Presumed mediastinal cysts have a low prevalence in canine thoracic computed tomographic studies.

Mediastinal cysts (MCs) are rare lesions that can arise from embryonic remnants of various mediastinal structures. MCs usually are incidental findings in dogs and cats; however, they can reach a mass-like appearance. The description of MCs on CT in dogs is limited. This retrospective, single-center, descriptive, prevalence study aimed to determine the prevalence of presumed mediastinal cysts (PMCs) in dogs and assess their CT characteristics. Dogs that underwent a thoracic CT scan from January 2019 to August 2021 were included. CT images were evaluated for the presence of PMCs by two diagnostic imaging interns, two last year diagnostic imaging residents, and a board-certified veterinary radiologist. Number, location, margins, shape, volume, size, mass effect, and attenuation values of PMCs were assessed. A total of 866 CT scans were reviewed, and 49 PMCs were identified. The prevalence of PMCs in dogs was 5.66%. English Bulldog and mixed-breed dogs were subjectively overrepresented; however, the possibility of population bias could not be excluded. PMCs were subjectively more frequently observed in male dogs. The PMCs were predominantly round, small, solitary fluid-filled findings localized in the cranioventral mediastinum, with well-defined margins, homogeneous attenuation, and no contrast enhancement. The median attenuation value was 6.32 HU (range: -20.16 to 23.45 HU) precontrast and 7.58 HU (range: -2.45 to 20.79 HU) postcontrast, and the median volume was 1.19 cm3 (range: 0.02-45.32 cm3). Although the prevalence of PMCs was low in our sample population, findings supported prioritizing a differential diagnosis of incidental PMC for dogs with the above imaging characteristics.

Delayed non-selective contrast-enhanced CT can reveal spontaneous contrast enhancement of the lymphatic system in cats.

The authors incidentally observed that, on delayed postcontrast CT studies, feline patients could show increased attenuation of the lymphatic system. The aim of the present study was to evaluate if the lymphatic system of feline patients undergoing intravenous administration of contrast medium might consistently enhance on delayed postcontrast CT studies. Feline patients that underwent CT examination for variable diagnostic purposes were included in this multicentric observational descriptive study. A 10-min delayed postcontrast whole-body CT series was obtained for all cats enrolled and the following anatomic structures were systematically assessed: mesenteric lymphatic vessels, hepatic lymphatic vessels, cisterna chyli, thoracic duct, and anastomosis of the thoracic duct with the systemic venous system. A total of 47 cats were included in the study. The mesenteric lymphatic vessels showed enhancement in the selected series in 39 of 47 (83%) patients and the hepatic lymphatic vessels in 38 of 47 (81%) patients. The cisterna chyli, thoracic duct, and the point of anastomosis of the thoracic duct with the systemic venous circulation were enhanced in 43 (91%), 39 (83%), and 31 of 47 (66%) cats, respectively. This study confirms the initial observation. The mesenteric and hepatic lymphatic system, the cisterna chyli, the thoracic duct, and its anastomosis with the systemic venous circulation of feline patients undergoing intravenous administration of iodinated contrast medium can show spontaneous contrast enhancement in non-selective 10-min delayed contrast-enhanced CT series.

Magnetic resonance imaging features of presumed intracerebral hemorrhages in dogs show a significant association between apparent diffusion coefficient peripheral layer and estimated bleeding age and commonly exhibit a T2 blackout effect.

