First published online July 5, 2006
British Journal of Radiology (2006) 79, 779-784
© 2006 British Institute of Radiology
doi: 10.1259/bjr/40749658
Virtual pulmonary arterioscopy in pulmonary embolic disease
C Hoskins, BSc, FRCR and
M Carpenter, BSc
Department of Diagnostic Imaging, Mayday University Hospital, Croydon CR7 7YE, UK
 |
Abstract
|
|---|
16 slice multidetector CT provides virtual endoscopic views of the inside of arteries, or any other hollow structures. This is performed non-invasively using post-processing of three-dimensional isotropic image data sets, acquired during standard CT examinations. These virtual endoscopic views are simultaneously correlated with the standard multiplanar reconstructions, with the ability to navigate a virtual camera through the hollow structure under study. Normal and abnormal volume rendered images of the pulmonary arteries are presented in correlation with the multiplanar reformats. The abnormal images show the volume rendered appearances of acute and chronic pulmonary embolic disease. It is also postulated that this technique has a problem solving role in the differential diagnosis of chronic mural emboli from extravascular structures such as adjacent lymph nodes or bronchiolar impaction. This technique may also have a role in medical education, providing clinicians and medical students with interactive three-dimensional representations of disease processes.
|
Introduction
|
|---|
The establishment of CT pulmonary angiography as the primary imaging technique for assessment of pulmonary embolic disease is widely recommended in many centres, as it combines the ability to directly visualize thrombus and evaluate the mediastinum and the lung parenchyma within one examination [1, 2]. Production of near isotropic data sets with 16+ slice multidetector CT has enabled the introduction and or refinement of numerous image processing techniques, avoiding the inherent distortion and artefacts associated with non-isotropic data [3].
Volume rendering (VRT) is one such technique that produces a 3D coloured image with depth perception. Unlike the earlier method of shaded surface display (SSD), VRT evaluates all the voxel intensities from the standard isotropic data set normally acquired for a routine examination. The VRT software assigns differing colours and degrees of transparency to CT number ranges within the volume and is thus able to display overlapping structures. It is also possible to position a virtual "endoscopic" camera within the arterial lumen guided by the multiplanar reconstruction (MPR) displays. This technique of "fly through" or perspective volume rendering (pVRT) allows visualization of the inner wall and any intraluminal contents of the arteries.
Development of VRT relied heavily on the advancement of computing power [4], but is now universally applied to virtual colonoscopy, virtual bronchoscopy and virtual angioscopy/interior vessel analysis [5]. Identified applications for virtual angioscopy include "fly through" coronary angiography [6] and qualitative assessment of carotid artery stenosis [7]. The prerequisites for good angioscopic VRT imaging are the reconstruction of thin overlapping slices and the existence of a strong interface between the Hounsfield Units of the structure of interest and its surroundings. For virtual angioscopy this interface is achieved with good contrast opacification of the blood vessels.
In this paper we show the pVRT appearance of the pulmonary arteries, both in normal cases and in cases of acute and chronic thromboembolic disease. These are correlated with the MPR reformats. We postulate that this post-processing non-invasive technique has the potential to be used as a problem solving tool in the differential diagnosis of chronic thromboembolic disease versus periarterial lymph nodes and bronchiolar impaction. We also briefly discuss a possible role in medical education and the visualization of disease processes.
|
Methods
|
|---|
Records were compiled of cases where a thoracic CT or pulmonary CT demonstrated acute or chronic pulmonary emboli, periarterial lymph nodes or a normal pulmonary arterial tree. These cases were protected on a workstation to enable future study.
CT scanning
CT pulmonary angiograms were performed using a Siemens Somatom Sensation 16 (Siemens AG Medical Systems, Germany). 100 ml of intravenous contrast was administered at a rate of 3 ml s–1 using an automated injector pump. Timing for the scan was influenced by the use of bolus triggering software (Siemens CARE Bolus) with a threshold of 100 HU and a region of interest positioned over the main pulmonary artery. Once triggered, a volume scan was undertaken following a 6 s delay through the whole chest caudocranially, using a slice collimation of 16 mm x 0.75 mm. Following raw data reconstruction (slice width of 1.0 mm, reconstruction interval of 0.7 mm) the resultant overlapping axial images were transferred to a workstation.
Data processing
Case images were reviewed using the interactive axial, sagittal and coronal MPR display with any additional reformatting (angulation, curved MPR, magnification) undertaken when required. Examinations were first pre-selected for adequate contrast enhancement of the pulmonary arteries, which was judged subjectively. MPR images of normal arteries, arteries with adjacent lymphatic tissue and arteries with features of acute or chronic emboli were created and saved. Chronic embolic features included mural thrombus and web formation.
From the selected images, pVRT or "fly through" views were acquired utilizing the standard post-processing software supplied with the scanner. The standard isotropic data set is manipulated to display 3D imagery by assigning different colour and degrees of transparency to different voxel Hounsfield Units within the volume. Whilst this rendering can be altered interactively by the user with regards to colour, opacity and brightness, our study employed the default setting. Further manipulation of the pVRT images is performed, locating areas of interest within the arteries by moving the virtual camera display on the MPR images. Subsequent views of the arterial lumen and emboli were obtained and saved.
|
Results
|
|---|
Figure 1
shows an axial reconstruction from a CT pulmonary angiogram with a diagrammatic representation of the virtual camera observing a normal main pulmonary arterial bifircation; together with a pVRT image of this area using different virtual camera positions within the arterial lumen.
