Abstract
OBJECTIVE To characterize CT findings in dogs with a presumptive diagnosis of chronic bronchitis, estimate the accuracy of thoracic CT for the diagnosis of chronic bronchitis in dogs, and determine interobserver agreement for this method.
DESIGN Retrospective case-control and cross-sectional study.
ANIMALS 26 dogs with confirmed chronic bronchitis and 20 control dogs with unremarkable results of thoracic CT and no recorded history of cough.
PROCEDURES Thoracic CT images of all dogs were interpreted for signs of chronic bronchitis by 2 observers who used specific criteria; observers also used the images to compute the bronchial wall thickness-to-pulmonary artery diameter (BWPA) ratio of the cranial lung lobes. Interobserver agreement was assessed for both diagnostic approaches. Performance of thoracic CT and the BWPA ratio specifically in the diagnosis of chronic bronchitis were evaluated, with the final diagnosis made by the attending internist as the reference standard. Associations between independent variables and the BWPA ratio for all dogs were assessed by linear regression.
RESULTS Accuracy of thoracic CT examination for the diagnosis of chronic bronchitis was 57%, sensitivity was 46%, and specificity was 90%. Interobserver agreement was moderate (κ = 0.50). The BWPA ratio had poor accuracy for discriminating dogs with chronic bronchitis from control dogs. Linear regression revealed that as dog body weight increased, BWPA ratios for the left and right cranial lung lobes decreased slightly but significantly.
CONCLUSIONS AND CLINICAL RELEVANCE These results suggested that thoracic CT and the associated BWPA ratio have limited value in the diagnosis of chronic bronchitis in dogs.
Chronic bronchitis, defined as chronic or recurring cough evident most days during 2 consecutive months in the preceding year with a lack of other specific bronchopulmonary disease, is a common cause of chronic cough in dogs.1 The diagnosis of chronic bronchitis is challenging and relies on exclusion of other causes of chronic cough through results of thoracic radiography, bronchoscopy, cytologic analysis and bacterial culture of BAL fluid samples, hematologic analysis, arterial blood gas analysis, pulmonary function tests, and fecal examination for respiratory parasites.2 Although bronchoscopy and BAL are important techniques in the diagnosis of this condition, neither technique is highly sensitive nor specific for bronchopulmonary diseases and both have limited value in the exclusion of other causes of chronic cough.3,4
The radiographic diagnosis of chronic bronchitis relies on identification of a bronchial pattern, demonstrated by a greater than typical number of visible bronchial walls or conspicuity of the bronchial walls.5 Radiography is a relatively inexpensive and noninvasive technique for assessment of the pulmonary parenchyma and, unlike bronchoscopy, can be performed on unanesthetized dogs. However, the sensitivity, specificity, and accuracy of thoracic radiography for the diagnosis of chronic bronchitis in dogs are reportedly only fair at 52% to 65%, 91%, and 65% to 74%, respectively.5
In humans, chronic bronchitis is often associated with other pulmonary conditions (eg, pulmonary emphysema) as a component of chronic obstructive pulmonary disease. The diagnosis of chronic obstructive pulmonary disease is usually made by spirometry, but thoracic CT is widely used to assess broncho-pulmonary lesions.6–9 The severity of bronchial wall thickening is associated with the severity of dyspnea, cough, and wheezing,10,11 and interobserver agreement for detection of bronchial wall thickening with CT is good.12
Accuracy of the BWPA ratio in thoracic CT images for the detection of bronchial wall thickening has been evaluated in dogs with a suspected diagnosis of chronic bronchitis, revealing a significant association between this ratio and a presumptive diagnosis of chronic bronchitis.13 A ratio ≥ 0.6 for the cranial lung lobes has a sensitivity of 77% and a specificity of 100% for the detection of bronchial wall thickening.13 The purpose of the study reported here was to characterize the CT features of chronic bronchitis in dogs and estimate the sensitivity, specificity, and accuracy of thoracic CT for the diagnosis of this disease by use of both subjective CT assessment and the BWPA ratio.