In humans, intracranial hematomas commonly exhibit a T2 blackout effect (BOE) with diffusion-weighted imaging (DWI) and apparent diffusion coefficient (ADC) maps. Published descriptions of comparable findings in dogs with intracerebral hemorrhages are lacking. Aims of this retrospective observational study were to describe the appearance of presumed intracerebral hemorrhages (PICHs) in DWI regardless of the underlying etiology and presence of T2 blackout effect (T2 BOE) in dogs and to test potential associations with DWI and ADC findings versus the estimated age of the hemorrhage. Magnetic resonance imaging studies of dogs with presumed PICHs based on the combined evaluation of T2*W gradient echo and/or susceptibility-weighted imaging, T1W and T2W images were enrolled. The age of the hemorrhage was estimated according to published criteria based on T1W and T2W sequences. The association between the age of the hemorrhage and the appearance of each finding (stratified or mixed), their signal intensities in DWI/ADC and the presence of the T2 BOE, was evaluated. A total of 35 PICHs were included: 13 of them were stratified and 22 had a mixed aspect. Only the ADC appearance of the peripheral layer in stratified PICHs was significantly associated with estimated age of the hemorrhage (p = .033), being hypointense in all hyperacute cases and hypo-/isointense in acute cases. The T2 BOE was present in 29/35 PICHs. The DWI sequences showed limited utility to date PICHs in this study population. As in humans, the T2 BOE was commonly seen in DWI and ADC maps of dogs with PICHs.

Susceptibility artifact morphology is more conspicuous on susceptibility-weighted imaging compared to T2* gradient echo sequences in the brains of dogs and cats with suspected intracranial disease.

Susceptibility-weighted imaging (SWI) has been found to be more reliable in the detection of vessels and blood products than T2*-weighted gradient echo (GE) in several human brain diseases. In veterinary medicine, published information on the diagnostic usefulness of SWI is lacking. The aim of this retrospective observational study was to investigate the value of SWI compared to T2*-weighted GE images in a population of dogs and cats with presumed, MRI-based diagnoses grouped as neoplastic (27), cerebrovascular (14), inflammatory (14), head trauma (5), other pathologies (4), or that were normal (36). Areas of signal void (ASV) were assessed based on shape, distribution, number, and conspicuity. Presence of ASV was found in 31 T2*-weighted GE and 40 SWI sequences; the conspicuity of lesions increased in 92.5% of cases with SWI. A 44.7% increase in the number of cerebral microbleeds (CMBs) was identified within the population using SWI (110) compared to T2*-weighted GE (76). Linear ASV presumed to be abnormal vascular structures, as are reported in humans, were identified in 12 T2*-weighted GE and 19 SWI sequences. In presumed brain tumors, abnormal vascular structures were detected in 11 of 27 (40.7%) cases on T2*-weighted GE and in 16 of 27 (59.3%) cases on SWI, likely representing tumor neovascularization; amorphous ASV interpreted as presumed hemorrhages on T2*-weighted GE were diagnosed as vessels on SWI in five of 27 (18.5%) cases. Since SWI shows ASV more conspicuously than T2*-weighted GE, the authors advocate the use of SWI in veterinary patients.

MRI and CT features of a disc herniation through a fractured lumbar vertebra causing lumbar spinal cord compression in a dog.

A 5-year-old intact male Saint Bernard was presented for paraparesis following a fall down the stairs. Magnetic resonance imaging revealed a chronic fragmentation of the dorsocaudal aspect of the second lumbar vertebral body (L2). Material migrated from the L2 to L3 intervertebral space through the fractured caudal vertebral endplate into the vertebral body and reached the vertebral canal, resulting in a severe extradural compressive myelopathy. Computed tomography confirmed the vertebral fragmentation. Histology of surgically removed compressive material revealed fibrocartilaginous material. The uncommon route of the extruded disc material through a fractured vertebral body makes this case of traumatic disc extrusion unusual.

Ultrasonographic features of intestinal lipogranulomatous lymphangitis in 10 dogs.

