanar reformatted images than with the "down-the-barrel" view obtained with bronchoscopy [11]. In our study, the sensitivity for revealing stenosis was higher for coronal reformatted images than for sagittal reformatted images, which may be explained by the fact that radiologists are less accustomed to interpreting the sagittal reformatted images. Besides depicting stenosis, multiplanar reformatted images also permit differentiation between intraluminal tumor growth and extraluminal airways compression and may give important information about adjacent structures, such as lymph nodes, and the extent of transmural tumors.
Although intraluminal findings can be clearly seen on virtual endoscopy, visualization and orientation may be difficult at times. Various navigational aids for virtual endoscopy have been described, including the multiview mode and collision avoidance [12]. Our study used the multiview mode, which allowed simultaneous viewing of virtual bronchoscopic images in the fly-through mode in addition to axial, coronal, and sagittal reformatted images on one screen for global orientation. Collision avoidance, a software tool that helps prevent penetration of the airways during subtotal stenosis, was not used in our study, which may explain the high number of impassable stenoses in our patients.
The ability of virtual bronchoscopy to depict tracheobronchial stenosis has been proven before with single-detector helical CT [2, 13]. A preliminary study concluded that virtual bronchoscopy could render stenosis in anatomic detail but that reconstruction time was slow and that further studies were needed to evaluate diagnostic potential of the modality [5]. Another study compared the diagnostic value of virtual bronchoscopy with those of axial CT images, multiplanar reformatted images, minimal-intensity-projection images, and images from flexible bronchoscopy obtained with a single-detector CT scanner using a thin collimation [6]. The results of that study were similar to our own: Virtual bronchoscopy was found to have excellent sensitivity and specificity for depicting tracheobronchial stenosis.
A limitation of single-detector helical CT is the relatively long scanning times needed for a narrow collimation; patients with malignant tracheobronchial stenosis may not be able to hold their breath for a sufficient length of time, causing breathing and motion artifacts that may lead to misinterpretation of the tracheobronchial lumen [11]. We found that multidetector CT was excellent in generating CT data sets that enabled high-resolution endoluminal visualization and multiplanar reformation because scanning time could be kept short despite thin collimation. Thus, breathing artifacts were not observed among our patients. Moderate cardiac motion artifacts that appeared as regular ring structures were present in the region of the bifurcation but did not impair the evaluation of tracheobronchial stenosis and were not observed in the peripheral bronchi. In the future, it can be expected that ECG-gated multidetector scanners may help to eliminate pulsation artifacts.
Virtual bronchoscopy uses surface rendering, which takes advantage of the natural contrast between the airway and surrounding tissues. Therefore, the level of thresholding is important in displaying accurate simulations. For reasons of standardization, we used the fixed threshold of -500 H used by Zeiberg et al. [7]. Our study focused on the proximal airways within the fourth-order bronchi, for which that threshold has been shown to be appropriate [7]. However, threshold values may have to be adapted for examination of segmental and subsegmental bronchi because in smaller bronchi, incorrect threshold values may be a considerable source of error [11].
Patients with incurable bronchial carcinoma often have severe bronchial stenosis and posttumoral atelectasis. In these patients, palliative treatment such as laser therapy, irradiation therapy, or stent implantation may become necessary; it is important that clinicians know the exact grade of tracheobronchial narrowing. In our study, misinterpretation of stenosis occurred more often with isolated interpretation of axial CT images than with virtual bronchoscopy. Compared with axial CT images and multiplanar reformatted images, images from virtual bronchoscopy yielded the highest number of correctly graded stenoses and provided important additional information. Besides the assessment of tracheobronchial stenosis, virtual bronchoscopy may also have a role in guiding transbronchial needle aspiration of lymph nodes [14, 15].