View larger version (49K):
[in this window]
[in a new window]
|
Figure 1. Axial CT pulmonary angiography of normal main pulmonary bifurcation with volume rendered(pVRT) image.
|
|
Figure 2 shows a sagittal multiplanar reformat demonstrating an acute embolus appearing as a filling defect in the pulmonary arterial branch. The pVRT image clearly shows the occluding clot and an adjacent patent vessel.
Figure 3 shows a chronic embolus appearing as a flattened eccentric defect in contrast filling at an obtuse angle with the vessel wall on the right side. There is a web in the left interlobar pulmonary artery. The pVRT image demonstrates the intraluminal web.
View larger version (75K):
[in this window]
[in a new window]
|
Figure 3. Multiplanar reconstruction and perspective volume rendering(pVRT) images of chronic emboli and webs.
|
|
Figure 4 shows chronic mural embolus in the left pulmonary artery on both coronal and sagittal MPRs, with the sagittal MPR demonstrating the position of the virtual camera. The pVRT image shows the intra-arterial nature of the embolus.
View larger version (112K):
[in this window]
[in a new window]
|
Figure 4. Multiplanar reconstruction and perspective volume rendering(pVRT) images of chronic embolus.
|
|
Figure 5 shows both axial and coronal views of a follow up scan in a patient with long term perivascular lymphatic tissue unchanged from previous scans. PVRT imaging confirms that this perivascular tissue is not intra-arterial.
View larger version (95K):
[in this window]
[in a new window]
|
Figure 5. Multiplanar reconstruction of perivascular lymphatic tissue with confirmatory perspective volume rendering(pVRT) image of arterial lumen.
|
|
|
Discussion
|
|---|
With the advent of 16 (and more) slice CT and subsequent acquisitions of near perfect isotropic volume data, there has been an explosion in the use of post-processing techniques. These include maximum intensity projections (MIP), SSD, VRT, pVRT and MPR reconstructions which have benefited from a remarkable increase in resolution as well as software improvements, which have allowed them to be used easily and quickly with the minimum of training. Post-processing with CT pulmonary angiography includes the standard multiplanar reformats as well as curved or other reformats individually tailored to the arterial branch under study. MIP algorithms have been most commonly used for the assessment of vascular structures, extracting the highest attenuating voxels along a specified direction and adding this to the projected image, providing a display demonstrating the blood vessels well but suppressing the background tissue. Resultant images are rotatable, allowing elimination of overlap. However, unlike VRT and pVRT there is no sense of depth as the MIP image is essentially a collapsed 3D structure imposed onto a 2D surface without perspective [4]. Furthermore, as high attenuation suppresses low attenuation, this reconstruction technique can obscure low attenuation thrombus. Chronic embolic features such as webs and mural thrombus are also difficult to see even with relatively thin slice MIP imaging [3].
Characteristic CT features of chronic pulmonary thromboembolic disease include webs, bands, intimal irregularities and eccentric flattened defects at an obtuse angle with the vessel wall. Thrombi can be simulated by periarterial lymph nodes [8].
While to our knowledge this volume rendering imaging has not been studied in relation to pulmonary embolic disease, we feel it may prove useful in problem solving cases where the differential lies between intravascular but perimural pathology, such as chronic embolus, and extravascular pathology, such as bronchiolar impaction or periarterial lymphatic tissue, situated adjacent to the arterial wall. It may also prove useful in the display of the intravascular findings of chronic PE such as mural thrombus and webs, which are often extremely subtle [3].
Volume rendering can play a role in medical education with the improved visual presentation of disease processes. Several CT image reformatting approaches such as sagittal/coronal, oblique, curves and variable thickness viewing helps to orientate referring clinicians and radiologists to particular anatomical structures and pathology, allowing selective display and enhancement of relevant findings [3, 5]. However, these methods still require the user to think in sections whereas volume rendered imagery provides realistic three-dimensional pictures of intraluminal disease processes, which if studied on an interactive CT workstation can provide excellent learning opportunities to health care students.
Received for publication January 6, 2006.
Revision received April 12, 2006.
Accepted for publication May 19, 2006.
|
References
|
|---|
- Schoepf UJ, Goldhaber SZ, Costello P. Spiral computed tomography for acute pulmonary embolism. Circulation 2004;109:2160–7.[Abstract/Free Full Text]
- Reidel M. Diagnosing pulmonary embolism. Postgrad Med J 2004;80:309–19.[Abstract/Free Full Text]
- Gruden JF, Tigges S, Baron MG, Pearlman H. MDCT pulmonary angiography: image processing tools. Semin Roentgenol 2005;40:48–63.[CrossRef][Medline]
- Edelman RR, Rabin DN. 3D imaging technologies. J Imaging Technology Management. www.ImagingEconomics.com, 2001
- Cody DD. AAPM/RSNA physics tutorial for residents: topics in CT. Imaging and Therapeutic Technology 2002. 22:1255–68.
- Van Ooijen PM, Oudkerk M, Van-Geuns RJ, Rensing BJ, De-Feyter PJ. Coronary artery fly through using electron beam computed tomography. Circulation 2000;102:E6–10.
- Do-Yeon Kim, Jong-Wong P. Visualisation of the internal carotid artery using MRA images. Magn Reson Imaging 2005;23:27–33.[CrossRef][Medline]
- Han D, Kyung SL, Franquet T, Muller L, Sung Kim T, Hojoong K, et al. Thrombotic and nonthrombotic pulmonary arterial embolism: spectrum of imaging findings. RadioGraphics 2003;23:1521–39.[Abstract/Free Full Text]
Top
Abstract
Introduction