Intestinal lipogranulomatous lymphangitis (ILL) is a granulomatous inflammation of the lymphatic vessels of the intestinal wall and mesentery characterized by lipogranulomas. The purpose of this retrospective, multi-center, case series study is to report the ultrasonographic features of canine ILL. Ten dogs with a histologically confirmed ILL undergoing preoperative abdominal ultrasound were retrospectively included. Additional CT was available in two cases. Lesion distribution was focal in eight dogs and multifocal in two. All dogs presented with intestinal wall thickening and two had a concomitant mesenteric mass adjacent to the intestinal lesion. All lesions were in the small intestine. Ultrasonographic features were altered wall layering with predominantly muscular and to a lesser extent submucosal layer thickening. Other findings included hyperechoic nodular tissue within the muscular, serosa/subserosal, and mucosal layers, hyperechoic perilesional mesentery, enlarged submucosal blood/lymphatic vessels, mild peritoneal effusion, intestinal corrugation, and mild lymphadenomegaly. The two intestinal to mesenteric masses presented heterogeneous echostructure, predominantly hyperechoic with multiple hypo/anechoic cavitations filled with mixed fluid and fat attenuation content on CT. Histopathological findings included lymphangiectasia, granulomatous inflammation, and structured lipogranulomas affecting mainly submucosa, muscularis, and serosa. The intestinal to mesenteric cavitary masses revealed severe granulomatous peritonitis with steatonecrosis. In conclusion, ILL should be considered as a differential diagnosis for dogs with this combination of ultrasonographic features.

MRI ischemic and hemorrhagic lesions in arterial and venous territories characterize central nervous system intravascular lymphoma in dogs.

Intravascular lymphoma (IVL) is characterized by the proliferation of large malignant lymphocytes within the lumen of blood vessels. This retrospective, multi-center, case series study aimed to describe the MRI features of confirmed central nervous system IVL in dogs and compare them with histopathological findings. Medical record databases from seven veterinary centers were searched for cases of histologically confirmed IVL. Dogs were included if an MRI was performed. The MRI studies and histopathology samples were reviewed to compare the MRI changes with the histopathological findings. Twelve dogs met the inclusion criteria (12 brains and three spinal cords). Imaging of the brains revealed multifocal T2-weighted/FLAIR hyperintense and T1-weighted iso-hypointense lesions, with variable contrast enhancement; areas of abnormal diffusion both in arterial and venous territories in diffusion-weighted imaging; and meningeal enhancement. On gradient echo images (GRE), the changes comprised tubular susceptibility artifacts, consistent with the "susceptibility vessel sign", and additional variably sized/shaped intraparenchymal susceptibility artifacts. Spinal cord lesions presented as fusiform T2-weighted hyperintensities with scattered susceptibility artifacts on GRE and variable parenchymal and meningeal contrast enhancement. On histopathology, subarachnoid hemorrhages and neuroparenchymal areas of edema and necrosis, with or without hemorrhage, indicating ischemic and hemorrhagic infarctions, were found. These lesions were concurrent with severely dilated meningeal and parenchymal arteries and veins plugged by neoplastic lymphocytes and fibrin. Due to the unique angiocentric distribution of IVL, ischemic and hemorrhagic infarcts of variable chronicity affecting both the arterial and venous territories associated with thrombi formation can be detected on MRI.

Ultrasonographic and CT diagnosis of a complete splenic duplication with right splenic torsion and presumed regional splenic vein hypertension in a dog.

An 8-year-old mixed breed dog, was referred for lethargy, abdominal discomfort, hematemesis, and melena. Abdominal ultrasound revealed mineralization of the "right splenic vein" with collateral spiraling and a suspected splenic duplication. Computed tomography confirmed the duplication along with torsion of the splenic pedicle. Exploratory laparotomy confirmed the splenic torsion associated with congestion of the gastrointestinal vessels, indicative of a chronic onset of secondary regional splenic vein hypertension.

Can MRI differentiate between ring-enhancing gliomas and intra-axial abscesses?

