High Resolution Computed Tomography of the Thorax
World Small Animal Veterinary Association World Congress Proceedings, 2007
John S. Mattoon, DVM, DACVR
Associate Professor, Radiology, College of Veterinary Medicine, Washington State University
Pullman, WA, USA

In veterinary medicine, survey thoracic radiography is the most frequently employed imaging modality to evaluate suspected lung disease. Unfortunately, radiographs often reveal no abnormalities, even in the presence of severe clinical signs, or non-specific findings. In human medicine CT is routinely used to identify abnormalities, which are not apparent on survey radiographs, to provide a more specific diagnosis for vague, diffuse radiographic abnormalities, or to characterize a lesion as benign or malignant. The introduction of helical and HRCT of the canine lung for the clinical usage has provided an additional, non-invasive, diagnostic tool which has higher sensitivity and specificity than traditional radiography, and can also provide quantitative tissue density.

CT is very diagnostic for other forms of pulmonary disease as well. Pulmonary masses can be precisely located and thus CT can aid in surgical planning. It is possible to differentiate pulmonary from pleural or esophageal origin masses. Lung lobe torsions, often difficult to diagnose on radiographs are amenable to CT diagnosis as the bronchus can easily be seen in an abnormal orientation. CT is very sensitive in evaluation of diffuse lung and airway disease in people and will has similar advantages over radiography in veterinary medicine.

In addition, CT of non-pulmonary thoracic anatomy and diseases has been described for the dog and cat. The majority of intrathoracic structures can be identified reliably, rendering thoracic CT a very valuable diagnostic tool. Evaluation of mediastinal disease is a hallmark contribution of CT, including cranial mediastinal masses, differentiating esophageal pathology, and assessment of hilar lymph nodes. The presence of pleural fluid diagnosed on thoracic radiographs often complicates diagnosis of intrathoracic disease. Pleural fluid may mask mass lesions of mediastinal, pleural or pulmonary origin. CT can be useful in these circumstances to help determine the origin of pleural fluid and to assess the extent of disease. The use of intravenous iodinated contrast material greatly aids in location of important vascular and other anatomic landmarks within the thorax. CT has also been used to identify the etiology of pneumothorax. CT is used regularly for assessment of extrathoracic diseases. Soft tissue masses of the thoracic wall can be evaluated for involvement of the ribs and vertebral column. A common example of this is presurgical planning for excision of CT vaccine-induced fibrosarcomas in cats. CT is also the most sensitive modality to evaluate the extent of rib lesions such as chondrosarcomas. CT is used for interventional procedures (needle aspirates or biopsies) when ultrasound cannot be used because of interspersed air.

Pulmonary CT

Conventional pulmonary/thoracic CT is performed with thick collimation (or slice thickness), typically 8-10 mm, using an incremental technique; that is, the patient on the CT table is advanced in the gantry for a length equal to the collimation while one cross-sectional image is acquired. The table then is stopped; it is advanced again when a second image is acquired, and so forth until the whole scan is completed.

Although conventional, incremental CT is used to evaluate pulmonary structures, helical and high-resolution thin-section computed tomography (HRCT) represent the most accurate noninvasive tools currently available for evaluating lung structures in people. Helical CT involves continuous acquisition of data while the patient is advanced at a constant rate through the CT gantry resulting in a contiguous set of images obtained without interscan delay, obviating potential misregistration. Helical CT offers two important advantages:

1.  The time of examination is greatly decreased: the entire thorax can be acquired in less than one minute.

2.  Multiplanar and three-dimensional reconstruction is improved: the step-like discontinuities that are a common to conventional CT are greatly reduced. Multidector CT evenbetters helical CT in this regard.

In addition to the development of helical CT, improvements in CT scanners have allowed HRCT (thin slice, 1-2 mm) to evolve during the last two decades. Currently HRCT makes it possible to acquire in vivo images with spatial resolution comparable to direct visualization of pathologic specimens. The principle effect of narrow collimation as used in HRCT is a significant reduction of volume averaging within the plane of section, and therefore increased spatial resolution when compared to a 10-mm collimated scan.

The reconstruction algorithm used to form CT images from raw data can be configured to provide either maximum contrast or maximum spatial resolution. In the high-contrast environment of the lungs, the spatial reconstruction is optimized for the highest spatial resolution. Visualization of the enhanced spatial resolution provided by thin collimation scans and the high-spatial-frequency algorithm requires appropriate configuration of the image display system. Typically, the highest resolution matrix size available is 512x512 lines, which yields a pixel size of 0.78 mm when using a field of view (FOV) of 40 cm. For optimal matching of the image display to the attainable spatial resolution, there should be two pixels for the smallest object resolved. Current scanners combined with the high-spatial-frequency algorithm can provide scan data with spatial resolution of 0.5 mm. The pixel size therefore needs to be reduced by targeting the image, or reducing the FOV. This procedure allows truly decreased pixel size and differs from magnification, which does not improve spatial resolution, since the pixels are merely enlarged.

One drawback of image targeting and use of a high-spatial-frequency algorithm is that the resultant image will show increased noise, or quantum mottle, as a result of reduced smoothing of the image data. Noise is reduced by increasing the kVp and mA settings (increasing patient dose and x-ray tube load).