In conclusion, multidetector CT virtual bronchoscopy is a reliable, noninvasive method for endoluminal assessment of central airways stenosis when combined with the interpretation of axial and multiplanar reformatted images. Axial slices remain indispensable because they are the primary imaging mode for the assessment of extraluminal thoracic findings. Multidetector CT—unlike single-detector CT—has the advantage of providing high-resolution multiplanar reformatted images because of its high z-axis resolution. Thus, in the future, when applications of virtual bronchoscopy are expanded to include the examination of segmental and subsegmental bronchi, use of multidetector CT may have considerable advantages related to a further reduction in collimation. It remains for future studies to elucidate these issues and their clinical implications.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  J. P. Williamson, A. L. James, M. J. Phillips, D. D. Sampson, D. R. Hillman, and P. R. Eastwood Quantifying tracheobronchial tree dimensions: methods, limitations and emerging techniques Eur. Respir. J., July 1, 2009; 34(1): 42 - 55. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. M. Horton, M. R. Horton, and E. K. Fishman Advanced Visualization of Airways with 64-MDCT: 3D Mapping and Virtual Bronchoscopy Am. J. Roentgenol., December 1, 2007; 189(6): 1387 - 1396. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Fluoroscopy-assisted thoracoscopic resection of pulmonary nodules after computed tomography-guided bronchoscopic metallic coil marking. J. Thorac. Cardiovasc. Surg., March 1, 2006; 131(3): 704 - 710.
|
|
|
|
 |
|
 D. Shitrit, P. Valdsislav, A. Grubstein, D. Bendayan, M. Cohen, and M. R. Kramer Accuracy of Virtual Bronchoscopy for Grading Tracheobronchial Stenosis: Correlation With Pulmonary Function Test and Fiberoptic Bronchoscopy Chest, November 1, 2005; 128(5): 3545 - 3550. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. M. Jones and T. Athanasiou Is Virtual Bronchoscopy an Efficient Diagnostic Tool for the Thoracic Surgeon? Ann. Thorac. Surg., January 1, 2005; 79(1): 365 - 374. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. Kosucu, A. Ahmetoglu, I. Koramaz, F. Orhan, O. Ozdemir, H. Dinc, A. Okten, and H. R. Gumele Low-Dose MDCT and Virtual Bronchoscopy in Pediatric Patients with Foreign Body Aspiration Am. J. Roentgenol., December 1, 2004; 183(6): 1771 - 1777. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 W K Chooi and S K Morcos High resolution volume imaging of airways and lung parenchyma with multislice CT Br. J. Radiol., December 1, 2004; 77(suppl_1): S98 - S105. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. Onodera, T. Omatsu, S. Takeuchi, N. Shinagawa, K. Yamazaki, T. Nishioka, and K. Miyasaka Enhanced Virtual Bronchoscopy Using the Pulmonary Artery: Improvement in Route Mapping for Ultraselective Transbronchial Lung Biopsy Am. J. Roentgenol., October 1, 2004; 183(4): 1103 - 1110. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W. De Wever, V. Vandecaveye, S. Lanciotti, and J.A. Verschakelen Multidetector CT-generated virtual bronchoscopy: an illustrated review of the potential clinical indications Eur. Respir. J., May 1, 2004; 23(5): 776 - 782. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. Lacasse, S. Martel, A. Hebert, G. Carrier, and B. Raby Accuracy of virtual bronchoscopy to detect endobronchial lesions Ann. Thorac. Surg., May 1, 2004; 77(5): 1774 - 1780. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. Shinagawa, K. Yamazaki, Y. Onodera, K. Miyasaka, E. Kikuchi, H. Dosaka-Akita, and M. Nishimura CT-Guided Transbronchial Biopsy Using an Ultrathin Bronchoscope With Virtual Bronchoscopic Navigation Chest, March 1, 2004; 125(3): 1138 - 1143. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. Hoppe, H.-P. Dinkel, B. Walder, G. von Allmen, M. Gugger, and P. Vock Grading Airway Stenosis Down to the Segmental Level Using Virtual Bronchoscopy Chest, February 1, 2004; 125(2): 704 - 711. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. M. Boiselle, G. Dippolito, J. Copeland, H. Kang, M. Ahmed, D. Weeks, E. Halpern, and S. N. Goldberg Multiplanar and 3D Imaging of the Central Airways: Comparison of Image Quality and Radiation Dose of Single- Detector Row CT and Multi- Detector Row CT at Differing Tube Currents in Dogs Radiology, July 1, 2003; 228(1): 107 - 111. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