Gliomas of the brain may appear as expansile ring-enhancing masses in MRI studies, mimicking the appearance of intra-axial abscesses. The aims of this study were to compare the MRI features of ring-enhancing gliomas and intra-axial brain abscesses in dogs and cats and to identify the characteristics that might help differentiate them. For this multicenter, retrospective, and observational study, the inclusion criteria were as follows: (a) a definitive diagnosis of glioma or abscess based on cytological or histopathological examination following CSF collection or surgical biopsy/necropsy, respectively; (b) MRI study performed with a high- or low-field MRI scanner, including a same plane T1W pre- and postcontrast, a T2W and a T2 FLAIR sequence in at least one plane. If available, delayed T1W postcontrast, T2*W GE, DWI/ADC, and SWI sequences were also evaluated. Sixteen patients were diagnosed with ring-enhancing gliomas, and 15 were diagnosed with intra-axial abscesses. A homogenous signal on T1W (P = 0.049) and T2W (P = 0.042) sequences, a T2W (P = 0.005) or T2*W GE (P = 0.046) peripheral hypointense halo, and an even enhancing capsule (P = 0.002) were significantly associated with brain abscesses. A progressive central enhancement on delayed T1W postcontrast sequences was correlated with ring-enhancing gliomas (P = 0.009). The combination of the following features was suggestive of brain abscess: homogeneous T1W or T2W signal intensity, a T2W or T2*W GE peripheral hypointense halo and an evenly enhancing capsule. Central progression of enhancement on delayed T1W postcontrast sequences was suggestive of glioma.

Gastrointestinal ultrasonographic findings in cats with Feline panleukopenia: a case series.

BACKGROUND: Feline panleukopenia virus (FPV) is very resistant and highly contagious and infects domestic cats and other felids. FPV is particularly widespread among sheltered cats, and is associated with high morbidity and mortality, causing severe gastroenteritis characterized by anorexia, lethargy, fever, dehydration, hemorrhagic diarrhea, and vomiting. There is currently no data on the ultrasonographic features of cats affected with FPV. This case series describes abdominal ultrasonographic findings in shelter cats with naturally-occurring FPV, and assesses whether are associated with clinical and laboratory findings. Cats affected by FPV were enrolled in the study if an abdominal ultrasound was performed within 12 hours of diagnosis. Clinical, laboratory and survival data were collected from medical records. Ultrasonographic examinations were reviewed for gastrointestinal abnormalities and their associations with the above data were explored. RESULTS: Twenty-one cats were included. Nine cats (42.9%) died and 12 (57.1%) recovered. Based on ultrasonography, the duodenum and jejunum showed thinning of the mucosal layer in 70.6% and 66.6% of cats, thickening of the muscular layer in 52.9% and 57.1% of cats, and hyperechogenicity of the mucosa in 41.2% and 33.3%. Jejunal hyperechoic mucosal band paralleling the submucosa and irregular luminal surface were both observed in 33.3% of the cats. Survival was positively associated with increased jejunal mucosal echogenicity (P = 0.003) and hyperechoic mucosal band (P = 0.003). Peritoneal free fluid was positively associated with vomiting (P = 0.002). CONCLUSIONS: This study provides ultrasonographic features of naturally-occurring FPV in cats, which, as expected, are compatible with gastroenteropathy. The most frequent findings were diffuse small intestine mucosal layer thinning, muscular layer thickening and mucosal hyperechogenicity, jejunal hyperechoic mucosal band and irregular luminal surface. Ultrasonographic features may be useful to complete the clinical picture and assess the severity of the gastroenteropathy in FPV cats. Prospective studies are needed to confirm ultrasonographic prognostic factors.

CT angiography identifies collaterals in dogs with splenic vein obstruction and presumed regional splenic vein hypertension.

