COMPUTED TOMOGRAPHIC FINDINGS IN 35 DOGS WITH NASAL ASPERGILLOSIS

The purpose of this study was to describe the computed tomographic (CT) features of nasal aspergillosis in dogs. Initial (n = 35) and follow-up (n = 12) CT images were available from 35 dogs. The most commonly encountered CT findings were (1) moderate to severe cavitary destruction of the turbinates with presence of a variable amount of abnormal soft tissue in the nasal passages, (2) non-specific thickening of the mucosa adjacent to the inner surface of bones of the frontal sinus, maxillary recess and nasal cavity and, (3) thickened reactive bone. The findings were consistent with a disease initially affecting one nasal cavity then progressing into the ipsilateral frontal sinus, the contralateral nasal cavity and the contralateral frontal sinus. Two dogs with associated nasal foreign body had a more localized invasion of the nasal cavity. Attenuation values and contrast enhancement were not specific. With follow-up examinations, a reduction in the amount of abnormal soft tissue was observed in all dogs except one, but this reduction could not be quantified.     

COMPUTED TOMOGRAPHIC IMAGING OF NASAL DISEASE IN 100 DOGS

One hundred dogs with nasal disease were evaluated by computed tomography (CT). Certain findings were significantly correlated with neoplasia. These included patchy areas of increased density within soft tissue opacity, destruction of part or all of the ethmoid bones, abnormal soft tissue in the retrobulbar spaces, destruction of one or both the lateral maxillae, destruction of the nasal bone or rostral dorsal maxilla, or hyperostosis of the lateral maxilla. Combinations of these findings were even more significant for neoplasia. No single finding or combination of findings was absolutely definitive of neoplasia. The results of this study are useful in indicating that biopsy results should be questioned if one or more of the above factors is present and a diagnosis of inflammation is made.     

COMMUNICATIONS OF NORMAL NASAL AND PARANASAL CAVITIES IN COMPUTED TOMOGRAPHY OF HORSES

Heads from 15 male and female horses of various breeds and ages were examined with computed tomography under general anesthesia in dorsal recumbency. The main interest was the evaluation of the entire paranasal sinus system. Special attention was paid to the location, size, and shape of the conchomaxillary, nasomaxillary, frontomaxillary, and sphenopalatinal openings. The frontomaxillary opening was the largest aperture and was found to be age-dependent in size mostly. Orientation and shape of the openings were sagittal or horizontal; the narrowest of them was the slit-like nasomaxillary aperture. The thickness of the vascularized mucous membranes in the nasal cavity was obtained up to 6 mm; the surface was smooth and well defined. The paranasal lining was invisible in computed tomography (CT) images because of its thin layer. CT proved to be a very useful method for documentation of the paranasal openings and for evaluation of the state of the mucous membranes in the nasal and paranasal cavities.     

COMPARATIVE PERFORMANCE OF RADIOGRAPHY AND COMPUTED TOMOGRAPHY IN THE DIAGNOSIS OF MIDDLE EAR DISEASE IN 31 DOGS

The purpose of this study was to compare computed tomography (CT) and radiography for diagnosing the presence and severity of middle ear disease in dogs with a history of chronic otitis externa. Thirty-one dogs undergoing a total ear canal ablation and bulla osteotomy were studied. Three normal dogs served as controls. All dogs were examined using radiography and CT. Three radiologists independently evaluated imaging studies in random order. A visual analog scale method was used for scoring certainty and severity of middle ear disease. Surgical findings were recorded intra-operatively. Bulla lining samples were submitted for histopathologic evaluation and scored by a single pathologist who also used a visual analog scale system. Findings from both imaging modalities agreed more closely with surgical findings than with histopathologic findings. With either surgical or histopathologic findings as the gold standard, CT was more sensitive than and as specific as radiographs for predicting presence and severity of middle ear disease. Observer performance with CT was more consistent than the performance with radiographs in the detection of changes that occur with middle ear disease. Both radiography and CT were more accurate for predicting the severity of the disease than its presence. Findings indicate that CT is more accurate and reliable than radiography in diagnosing middle ear disease for dogs having concurrent otitis externa, but only when severity of disease is moderate or high. With low severity of disease, diagnostic certainty for both modalities becomes more variable.     

