Chest: Multidetector Computed Tomography of the Tracheo-Bronchial Tree with Three-Dimensional Volume Rendering-a Problem Solving Tool

Leo P. Lawler, M.D.1 Edward F. Haponik, M.D.2 Elliot K. Fishman, M.D., FACR1
1From the Russell H. Morgan Department of Radiology and Radiological Science
2From the Division of Pulmonary and Critical Care Medicine,
Johns Hopkins University
Corresponding author:    Elliot K. Fishman, M.D., FACR    The Russell H. Morgan Department of Radiology and

Radiological Science,
Johns Hopkins University,
601 North Caroline Street/Room 3253,
Baltimore, Maryland 21287
Phone: (410) 9556500
Fax:    (410) 6143896
E-mail: [email protected]

OBJECTIVE. We describe the combined use of thin section multidetector computed tomography (MDCT) with three-dimensional volume rendering to image the airways as a problem solving tool.

CONCLUSION. New advances in multidtector CT technology now compliment similar recent progress in volume rendering technology. This powerful combination allows easy management of the very large, high quality, multidetector data sets as well as providing a platform for current software to interactively exploit these data sets’ potential for clinical use. We believe MDCT and volume rendering techniques represent a new plateau in airway imaging.

KEYWORDS. Airway, computed tomography, volume rendering

INTRODUCTION. As in other applications of three dimensional computed tomography, the advent of multidetector computed tomography (MDCT) scanning has led to the ability to acquire images with both large areas of coverage and high resolution, making them ideally suited to three-dimensional (3D) reconstruction. This technology has been successfully applied in angiography and orthopedics but its role in airway imaging has yet to find widespread acceptance. These new techniques allow near isotropic, high quality data sets to be obtained and now permit real time study of high fidelity three-dimensional renderings.

We have recently employed these technological advances in several clinical cases and think the findings illustrate a landmark in airway imaging and suggest clear potential for future use of the tracheo-bronchial 3D CT in patient management.





MATERIALS AND METHODS. Four patients were referred for CT scanning with three-dimensional imaging to solve difficult clinical dilemmas. There were two women and two men, mean age 57 years, range 48 to 75 years. All patients were attending the pulmonology clinic at our institution. One patient had suffered an acid inhalation injury, one had suffered adult respiratory distress syndrome (ARDS) post partial lobectomy surgery, one patient had a single lung transplant for usual interstitial pneumonitis (UIP) complicated by bronchial anastamotic stricture and the final case had been recently diagnosed with usual interstitial pneumonitits (UIP). At the time of referral for imaging all patients were suffering increasing dyspnea. All patients had had prior clinical evaluation, pulmonary function tests, chest radiographs and tissue biopsy where indicated. All patients had had prior conventional computed tomography either at our institution or elsewhere. In all patients the specific clinical questions related to airways. In two patients the issues were airway stenosis and airway stent management and in the other two the clinical questions pertained to bronchiectasis.

Imaging was performed in all patients with the same protocol using a Siemens Plus 4 Volume Zoom (Siemens Medical Systems, Iselin, New Jersey). We used 140 KVp, 100 mAs and a rotation time of 0.5 seconds. 1-mm collimation, 1.25-mm slice thickness and 1mm data reconstructions were used. The table speed per rotation was 6mm with a pitch of 6. No contrast was administered.

Immediately after data acquisition the data sets were transferred to a prototype Siemens 3D Virtuoso (Siemens Medical Systems, Iselin, NJ) workstation. Each case was analyzed by varying the trapezoidal transfer functions to highlight tissues of interest and interactive real-time change in projections to focus on specific regions of interest. Clip planes are used to selectively edit out tissues as desired by the user.

RESULTS. Our first patient was a 48 year old male who had suffered multiple tracheobronchial stenoses from acid inhalation. An initial three-dimensional CT (3D CT) showed a patent stented airway just beyond a narrow left main stem orifice (Fig1A). It also revealed long segment of narrowing of the lower two thirds of the trachea, significant narrowing of the right bronchus main stem bronchus/intermedius and narrowing of the right middle lobe bronchus origin (Fig1A.). A subsequent nuclear medicine ventilation study demonstrated poor ventilation of both lungs and significant tracer retention. A repeat 3D CT three weeks later showed patent tracheal and right middle lobe stented airways but suggested that granulation tissue was causing a flap of tissue to partially occlude the left main stem stent (Fig1B, C.). This was confirmed and treated bronchoscopically and a follow-up 3D CT showed a new stent traversing the granulation tissue and a patent lumen as well as a new right main-stem stent (Fig1E.).