Occlusion of the splenic vein, without occlusion of the portal vein, can lead to a localized, regional splenic hypertension, referred as sinistral or left-sided portal hypertension in the human radiology literature. In people, may cause gastrointestinal hemorrhages from the esophageal and gastric varices and the primary pathology usually includes pancreatitis and pancreatic neoplasms. The final diagnosis of localized splenic hypertension necessitates accompanying normal liver functions and a patent extrahepatic portal vein. Following obstruction, the resultant elevated splenic bed venous pressure causes formation of collateral routes, the extent of which depends upon the level and degree of obstruction. In this retrospective descriptive study, authors assessed the collateral pathways in dogs with isolated splenic vein occlusion and possible regional splenic vein hypertension. Out of the 46 patients initially recruited, 25 were excluded due to the presence of concomitant portal thrombosis and direct/indirect CT signs of portal hypertension. The remaining 21 dogs had clinicopathological tests suggesting normal liver function. The causes of obstruction identified included splenic pedicle torsion, tumoral splenic vein invasion, and splenic vein thrombosis. Four of 21 dogs with isolated splenic vein obstruction showed collateral pathways through the left gastroepiploic vein (4/4), left gastric vein (2/4), and splenogonadal vein (1/4). The diagnosis of isolated, regional splenic hypertension should be based on clinical, biochemical, and radiological evaluation. Computed tomography is an excellent tool to assess the collateral patterns and to determine the underlying cause.

Radiography and ultrasonography of pneumatosis intestinalis in a cat.

An adult cat was presented for acute history of vomiting and collapse. Radiographs showed the presence of air within small intestinal walls and arborizing gas patterns within the liver, compatible with pneumatosis intestinalis and presumed portal venous gas, respectively. An abdominal ultrasound the following day was suggestive of gas within the intestinal wall, however, gas within the hepatic vasculature, parenchyma, or biliary tree was not evident. Due to progressive clinical deterioration of the patient, the owners elected humane euthanasia. Necropsy revealed severe necrotizing hemorrhagic enterotyphlocolitis secondary to Clostridium difficile toxin.

Diagnostic imaging for the assessment of acquired abdominal vascular diseases in small animals: A pictorial review.

Advances in interventional radiology and surgical techniques now allow complex abdominal diseases to be more successfully treated in small animals. Abdominal vascular alterations, acquired as individual process or as complication of other lesions such as neoplasia, can be life-threatening or at least greatly limit curative interventions of underlying diseases. Computed tomography (CT) and high-definition ultrasonography are now readily available in veterinary referral centers. Yet, there is little information currently available on the use of these modalities for the diagnosis and characterization of these vascular alterations. The purpose of this article is to review the CT and ultrasonographic findings of acquired vascular diseases in the abdomen of dogs and cats, using both the veterinary and human medicine literature as references, and highlighting essential concepts through figures.

Canine insulinomas appear hyperintense on MRI T2-weighted images and isointense on T1-weighted images.

Clinical and imaging diagnosis of canine insulinomas has proven difficult due to nonspecific clinical signs and the small size of these tumors. The aim of this retrospective case series study was to describe MRI findings in a group of dogs with pancreatic insulinomas. Included dogs were presented for suspected pancreatic insulinoma, MRI was used to assist with localization of the primary lesion, and the diagnosis was confirmed with surgical exploratory laparotomy and histopathology. The MRI studies for each dog were retrieved and the following data were recorded: T1-weighted and T2-weighted signal intensities, type of contrast enhancement, size and location of the primary lesion, and characteristics of metastatic lesions (if present). A total of four dogs were sampled. In all patients, the insulinoma displayed high-intensity signal on T2-weighted fat saturation images, similar to human studies. On postcontrast T1-weighted fat saturation images, the tumors were primarily isointense to normal pancreatic tissue, in contrast to human studies where a low-intensity signal is typically identified. Abnormal islet tissue was detected with MRI in all four dogs and metastases were identified in three dogs. Variations in the MRI appearance of primary and metastatic lesions were identified and could have been related to the variation of tissue composition, including the presence of neoplastic cells, hemorrhage, and fibrovascular stroma, and to the transformation of this tissue throughout the disease process.

A "spaghetti sign" in feline abdominal radiographs predicts spleno-systemic collateral circulation.

