Chest:Volume Rendering of the Pediatric Airway- A Problem Solving Tool

INTRODUCTION.

Background.

Despite the demise of tracheobronchography there remains a demand for the types of images it produced, that is, comprehensive, whole lung images of the central and peripheral airways in anterior and oblique projections. Conventional axial CT and magnetic resonance imaging have largely filled the demand for non-invasive imaging to compliment direct visualization by tracheobronchoscopy (1). However there has remained a compelling demand for something more than the two-dimensional images we routinely produce and interpret.

The journey through three-dimensions has seen landmarks of shaded surfaced display, maximum intensity projection and volume rendering. Recent advances in slip-ring technology and multi-slice technology together with similar advances in three-dimensional volume rendering are setting a new standard. We employed these techniques for three pediatric patients recently referred for airway imaging.

Objective.



To describe advances in computed tomography (CT) and three-dimensional volume rendering as problem-solving tools applied to the pediatric patient with airway disease.

 CONCLUSION.

We have reached a benchmark in pediatric airway imaging where the combination of volume data acquisition by CT and three-dimensional volume rendering can now readily provide, high quality images when clinical dilemmas demand a comprehensive non-invasive study of the tracheo-bronchial tree.

MATERIALS AND METHODS.


Three pediatric patients were referred for CT scanning with three dimensional airway imaging to resolve clinical questions. There were two girls aged 15 months and 13 years and a boy aged 4 years old. Either a Siemens Plus 4 single detector spiral scanner or a Siemens Plus 4 Volume Zoom (Siemens Medical Systems, Iselin, NJ) were used. For the spiral CT slice thickness was 3 mm with a pitch of one and for the MDCT there was a slice thickness of 1.25 mm with a pitch of 6. In two patients intravenous Omnipaque 350 (Nycomed Amersham, Princeton New Jersey) was administered with a dose of 0.5cc /kg at a rate of 2cc per second. Standard axial film interpretation was performed. Data acquisition sets were immediately transferred to a prototype Siemens 3D Virtuoso (Siemens Medical Systems, Iselin, NJ) workstation and volume renderings were also performed.

RESULTS.


Our first patient was a ventilator dependant 15 month old female with spondylothoracic dystrophy with an abnormal chest wall, lung hypoplasia and tracheo-bronchomalacia. A bronchoscopy revealed significant narrowing of the trachea on the left just above the carina. There was bulging of the tracheal wall and there was concern there might be an extrinsic mass or vascular ring. She was referred to evaluate the length of this narrowing and to image for a possible extrinsic mass. Discrete, severe narrowing of the origin of the left main stem bronchus, not seen on the axial images, was noted (Fig1). No extrinsic mass was present. A second 4 year old patient was referred with bronchopulmonary dysplasia and pulmonary hypertension and was status post a D-transposition, Jatene arterial switch repair. The patient had multiple episodes of hemoptysis and though bronchoscopy had shown severe narrowing of the left main stem bronchus. Imaging showed the aortic arch was focally compressing the left mainstem bronchus and was occupying much of the space normally filled by pulmonary artery. The volume rendering showed the airway to be severely narrow, in a focal segment related to the arch (Fig 2A,B).

Our final patient was a 13-year-old referred for pulmonary embolism evaluation. Pulmonary embolism was excluded however a lucent right upper lobe with a paucity of vessels was noted and subsequent three-dimensional imaging defined an unsuspected right upper lobe aberrant bronchus arising directly from the right mainstem bronchus(Fig 3A,B).

DISCUSSION.


When presented with a pediatric patient for evaluation of possible airway pathology the following questions should come to mind; (i) is the condition congenital or acquired? (ii) What is the location, degree and the extent in the segment affected and what is the whole lung distribution if in multiple airways? (iii) Is there an associated vascular anomaly? (iv) Is the luminal narrowing or dilatation fixed or dynamic? (v) What information is required if surgical intervention is planned? The pediatric patient with airway pathology represents a distinct diagnostic challenge, since these patients cannot cooperate with pulmonary function testing and may be sufficiently compromised from a respiratory standpoint that bronchoscopy may not be possible. Furthermore lesions involving more peripheral airways may not be definable by bronchoscopy. Lumen compromise arises either intrinsically (foreign body, mass or mucosal thickening) or extrinsically (tumor, vascular). Obstructing lesions can be dynamic such as tracheomalacia or fixed such as a stenosis secondary to scarring, complete tracheal rings or tumors.

Developmental abnormalities may result in supernumerary or ectopic bronchi with the most frequent being the tracheal bronchus seen in 1% of the population. Other anomalies include bridging bronchus, cardiac bronchus and esophageal bronchus. These anomalies have been implicated in recurrent infection and air trapping. Conventional spiral CT can pick up these branches but 3D CT is ideal as it is easier to differentiate the branches and volume rendering provides a roadmap. of what the surgeon of bronchoscopist will encounter.