COMPARISON OF RADIOGRAPHY AND COMPUTED TOMOGRAPHY FOR THE DIAGNOSIS OF CANINE NASAL ASPERGILLOSIS

To compare the radiographic and computed tomographic (CT) findings and to evaluate the sensitivity of radiography and CT for diagnosis of nasal aspergillosis in dogs, the radiographic and CT studies of 48 dogs with chronic nasal disease were reviewed separately. The radiographic and CT findings were recorded, and a diagnosis was made. The results obtained in the dogs with nasal aspergillosis (n = 25) were used. Based on definite aspergillosis as diagnosis, CT had a sensitivity of 88% and radiography of 72%. Considering definite and probable aspergillosis as equivalent, CT had a sensitivity of 92% and radiography of 84%. The sensitivity was higher in dogs with lesions affecting the entire nasal cavity and frontal sinus on at least one side (n = 20) with a sensitivity of 100% for CT and 90-95% for radiography than in dogs with lesions restricted to the nasal cavities (n = 5) where CT had a sensitivity of 60-80% and radiography of 0-40%. CT was superior to radiography for evaluation of the nasal cavities (mucosal thickening along the nasal bones, surrounding bone hyperostosis/lysis), frontal sinuses (mucosal thickening along the frontal bone, fluid/soft tissue, frontal bone hyperostosis/lysis), and differentiation between a cavitated-like or a mass-like process. This study suggests that CT is more sensitive than radiography for diagnosis of nasal aspergillosis in the dog because of a better demonstration of some changes suggestive of nasal aspergillosis. A diagnosis of a nasal aspergillosis restricted to the nasal cavities or associated with an FB is challenging, even with the use of CT.     

PREVALENCE OF EAR DISEASE IN DOGS UNDERGOING MULTIDETECTOR THIN‐SLICE COMPUTED TOMOGRAPHY OF THE HEAD

Previous reports describing the prevalence of ear diseases in dogs have primarily been based on dogs presenting with clinical signs of disease. The prevalence of subclinical ear disease remains unknown. The purpose of this cross‐sectional retrospective study was to describe the prevalence of lesions consistent with middle and external ear disease in dogs presented for multidetector computed tomography (CT) of the head and/or cranial cervical spine at our hospital during the period of July 2011 and August 2013. For each included dog, data recorded were signalment, CT findings, diagnosis, and treatment. A total of 199 dogs met inclusion criteria. Nineteen dogs (9.5%) were referred for evaluation of suspected ear disease and 27 dogs (13.5%) had histories or physical examination findings consistent with otitis externa. A total of 163 dogs (81.9%) had CT lesions consistent with external ear disease (i.e. ear canal mineralization, external canal thickening, and/or narrowing of the external canal). Thirty‐nine dogs (19.5%) had CT lesions consistent with middle ear disease (i.e. soft tissue attenuating/fluid material in the tympanic bullae, bulla wall thickening or lysis, and/or periosteal proliferation of the temporal bone). Findings from this study indicated that the prevalence of external and middle ear disease in dogs could be higher than that previously reported.     

A COMPARISON OF RADIOGRAPHIC AND COMPUTED TOMOGRAPHIC FINDINGS IN 31 DOGS WITH MALIGNANT NASAL CAVITY TUMORS

Nasal cavity radiographs and CT images from 31 dogs with nasal cavity cancer were compared. All dogs had abnormal clinical signs relating to -nasal cancer and histologic confirmation of malignant nasal cavity neoplasia. No dog had cyto reductive surgery prior to imaging. All radiographic and CT examinations were abnormal. CT was more accurate than radiographs in identifying unilateral versus bilateral nasal cavity disease and tumor extension into adjacent structures such as the cranial cavity, hard palate, and pterygopala-tine fossa. The improved accuracy of CT in these respects was not of benefit in the confirmation of nasal cavity disease because radiographs were abnormal in every instance. However, CT may be useful for more accurate tumor staging, predicting possible treatment-related complications, and planning of surgery and radiation therapy. It was also determined that one dorsally located radiation therapy portal bounded laterally by the medial ocular canthi, as described in previous reports, would not have been adequate for encompassing all abnormal tissue in 28 of the 31 dogs evaluated.     