A well-defined, tortuous, and tubular soft tissue opaque structure has been observed within the left mid abdominal cavity in digital radiographs of some cats. Authors hypothesized that this radiographic "spaghetti sign" is a predictor for spleno-systemic collateral circulation. The purpose of this multicenter, retrospective, descriptive study was to further characterize radiographic "spaghetti signs" in a group of cats and compare these signs with results of ultrasonography or computed tomography (CT). A multiinstitutional database was reviewed for radiographic reports describing the presence of a tubular tortuous structure within the left abdominal cavity of feline patients. Patients were included if additional abdominal ultrasound or CT studies were available. Imaging studies and medical records were retrieved and reviewed by a board-certified veterinary radiologist, and characteristics were recorded. Eighteen cats met the inclusion criteria. The presence of a tortuous vessel consistent with spleno-systemic collateral circulation was observed by ultrasonography or CT examinations in all patients with a "spaghetti sign" on abdominal radiographs. All patients were spayed females. One cat showed ultrasonographic imaging features consistent with portal hypertension. In conclusion, findings supported the hypothesis that a radiographic "spaghetti sign" is a predictor of spleno-systemic collateral circulation in feline patients and should be distinguished from other signs such as focal loss of serosal detail or a space occupying lesion.

Multidetector computed tomographic pulmonary angiography in a cat with fatal heartworm disease.

A 17-month-old neutered male domestic shorthair cat was referred for a computed tomographic (CT) study of the thorax due to respiratory distress. Multidetector CT angiography showed a multifocal interstitial ground glass opacity, tortuous and blunted pulmonary arteries consistent with thromboembolism with perivascular lung infiltration and hypoventilation in multiple lung lobes. A blood antigen test was positive for Dirofilaria immitis. The cat's clinical condition rapidly declined and the owners elected euthanasia. The histopathologic examination confirmed heartworm disease with parasitic pulmonary thromboembolism.

The computed tomographic "tree-in-bud" pattern: Characterization and comparison with radiographic and clinical findings in 36 cats.

In humans, a CT "tree-in-bud" pattern has been described as a characteristic of centrilobular bronchiolar dilation, with bronchiolar plugging by mucus, pus, or fluid. Aims of this retrospective, descriptive, multi-center study were to characterize the CT appearance of a "tree-in-bud" pattern in a group of cats, and compare this pattern with radiographic and clinical findings. Databases from four hospitals were searched during the period of January 2012 to September 2015 and cats with thoracic radiographs, thoracic CT scans and CT reports describing findings consistent with a "tree-in-bud" pattern were included. Images were reviewed by two veterinary radiologists and characteristics were recorded based on consensus. Clinical findings were recorded by one observer from each center. Thirty-six cats met inclusion criteria. Six cats were asymptomatic, 12 were diagnosed with bronchial disease and 23 were suspected to have bronchial disease. Right cranial and right caudal lung lobes were most commonly affected on both imaging modalities. Localization of the "tree-in-bud" pattern was most often peripheral. On radiographs, the CT "tree-in-bud" pattern often appeared as soft-tissue opacity nodules; their number and affected pulmonary segments were often underestimated compared with CT. In conclusion, the "tree-in-bud" pattern should be considered as a differential diagnosis for radiographic soft tissue opaque nodules in feline lungs. Based on lesion localization and presence or suspicion of a concomitant bronchial disease for cats in this sample, authors propose that the CT "tree-in-bud" pattern described in humans is also a characteristic of bronchial or bronchiolar plugging and bronchial disease in cats.

Vascular conspicuity differs among injection protocols and scanner types for canine multiphasic abdominal computed tomographic angiography.