Our second patient was a 50-year-old male patient who suffered an anastomotic stricture due to bronchomalacia after left lung transplant. 3D CT showed optimal positioning of a left main stem stent, a patent airway and an enlarged trachea (Fig 2A,B).

The third case was a 75-year-old female patient recently diagnosed with usual interstitial pneumonitis but with a significant obstructive component. A 3D CT revealed dilated central bronchi consistent with bronchiectasis (fig3A). The distribution and extent of bronchiectasis and bilateral interstitial fibrosis and honeycomb pattern was demonstrated (Fig3B).

Our last patient was a 58 year old female with a history of ARDS after left lower lobe resection for breast metastasis. 3D CT showed tracheomegally and saccular left lower lobe bronchiectasis (Fig 4A). Again a clear picture of the extent of bronchiectasis and associated lung parenchyma change was formed (Fig4B).

DISCUSSION. To harness the full potential of the latest technology in 3DCT at this time requires the use of MDCT and advanced volume rendering. In any three-dimensional imaging the final result is only as good as the data acquisition. This is where the advances in MDCT represent a large step in the quality of three-dimensional imaging. Likewise high quality data sets are of little practical use if there are not robust workstations and graphics software to handle the large information and to use it to its full potential.

MDCT provides subsecond, true volume potentially isotropic acquisitions with large coverage and narrower collimation giving better resolution. Thin-section volumetric studies of large imaging thoracic volumes are tailor-made for volume rendering. Most radiologists agree that thin sections are optimal for airway imaging. Unlike in conventional spiral CT we no longer have to compromise between z-axis coverage and breath holding. High pitch and high-speed gantry rotation allow greater table speed and coverage with less than 10 seconds needed for a routine thoracic study with narrow collimation. Airway disease often affects multiple disparate branches as in our patients. Thus it is important to be able to perform whole lung volume rendered images with ease, as opposed to being confined to a pre-elected small area of interest as in the past. The MDCT high temporal resolution limits breathing and cardiac motion artifacts that commonly accompany patients presenting for airway imaging and MDCT also results in diminished helical artifacts in 3D imaging. MDCT with its ability to reconstruct slices of varying thickness after scanning allows one to produce relatively thick slices for axial interpretation and to create relatively much thinner slices for 3D volume rendering.

Volume rendering converts the volume of information acquired by MDCT into a simulated three-dimensional form that surpasses the previous techniques of shaded surface display (SSD) and maximum/minimum intensity projection (MIP) which are limited by threshold voxel selection and inability to maintain three dimensional relationships respectively (1,2,3). Also MIP is unable to show airways.

The final, volume rendered 3D image is the computed sum of voxels along a ray projected through the data set in a specific orientation and thus all the MDCT data is utilized to form the final image. The volume rendering technique with its attendant percentage-based probabilistic classification assigns a continuous range of values to a voxel allowing the percentage of different tissue types to be reflected in the final image while maintaining three-dimensional spatial relationships. This, unlike the all-or-none thresholding binary classification, is far less susceptible to volume averaging artifacts which are particularly a problem with airway imaging due to the air-soft tissue interface and the small caliber more peripheral branches. The trapezoidal transfer function (1,2) creates a histogram of displayed houndsfield units permitting control over window width and level, opacity and brightness so that particular combinations of tissues may be displayed. In our cases we were thus able to generate images that showed either the airways, stents or both to best advantage.

The development of clip planes removes the need for labor intensive manual drawing of regions of interest and has made it practical in airway imaging to the entire lung with ease. This removes the potential for inaccuracies of manual drawing and has made central as well as peripheral airways equally easy to visualize. This is of great advantage when one has to follow the tortuous airways of bronchiectasis as in two of our cases. The real-time interactivity represents a significant advance in 3DCT volume rendering, particularly, for airway imaging and compliments clip plane editing where the complex, numerous branching structures often have to be rotated in multiple projections to allow their full delineation. This not only facilitates the radiologist’s processing and interpretation of the data but also helps communication with non-radiologists. Volume rendering has higher fidelity to the high quality MDCT patient data and therefore should lead to more accurate clinical utility in airway imaging.