PULMONARY ANGIOGRAPHY WITH 64‐MULTIDETECTOR‐ROW COMPUTED TOMOGRAPHY IN NORMAL DOGS

Pulmonary angiography using 64‐multidetector‐row computed tomography (MDCT) was used to evaluate pulmonary artery anatomy, and determine the sensitivity of pulmonary artery segment visualization in four Beagle dogs using images reconstructed to 0.625 mm and retro‐reconstructed to 1.25 and 2.5 mm slice thickness. Morphologically, characteristic features included a focal narrowing in the right cranial pulmonary artery in all dogs, which should not be mistaken as stenosis. While the right cranial pulmonary artery divided into two equally sized branches that were tracked into the periphery of the lung lobe in all dogs, only a single left cranial (cranial portion) lobar artery was present. Compared with 1.25 and 2.5 mm retro‐reconstructions, 0.625 mm reconstructions allowed for detection of significantly (P≤0.05) more pulmonary artery segments and sharper depiction of vessel margins. Clinical applications such as prevalence and significance of diameter changes, and detection of pulmonary arterial thrombembolism on lobar and sublobar level, using pulmonary angiography with 64‐MDCT applying 0.625 mm reconstruction slice thickness remain to be established.     

RESULTS OF COMPUTED TOMOGRAPHY IN DOGS WITH SUSPECTED WOODEN FOREIGN BODIES

Detection of wooden foreign bodies in dogs can be challenging. A retrospective, cross‐sectional study was done to describe computed tomographic (CT) signs associated with wooden foreign bodies, and to estimate the accuracy of CT for detection of wooden foreign bodies. Patient records and CT images were reviewed for 72 dogs that had a history of suspected stick injury and CT of the affected body part, or possible wooden foreign object reported on CT, and had surgical exploration during the same period of hospitalization. Duration of clinical signs was acute in 48 (67%) dogs and chronic in 24 (33%). Wood was removed from 55 dogs, including a piece of a tree or shrub in 33 (60%) instances, kebab stick in 8 (15%), piece of bamboo garden cane in 2 (4%), cocktail stick in 2 (4%), thorn in 1 (2%), and unidentified wood in the remaining nine instances. Based on review of CT images with knowledge of the surgical findings, sensitivity of CT for wooden foreign bodies was 79% (95% CI 65%–89%), specificity 93% (78%–98%), positive likelihood ratio 11.5 (2.9–44.1), and negative likelihood ratio 0.23 (0.13–0.41). Wooden foreign bodies were predominantly rectangular or linear, with median length 48 mm (range 2–270 mm), median thickness 3 mm (range 1–22 mm), and median attenuation 111 HU (range ?344 to +640 HU). A CT finding of gas in soft tissues was significantly associated with acute cases, whereas suspected foreign material, cavitary lesions, fat stranding, and periosteal reaction on adjacent bones were associated with chronic cases.     

MYELOGRAPHY VS. COMPUTED TOMOGRAPHY IN THE EVALUATION OF ACUTE THORACOLUMBAR INTERVERTEBRAL DISK EXTRUSION IN CHONDRODYSTROPHIC DOGS

Myelography and/or computed tomography (CT) are commonly used to evaluate acute intervertebral disk extrusion in chondrodystrophic dogs. The purpose of this study was to evaluate the sensitivity of myelography and two different CT protocols in determining level and lateralization of acute thoracolumbar intervertebral disk extrusion in comparison with surgical findings, investigate interobserver variability, and determine examination duration times. Results of conventional CT, helical CT, and myelography were compared with surgical findings in 19 chondrodystrophic dogs with acute thoracolumbar intervertebral disk extrusion. Agreement among investigators was determined for different imaging modalities, and examination times were recorded. In the diagnosis of level of disk extrusion there was agreement of myelography, conventional CT, and helical CT with surgical findings in 94.7%, 100%, and 94.7% of dogs, respectively ( P =0.144). In the diagnosis of level and lateralization of disk extrusion there was agreement of myelography, conventional CT, and helical CT with surgical findings in 78.9%, 87.4%, and 85.3% of dogs, respectively ( P =0.328). Interobserver agreement was very good for all imaging modalities (myelography κ=0.87; conventional CT κ=0.86; helical CT κ=0.95). There were significant differences in median examination duration times between helical CT (4 min), conventional CT (8 min), and myelography (32 min) ( P <0.001). Both helical and conventional CT appear to be accurate for evaluation of acute thoracolumbar intervertebral disk extrusion in chondrodystrophic dogs and are faster to perform than myelography.     