Multiphasic multidetector computed tomographic angiography is a standard diagnostic test for canine abdominal vascular disorders. Three imaging protocols have been previously described. The test-bolus protocol allows precise timing but can be time consuming to perform. Bolus-tracking software is fast and easy to use but can be problematic for exact timing of vascular phases. A recently described fixed-injection-duration protocol is not influenced by body weight and provides a wider temporal window for arterial acquisitions. Objectives of this retrospective and prospective, multicentric, method comparison study were to determine which of the three multidetector computed tomographic angiography protocols allows best vascular conspicuity of the canine abdomen and to assess the influence of different multidetector computed tomography (CT) scanners on study quality. Triple-phase multidetector computed tomographic angiography canine abdominal studies from 30 dogs were retrospectively retrieved from three different institutions. Each institution performed one of the three computed tomographic angiography protocols (4-row and 16-row multidetector CT). Prospectively, the three protocols were also acquired with similar conditions on a 64-row MDCT in 21 dogs. Main abdominal vessels were scored by blinded readers for each phase. The fixed-injection-duration protocol had the best combined arterial and portal vascular conspicuity on scanners of limited speed, while the test-bolus protocol provided the best overall vascular conspicuity on 64-row multidetector CT scanner. The quality of arterial studies performed on 64-row MDCT scanner was improved compared to the ones performed on four- to 16-row multidetector CT scanners. Findings supported the fixed-injection-duration protocol as the best compromise between an ideal portal vascular enhancement and an easily reproducible protocol on scanners with low and high number of detector rows.

Canine and feline abdominal arterioportal communications can be classified based on branching patterns in computed tomographic angiography.

Arterioportal vascular anomalies are communications between the splanchnic arteries and the portal system that represent a rare cause of presinusoidal portal hypertension in small animals. There is little information concerning the imaging findings of arterioportal communications in small animals and no classification could be found for radiologists and surgeons. The aims of this retrospective descriptive multicentric study were to describe the computed tomographic characteristics of arterioportal communications in a group of cats and dogs, and to propose a classification based on computed tomography (CT) angiographic anatomy. Computed tomography databases from multiple veterinary hospitals were searched for cats and dogs with a diagnosis of arterioportal communication. A total of 36 animals (33 dogs, three cats) met the inclusion criteria. There were 32 intrahepatic arterioportal malformations and four extrahepatic fistulae. The intrahepatic arterioportal malformations were classified as right divisional (11/32) and left divisional (21/32), and the left divisional were subclassified as left medial (16/21) and left lateral (4/21). One patient showed multiple intrahepatic arterioportal communications with concomitant left medial and left lateral conformations. Two patients with intrahepatic arteriovenous malformation showed concomitant congenital intrahepatic shunts. The proposed anatomical classification based on CT angiography could allow veterinary radiologists to have a more systematic approach and help improve the radiologist-surgeon communication.

RADIOGRAPHIC APPEARANCE OF PRESUMED NONCARDIOGENIC PULMONARY EDEMA AND CORRELATION WITH THE UNDERLYING CAUSE IN DOGS AND CATS.

Noncardiogenic pulmonary edema is an important cause of respiratory disease in dogs and cats but few reports describe its radiographic appearance. The purpose of this retrospective case series study was to describe radiographic findings in a large cohort of dogs and cats with presumed noncardiogenic pulmonary edema and to test associations among radiographic findings versus cause of edema. Medical records were retrieved for dogs and cats with presumed noncardiogenic edema based on history, radiographic findings, and outcome. Radiographs were reviewed to assess lung pattern and distribution of the edema. Correlation with the cause of noncardiogenic pulmonary edema was evaluated with a Fisher's exact test. A total of 49 dogs and 11 cats were included. Causes for the noncardiogenic edema were airway obstruction (n = 23), direct pulmonary injury (n = 13), severe neurologic stimulation (n = 12), systemic disease (n = 6), near-drowning (n = 3), anaphylaxis (n = 2) and blood transfusion (n = 1). Mixed, symmetric, peripheral, multifocal, bilateral, and dorsal lung patterns were observed in 44 (73.3%), 46 (76.7%), 55 (91.7%), 46 (76.7%), 46 (76.7%), and 34 (57.6%) of 60 animals, respectively. When the distribution was unilateral, pulmonary infiltration involved mainly the right lung lobes (12 of 14, 85.7%). Increased pulmonary opacity was more often asymmetric, unilateral, and dorsal for postobstructive pulmonary edema compared to other types of noncardiogenic pulmonary edema, but no other significant correlations could be identified. In conclusion, noncardiogenic pulmonary edema may present with a quite variable radiographic appearance in dogs and cats.