Materials and Methods
The study protocol was approved by the Veterinary School Research Ethics Committee of the University of Liverpool. The clinical records database of the University of Liverpool's Small Animal Teaching Hospital was reviewed to identify all dogs with a record of chronic cough between January 1, 2007, and November 30, 2014 (ie, case dogs). Dogs were included in the study if they had a history of chronic cough; thoracic CT, bronchoscopic, and BAL fluid cytologic analysis findings compatible with chronic bronchitis; no evidence of clinically important cardiac disease identified on physical examination or cardiac ultrasonography; and a final diagnosis of chronic bronchitis by the attending clinician. Dogs with BAL fluid and fecal analysis findings compatible with parasitic, bacterial, or fungal infection or eosinophilic bronchopneumopathy were excluded. In addition, a group of 20 control dogs with no recorded history of coughing and with thoracic CT findings considered unremarkable at the time of initial recording was randomly selected from among 43 dogs in the patient database that had these inclusion criteria.
Data collection and CT scan review
Data were collected for each dog regarding body weight, reproductive status, age, and concurrent diseases at the time of the visit that qualified them for inclusion as well as breed and sex. Thoracic CT scans were acquired with a helical 4-slice CT scannera or helical 80-slice CT scannerb by use of a standardized institutional protocol. Examinations were performed with the dogs positioned in sternal recumbency and anesthetized to allow preacquisition manual hyperventilation of the lungs or with the dogs heavily sedated. No manual lung inflation was performed during image acquisition. Slice thickness, pitch, and collimation varied among scanning sessions. Images were reconstructed by use of lung and soft tissue reconstruction algorithms. For some dogs (n = 19), only images acquired after administration of iodinated contrast mediumc (2 mL/kg [0.9 mg/lb]) via cephalic vein were available.
Patient identification data were removed from CT images for both groups. Images were then independently interpreted with the aid of medical imaging softwared by a board-certified veterinary radiologist (IF) and an observer experienced in reading CT images (LRM). Both observers were blinded to the final diagnosis and number of dogs in each group and were asked to classify each dog as to whether it had chronic bronchitis. A subjective CT-based diagnosis of chronic bronchitis was made when bronchial wall thickening was observed in the absence of any other clinically relevant intrathoracic disease. For each dog, observers were asked to measure the BWPA ratio of the cranial lung lobes as described elsewhere.13 A subjective CT scoring system was used to assess the type and severity of the bronchial wall thickening (Figure 1). Lesions were characterized as generalized, multifocal, or focal. Presence or absence of mucoid plugging, bronchial or tracheal narrowing or flattening, and bronchiectasis was noted. Presence or absence of areas of pulmonary consolidation or atelectasis, lymphadenopathy, pleural effusion, tracheal exudate or other tracheal lesions, pleural nodules or thickening, and pneumothorax was also noted. Pulmonary arterial and venous size was subjectively evaluated.
Transverse CT images (lung reconstruction algorithm and lung window) of the thorax of a dog without bronchial thickening (A), a dog with grade 1 bronchial thickening (B), a dog with grade 2 bronchial thickening (C), and a dog with grade 3 bronchial thickening (D).
Citation: Journal of the American Veterinary Medical Association 253, 6; 10.2460/javma.253.6.757
Statistical analysis
Statistical analyses were performed with statistical software.e Descriptive statistics were calculated to summarize characteristics of dogs in the chronic bronchitis and control groups. Data distribution for continuous variables was assessed through graphical analysis and the Kolmogorov-Smirnov test. Normally distributed continuous data are reported as mean (SD), and nonnormally distributed data are reported as median (IQR). Percentages and associated 95% CIs were computed for categorical data. The distribution of continuous variables was assessed through graphical analysis and the Kolmogorov-Smirnov test.
In the diagnostic test evaluation portion of the study, the primary outcomes considered were the performance of subjective CT evaluation or specific CT measurements (ie, BWPA ratio) for the diagnosis of chronic bronchitis. Accuracy, sensitivity, specificity, and positive and negative predictive values were calculated, with final diagnosis made by the attending veterinary internist as the reference standard. An ROC curve was created, and area under the curve was calculated for the left and right cranial BWPA ratios. Interobserver agreement for the subjective CT-based diagnosis of chronic bronchitis (a categorical variable) was assessed through calculation of the Cohen κ statistic. Interobserver agreement for continuous variables (left and right cranial bronchial wall thickness-to-pulmonary artery ratio) was assessed through calculation of intraclass correlation coefficients (2-way single measurement for absolute agreement) and associated 95% CIs.