Latest developments show that volume rendering of such data are no longer a cumbersome luxury but are an efficient, easy to use necessity for interpreting the routine 300-500 images possible with MDCT.

From a clinical perspective it is fair to say that for many lesions 3DCT does not have a greater sensitivity than conventional axial images but it does confer great advantage in describing spatial relationships of airway disease and communicating this to the clinical service (4,5,6). A pre-procedure bronchoscopic map has clear potential in defining the number, location and extent within an individual bronchus as well as the overall lobar and segmental distribution of both areas of stenoses and bronchiectasis. Abnormal luminal caliber represents a subject where the two dimensional axial plane limits the observers’ (radiologist and non-radiologist) ability to appreciate these characteristics and volume rendering with faithful mapping of the airways is better able to compliment the bronchoscopist’s efforts. There are many indications now for stent placement including malignant neoplasm, tracheal stenosis of many causes and tracheomalacia. Stent choice includes mutliple factors but the length is dictated in part by the disease extent and luminal diameter. The goal is to traverse the diseased portion leaving the stent ends to be embedded in normal airway mucosa. If too short a second overlapping coaxial stent may be required and if too long it may protrude and cover airway openings (7). Volume rendering is ideally suited to this purpose. After stent placement it is the ability of thee-dimensional imaging to monitor for migration, compression and foreign body reaction with granuloma formation that makes volume rendering an ideal choice for non-invasive follow-up. In our cases of bronchial stenoses, multidetector and volume rendered 3DCT created images that clearly defined central airway stenoses, demonstrated well the site of stent placement and allowed non-invasive follow-up to forewarn of a potential stent block which was treated before it led to respiratory compromise. For patients with obstructive airway disease due to bronchiectasis it is the three dimensional ability to generate the faithful airway model of the patients disease severity and distribution that helps clinicians conceptualize the airway disease that accounts for the abnormal pulmonary function tests. This is of value in the initial evaluation of respiratory compromise but should also provide a basis for follow-up of disease dynamics over time. For our patients with bronchiectasis we were able to create a faithful rendering of the disease that defined their clinical condition. All the information on an individual’s airways is of course obtained above and beyond the concomitant acquiring of normal data regarding mediastinum and lung parenchyma etc.

The four cases discussed here illustrate the many ways the newest technology is allowing radiologists to interact with the data they are acquiring. MDCT has contributed greatly to advances in the volume formation step of three-dimensional imaging and the resultant data can now be better classified and projected. It would appear to us that 3DCT has reached a complimentary role to conventional tracheobronchoscopy in the evaluation of the airway just as it compliments conventional angiography. These cases illustrate diseases affecting airway caliber but volume rendering also has far wider implications in thoracic imaging (8).



REFERENCES:

1. Calhoun PS, Kuszyk B, Heath DG, Carley JC, Fishman EK. Three-dimensional volume rendering of spiral CT data: Theory and method. Radiographics 1999; 19:745-764

2. Remy-Jardin M, Remy J, Artaud D, Friburg M, Naili A. Tracheobronchial tree: Assessment with volume rendering –Technical aspects. Radiology 1998; 208:393-398

3. Fishman EK, Magid D, Ney DR, Chaney EL et al. Three-dimensional imaging. Radiology 1991; 181:321-337

4. Naidich DP, Grudrn JF, McGuinness G, McCauley DI, Bhalla M. Volumetric (helical/spiral) CT (VCT) of the airways. Journal of Thoracic Imaging 1997; 1211-28

5. Remy-Jardin M, Remy J, Artaud D, Fribourg M et al. Volume rendering of the tracheobronchial tree: Clinical evaluation of bronchographic images. Radiology 1998; 208:761-770

6. Ney DR, Kuhlman JE, Urban RH, Ren H, Hutchins GM, Fishman EK. Three-dimensional CT-Volumetric reconstruction and display of the bronchial tree. Invest Radiol 1990; 25:736-742

7. Mehta AC, Dasgupta A. Airway stents. Clinics in Chest Medicine; 1999; 20: (1): 139-151

8. Johnson PT, Fishman EK, Duckwall JR, Calhoun PS, Heath DG. Interactive thee dimensional volume rendering of spiral CT data: Current applications in the thorax. Radiographics 1998;18:165-187