FAILURE PATTERNS FOLLOWING COBALT IRRADIATION IN DOGS WITH NASAL CARCINOMA

The pattern of tumor recurrence was assessed in 24 dogs receiving cobalt radiation therapy for nasal carcinoma. Dogs were evaluated using nasal cavity computed tomography prior to treatment, and at 1, 3, 6 and 12 months after treatment, and at 6-month intervals thereafter if still alive. Dogs were treated with various combinations of total dose, and fraction size. Total doses were normalized to equivalent doses given in 3.0 Gy fractions. The extent of tumor regression or duration of tumor control were not dependent on absolute total dose, normalized total dose, or tumor type. The median duration of local control in all dogs was 312 days. Marked tumor regression was observed in 11 of the 24 dogs. Median duration of local control was significantly longer in dogs with marked tumor regression in comparison to dogs without tumor regression; 389 vs. 161 days respectively. When tumor recurrence was documented in dogs having tumor regression, the location of the recurrence was in the nasal cavity. No tumor recurred in a sinus or periorbital region, and only one geographic miss was detected. Tumor recurrence in the irradiated volume, including dogs with and without marked regression, was documented in 13 of the 24 dogs. The high local failure rate, coupled with the recurrence pattern in these dogs, suggests there may be an opportunity for improvement in local control through use of shrinking field techniques.     

COMPUTED TOMOGRAPHY AND BLOOD GAS ANALYSIS OF ANESTHETIZED BLOODHOUNDS WITH INDUCED PNEUMOTHORAX

Increasingly severe degrees of pneumothorax were produced in 6 adult anesthetized bloodhounds. Computed tomography (CT) of the thorax was performed on each dog to evaluate the effects of pneumo thorax on thoracic and on pulmonary cross-sectional area (TA and PA). Arterial PO₂ (PaO₂) and PCO₂ (PaCO₂), heart rate (HR), and mean arterial blood pressure (MAP) were determined and related to the severity of pneumothorax. Volumes of air equal to 1, 1.5 and 2 times functional residual capacity of the lung produced approximately 33%, 40%, and 50% reductions in pulmonary area respectively. These amounts of atelectasis correspond to a radiographically "moderate" degree of pneumothorax. As severity of pneumothorax increased, thoracic area consistently increased, PaO₂ consistently decreased, and PaCO₂ consistently increased, with all being statistically significant relationships (p<0.0001); but HR and MAP were variable and showed no statistical correlation to the degree of pneumothorax (p>0.2).     

CONTRAST-ENHANCED COMPUTED TOMOGRAPHY AS A PREOPERATIVE INDICATOR OF VASCULAR INVASION FROM ADRENAL MASSES IN DOGS

The purpose of this study was to determine if contrast-enhanced computed tomography (CT) accurately reflected vascular luminal invasion by adrenal masses in dogs. Medical records of 15 dogs with 17 adrenal masses that underwent preoperative CT imaging were reviewed. Presence or absence of vascular invasion was confirmed by surgery or necropsy/histology. Vascular invasion was identified correctly using contrast-enhanced CT in 11/12 dogs. The sensitivity and specificity of contrast-enhanced CT for vascular invasion compared with surgery or necropsy was 92% and 100%, respectively. The positive predictive value and negative predictive value of contrast-enhanced CT was 100% and 90%, respectively. The accuracy was 95%. In one dog, invasion of the phrenicoabdominal vein was not identified on CT images. Six of eight masses with vascular invasion where a histologic diagnosis was obtained were malignant. Four of four pheochromocytomas invaded adjacent vasculature. Contrast-enhanced CT provided accurate preoperative assessment of adrenal masses. Vascular invasion by adrenal masses in this study occurred by way of the lumen of the phrenicoabdominal vein with subsequent intraluminal extension into other veins, rather than by erosion through vessel walls.     