The most common cause of abnormal airway dilatation in childhood is probably bronchiectasis usually related to cystic fibrosis. Scoring systems based on conventional chest radiographs and/or CT have been developed to monitor responses to developing therapies. These technologies are limited but the ability to produce whole lung images of the entire tracheobronchial tree in isolation has clear potential for modeling both the disease and response to therapy.

An ideal imaging test to evaluate conditions described above should have the following capabilities; (a) can evaluate the central (endoscopically visible) and the peripheral tracheobronchial tree. (b) Be able to simultaneously provide information on mediastinum, lung parenchyma, thoracic cage and functional information if required (c) acquire the information fast, limiting radiation dose, contrast load and motion or respiratory artifact (d). If generating three dimensional images, be able to do so in a timely fashion with high fidelity using all the data obtained. (e) In addition the results of these images should be presented in a form that provides knowledge not obtainable by any other means and provide a perspective that compliments their surgical or bronchoscopic approach. (f) The technique should allow rotation in different views and permit removal of unwanted data without laborious region of interest drawings.

Spiral CT and volume rendering has all these attributes. The axial spiral imaging and volume rendering combination allow a complete management algorithm to be pursued with a single test to assess the child with stridor, chest wall abnormalities, vascular anomalies etc. Conventional axial CT has largely fulfilled the imaging role of the airway that was left vacant with the demise of tracheobronchography. Non-invasive imaging is particularly important in the pediatric age group where airways may be too small to allow safe scope passage and where doing so may cause inflammation in already small airways. There has remained a demand for whole lung bronchography-like images, which has fueled developments in CT and MRI with multiplanar reconstructions and three-dimensional renderings (2,3,4,5). The complex continuous tree-like structure that is the bronchial tree, however, defies definition by any single plane and the extrapolation of airway information is limited by radiologist and non-radiologist’s abilities to conceptualize this 2D information as a 3D structure. It makes sense that we endeavor to produce imaging representations that simulate the perspective the surgeon or the bronchoscopist will encounter. The airway images are produced from the same data set as the original lung images so there is no increased radiation dose. If however dynamic imaging is contemplated the patient will be scanned in inspiration and expiration with the expected increase in exposure. Multidetector CT with an adaptive array design maximizes dose efficiency by minimizing the detectors used.

By CT we may define if indeed the problem is localized to the airway or not. It may be clarified if it is intraluminal, epithelial based, intramural or extrinsic. Its location, extent, degree and distribution may be charaterized as well as the assessment of any associated anomalies of mediastinum, lung parenchyma, pleura and chest wall.

We can now not only acquire the data as a volume with slip-ring technology but with advances in multidetector CT and volume rendering technology we may do so very quickly, with high spatial resolution and use this complete data set to reconstruct our images. The faster scanning (acquisition times less than 5 seconds) with subsecond rotation times and multiple detectors has significantly decreased respiratory and motion artifact in children, lessened thus the need for sedation and allowed reduction in contrast volumes administered. Frush et al. (1) report that 97% of children under 6 years of age were successfully imaged without sedation using multislice technology. One does not have to compromise between z-axis coverage and high-resolution images and a 24-cm chest may be covered in 8 seconds with a pitch of 6. Volume rendering techniques assign a continuous range of values to a voxel so that the percentage of each tissue present is represented in the final image and spatial relationships are maintained (6). The trapezoidal transfer function creates a histogram of the displayed houndsfield units enabling manipulation of window width, level, opacity and brightness so particular pertinent combinations of tissues may be projected. Clip plane editing and interactive real time rotation of the image allow complex branching structures to be followed (7). CT, which once rivaled MRI for the airway, now surpasses it in many instances with shorter acquisition times, better lung parenchyma imaging, better post-stent imaging and superb three-dimensional imaging of airway and angiography. With a baseline established one also has the means for easy non-invasive follow-up.



REFERENCES.



Frush DP, Donnelly LF, Chotas HG (2000). Contemporary pediatric thoracic imaging AJR 175:841-851



Gustafson LM, Liu JH, Link DT, Strife JL, Cotton RT (2000). Spiral CT versus MRI in neonatal airway evaluation. Int J Pediatr Otorhinolaryngol. 52(2):197-201



Nicotra JJ, Mahboubi S, Kramer SS (1997). Three-dimensional magnetic resonance imaging of the pediatric airway. Int J Pediatr Otorhinolaryngol. 41(3):299-305



Donnelly KJ, Bank ER, Parks J, Gussack et al. (1994) Three-dimensional imaging evaluation of the pediatric tracheobronchial tree. Laryngoscope. 104(12):1425-30



Katz M, Konen E, Rozenman J, Szeinberg A, Itchak Y (1995). Spiral CT and 3D image reconstruction of vascular rings and associated tracheobronchial anomalies. J Comput Assist Tomogr 19(4):564-8



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



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