Detailed studies describing normal human lung anatomy and pathologic findings,as well as quantitative and densitometric studies using helical and HRCT are available in the medical literature. Dogs have been used in experimental studies as models to investigate pulmonary disease processes, such as acute pulmonary vascular occlusion, pulmonary infarction, pneumothorax and pulmonary edema, in an effort to better understand human diseases. The canine lung has also been used as a model to evaluate the accuracy of helical CT in the detection of pulmonary metastases from osteosarcoma. Differences in subgross anatomy between humans and dogs exist, however, making direct comparison of disease processes difficult at best.

Until recently, HRCT had been used mainly in experimental studies in veterinary medicine, but several recent reports have described HRCT of canine lung and protocols for helical CT and HRCT have now been developed.

The mean HU value of normal canine lung has been reported to be between approximately -850 to-700 HU. Pulmonary vasculature, airways and interstitium can be readily assessed. Interestingly, caudal lobar pulmonary veins have been shown to be significantly larger than corresponding arteries using CT. A novel classification scheme for HRCT of canine lung was developed, dividing the lung into a 1 mm pleural region (Zone 1), a subpleural region encompassing the outer 5% of lung lobe area (Zone 2), and the peribronchovascular region (Zone 3). Types of pulmonary abnormalities were then classified into four groups: linear and reticular opacities, nodules, increased lung density, and decreased lung density.

Lung is viewed in a "lung window", with window levels set at approximately -500 to-700 HU, window widths of 1200-1500 HU. This setting maximizes pulmonary parenchymal visualization but renders the mediastinal structures non-viewable. For mediastinal viewing (or soft tissue masses of pulmonary origin), a "mediastinal Window" is chosen with window levels of -50 to 40 HU and window widths of 400-500 HU.

Anesthetic Considerations

Anesthetic considerations are important during CT evaluation of the thorax. Foremost, respiratory movement must be managed. Secondly, if the lung is the primary organ of interest, it must be properly inflated. Under ideal circumstances, the patient is anesthetized and maintained in sternal recumbency. This assures that the lungs will not become atelectic due to recumbency (atelectic lung cannot be differentiated from pulmonary pathology). During the scan the lungs are inflated and maintained at 15-20 cm H2O by positive pressure ventilation. This ensures good aeration and eliminates respiratory motion artifact. Breath holds of 30 seconds can be safely maintained. Most medium-sized patients can be imaged in 2-3 breath holds. A feline thorax can often be scanned in one breath hold when a helical CT scanner is used. Sedation can be used in some patients that are not breathing rapidly and a quality scan still obtained, although some slices may be compromised by motion. In critical cases in which anesthesia is not advisable thoracic CT scans may still be made and provide diagnostic information.


1.  De Rycke LM, Gielen IM, Simoens PJ, van Bree H. computed tomography and cross-sectional anatomy of the thorax in clinically normal dogs. Am J Vet Res. 200;66:512-524.

2.  Henninger W. Use of computed tomography in the diseased feline thorax. J Small Anim Pract. 2003;44:56-64.

3.  Johnson VS, Schwarz T, Sullivan M. High resolution computed tomography (HRCT) of the normal canine lung. In: Proceedings of the Meeting of the American College of Veterinary Radiology, Montreal. 2004, page 58.

4.  Johnson VS, Ramsey IK, Thompson H, Cave TA, Barr FJ, Rudorf H, Williams A, Sullivan M. thoracic high-resolution computed tomography in the diagnosis of metastatic carcinoma. J Small Anim Pract. 2004;45:134-143.

5.  Johnson VS, Corcoran BM, Wotton PR, Schwarz T, Sullivan M. Thoracic high-resolution computed tomographic findings in dogs with canine idiopathic pulmonary fibrosis. J Small Anim Pract. 2005;46:381-388.

6.  Morandi F, Mattoon JS, Lakritz J, Turk JR, Wisner ER. Correlation of helical and incremental high-resolution thin-section computed tomographic imaging with histomorphometric quantitative evaluation of lungs in dogs. Am J Vet Res, 2003;64:935-944.

7.  Morandi F, Mattoon JS, Lakritz J, Turk JR, Jaeger JQ, Wisner ER. Quantitative evaluation of an acute inflammatory response of the canine lung: correlation between helical and incremental high-resolution thin-section pulmonary computed tomography and histomorphometry. Am J Vet Res 2004;65:1114-1123.

8.  Riverio MA, Ramirez JA, Vazquez JM, Gil F, Ramirez G, Arencibia A. Normal anatomical imaging of he thorax in three dogs: computed tomography and macroscopic cross-sections with vascular injections. Anatomica histologia Embryoligia. 2003;34:215-219.

9.  Samii VF, Biller DS, Koblik PD. Normal cross-sectional anatomy of the feline thorax and abdomen: comparison of computed tomography and cadaver anatomy. Vet Radiol Ultrasound.1998;39:504-511.

Speaker Information
(click the speaker's name to view other papers and abstracts submitted by this speaker)

John S. Mattoon, DVM, DACVR
Washington State University