COMPUTED TOMOGRAPHIC FINDINGS OF DOGS WITH CUBITAL JOINT LAMENESS

The purpose of this study was to identify the incidence of abnormal computed tomography (CT) findings in dogs with lameness of one or both elbows. CT examinations of 102 elbows in 51 dogs were reviewed individually and by group consensus by 3 examiners and graded for various abnormalities. Frequently encountered abnormalities included abnormal shape and sclerosis of the medial coronoid process of the ulna, irregularity of the radial incisure of the ulna, and ulnar trochlear notch sclerosis. Other abnormalities included distinct/separate or fissure/in situ fragmentation of the medial coronoid; lucency of the radial incisure; ununited anconeal process; sclerosis, lucency, or flattening of the medial aspect of the humeral condyle; osteophyte formation; and joint incongruity. In conclusion, CT of the canine cubital joint is useful in identification of various lesions beyond simple fragmentation of the medial coronoid process of the ulna.     

EVALUATION OF THE GASTROINTESTINAL TRACT IN DOGS USING COMPUTED TOMOGRAPHY

Abdominal computed tomography (CT) studies of 19 dogs with no history or clinical signs of gastrointestinal disease, and two dogs with a histological diagnosis of gastrointestinal neoplasia were examined retrospectively. Gastrointestinal segments were evaluated subjectively for conspicuity, contrast enhancement, and wall layering after contrast medium administration. In dogs without gastrointestinal disease, there were 62.8% of gastrointestinal segments (serosa to serosa) and 77.7% of gastrointestinal walls (serosa to mucosa) visualized. Wall layering on postcontrast images was seen in 21.8% of gastrointestinal segments. There was significant association between gastrointestinal diameter and wall thickness. There was significant association between weight and gastrointestinal wall thickness in the following regions: gastric fundus, gastric body, gastric pylorus, gastric pyloric antrum, duodenal cranial flexure, jejunum and ascending colon, and between patient weight and gastrointestinal diameter in cranial duodenal flexure, descending duodenum, transverse duodenum, ascending duodenum, and jejunum. Measurements acquired from CT studies correlated well with previously published normal reference ranges for radiographic and ultrasonographic studies. Gastrointestinal neoplasia, diagnosed in two dogs, had a gastrointestinal wall thickness greater than the range of the dogs without gastrointestinal disease. Computed tomography offers identification of the gastrointestinal tract segments in dogs, allows for evaluation of gastrointestinal diameter and aids in investigation of gastrointestinal wall thickness.     

DETERMINATION OF GLOMERULAR FILTRATION RATE IN DOGS USING CONTRAST-ENHANCED COMPUTED TOMOGRAPHY

The purpose of this project was to establish a procedure and reference values for glomerular filtration rate (GFR) using contrast-enhanced computed tomography (CT) in eight healthy dogs. A single section of the kidney was scanned sequentially after bolus injection (3 ml/s) of iohexol (300 mg/kg). Time-attenuation curves were constructed and the GFR per volume of kidney was calculated using Patlak graphical analysis software. The GFR was then converted from contrast clearance per unit volume (ml/min/ml) to contrast clearance per body weight (ml/min/kg). Individual kidney and global GFR were calculated using both CT and nuclear scintigraphy. Global GFR for each dog was also determined by plasma iohexol clearance. Contrast-enhanced CT underestimated the global GFR compared with the other two methods. The average global GFR was 2.57 +/- 0.33 ml/ min/kg using functional CT and 4.06 +/- 0.37 ml/min/kg using plasma iohexol clearance. There was significant (P < 0.05) interobserver variability of CT GFR of the right kidney and total GFR. There was decreased interobserver variability for the left kidney. There was no difference in the intraobserver variability for CT-determined individual kidney and global GFR. There was no difference between the motion corrected and nonmotion corrected values for individual and global CT GFR. Nuclear scintigraphy produced a slightly higher coefficient of variation than contrast-enhanced CT, 2.9% and 1.0%, respectively. It is hypothesized that altered renal blood flow, hematocrit of the small vessels, and nephrotoxicity play a role in the underestimation of GFR by contrast-enhanced CT.     

COMPUTED TOMOGRAPHIC IMAGING OF INFILTRATIVE LIPOMA IN 22 DOGS

Twenty two dogs with an infiltrative lipoma had computed tomographic (CT) images acquired to evaluate the extent of local disease. Ten dogs had undergone at least one cytoreductive surgical procedure (range = 1-3; median = 2) prior to imaging. Twenty dogs had measurable disease on CT images; 2 dogs had diffuse disease at a previous surgical site that could not be measured. Tumor volume (n = 20) ranged from 20 to 5,632 cm3 (median = 345 cm3; mean = 996 cm3). None of the dogs had evidence of bone involvement on the CT images; 2 of the 22 dogs had tumors that did not come into direct contact with osseous structures. All dogs with measurable disease had evidence of a fat opacity mass with variable degrees of muscle infiltration. Eleven of 22 dogs were given intravenous contrast medium prior to image acquisition and there was not evidence of enhancement of the infiltrative lipoma in any dog. Based on CT images, tumors were classified as well-defined in 9 dogs, moderately well-defined in 4, not well-defined in 3 and a mix of well-defined and not well-defined in 6 dogs. Tumors tended to be less well-defined in regions where the infiltrative lipoma interdigitated with normal body fat. It appears CT imaging allows adequate discrimination of tumor with the caveat that differentiation of normal fat from infiltrative lipoma can be problematic.     

HEPATIC VOLUME ESTIMATION USING QUANTITATIVE COMPUTED TOMOGRAPHY IN DOGS WITH PORTOSYSTEMIC SHUNTS

The purpose of this study was to use quantitative computed tomography (CT) to estimate liver volume in dogs with a portosystemic shunt and to compare the liver volume in normal dogs to dogs with a shunt. Twenty-one dogs with a portosystemic shunt underwent contrast-enhanced abdominal CT for shunt characterization and preoperative planning. Six dogs without clinical signs relating to liver disease were used as a control group. In addition, liver volume was compared before and 2-4 months after surgical shunt attenuation in three dogs. All studies followed established clinical imaging protocols. Liver margins were defined on each image using an operator-defined region of interest and hepatic volume renderings were produced from which the liver volume was quantitatively estimated. There was a statistically significant association between liver volume and body weight in control and shunt dogs (r = 0.909 and 0.899, respectively, P < 0.01). Liver volume normalized to body weight was 15.5 +/- 5.2 cm3/kg in affected dogs and 24.5 +/- 5.6 cm3/kg in control dogs. Based on postligation CT studies in three affected dogs, liver volume increased by 43%, 51%, and 62%. Hepatic volume estimation may be a clinically useful parameter in the initial and postsurgical evaluation of dogs with portosystemic shunts.     

COMPUTED TOMOGRAPHY VARIATIONS IN MORPHOLOGY OF THE CANINE ATLAS IN DOGS WITH AND WITHOUT ATLANTOAXIAL SUBLUXATION

Computed tomography images of 120 dogs were reviewed to characterize variations in atlas morphology, and to identify breed‐specific morphologic features. The neural arch of the atlas was thicker in large dogs and male dogs than in small dogs, having a layer of trabecular bone between the inner and outer layers of compact bone. The transverse processes of the atlas were relatively longer in large dogs than in small dogs. Twelve (10%) dogs had incomplete ossification of the atlas. Incomplete ossification of the atlas was associated with gun dogs. Eight dogs had atlantoaxial subluxation. All eight dogs with atlantoaxial subluxation had cervical signs, whereas none of the seven dogs with incomplete ossification of the atlas unaccompanied by atlantoaxial subluxation had clinical signs referable to that area. Of five dogs with both atlantoaxial subluxation and incomplete ossification of the atlas, four had osseous defects affecting both the intercentrum and neural arch, and one had only an osseous defect affecting the neural arch. There was a strong association between incomplete ossification of the atlas and atlantoaxial subluxation (odds ratio 35.0, 95% CI 7.0–175, P=0.00002), which supports the hypothesis that incomplete ossification of the atlas predisposes dogs to atlantoaxial subluxation.     

DIAGNOSIS OF ARTERIOPORTAL FISTULAE IN FOUR DOGS USING COMPUTED TOMOGRAPHIC ANGIOGRAPHY

Arterioportal fistulae are rare congenital anomalies of the hepatic vasculature. Diagnosis is conventionally made by selective angiography or ultrasonography. This report describes use of a dual-phase computed tomographic (CT) angiographic technique to diagnose arterioportal fistulae in four dogs. Advantages of this method include a noninvasive peripheral injection of contrast medium, ability to diagnose multiple acquired extrahepatic shunts, and observation of hemodynamic changes such as hepatofugal blood flow and reduced circulation to the caudal abdomen. The hepatic vasculature including arteries, veins, and portal veins can be completely evaluated. Dual-phase CT angiography is a safe and minimally invasive method of diagnosing arterioportal fistulae in dogs.