Differences in continuous variables (age and weight) between dog groups were assessed with the Mann-Whitney U test. The BWPA ratios were compared between the left and right cranial lung lobes with the Wilcoxon matched pairs signed rank test. Associations between independent variables (age, weight, and sex) and BWPA ratios were assessed by means of linear regression. Variables yielding a value of P < 0.25 on preliminary univariable analysis were considered for inclusion in a final multivariable model. If 2 variables had a correlation coefficient > 0.70, the variable with the smallest P value on univariable analysis was selected for incorporation in the multivariable model. A manual backward stepwise procedure was used for model creation, and variables yielding a value of P < 0.05 were retained. Confounding factors were detected by assessing parameter estimates for substantial changes (> 20%) following their removal from the model. Interaction terms were considered for any combinations of variables remaining in the final models.
Results
Twenty-six dogs evaluated over the 8-year inclusion period met the criteria for having chronic bronchitis, yielding a total of 46 dogs (including 20 control dogs with no recorded history of cough) that were included in the study. Dogs were classified as Labrador Retriever (n = 7), mixed-breed dog (6), Cocker Spaniel (5), English Springer Spaniel (4), and various other breeds (≤ 2 each). Among dogs with chronic bronchitis, neutered males were most common (n = 13 [50%]), followed by neutered females (8 [31%]), sexually intact females (3 [12%]), and sexually intact males (2 [8%]). Median age of dogs with chronic bronchitis was 8.2 years (IQR, 5.7 to 10.1 years), and median body weight was 21.3 kg (46.9 lb; IQR, 13.9 to 40.8 kg [30.6 to 89.8 lb]). For control dogs, these values were 8 years (IQR, 5 to 10 years) and 28.5 kg (62.7 lb; IQR, 14.3 to 38.5 kg [31.5 to 84.7 lb]). No significant difference was identified between groups in age (P = 0.63) or body weight (P = 0.55).
Clinical and CT findings
All dogs with a diagnosis of chronic bronchitis had negative results of bacterial culture of BAL fluid samples. One dog had a fecal parasitologic analysis performed, and the result was negative. Seven dogs had concurrent disease, including anal sac adenocarcinoma (n = 1), atypical hypoadrenocorticism (1), systemic hypertension (1), lymphoplasmacytic rhinitis (1), hypothyroidism (1), diabetes mellitus (1), and iatrogenic hyperadrenocorticism (1). Two dogs had evidence of bronchomalacia on bronchoscopic examination.
For control dogs (with unremarkable results of thoracic CT examination), the CT examination had been performed because of tumor staging (n = 10), rectal bleeding (1), immune-mediated polyarthritis (1), bilateral nasal discharge (1), abdominal pain (1), suspected pulmonary hypertension (1), hyperadrenocorticism (1), and weight loss (1). For dogs with chronic bronchitis, 8 (31%) had no abnormalities identified on thoracic CT examination. The distribution of pulmonary parenchymal lesions in dogs with chronic bronchitis was classified as no lesions for 13 (50%) dogs, multifocal for 9 (35%) dogs, generalized for 2 (8%) dogs, and focal for 2 (8%) dogs. Although the pulmonary parenchymal lesions represented atelectasis in most dogs, CT images of the 2 dogs with generalized lesions had patchy, ground-glass attenuation of the pulmonary parenchyma (Figure 1). Degree of bronchial thickening was classified as none for 14 (54%) dogs with chronic bronchitis, mild (grade 1) for 9 (35%) dogs, moderate (grade 2) for 2 (8%) dogs, and severe (grade 3) for 1 (4%) dog. The distribution of bronchial lesions in affected dogs (n = 12) was classified as multifocal for 7 (27%) dogs and generalized for 5 (19%) dogs with chronic bronchitis. Mucoid plugging with (n = 2 [8%]) or without (1 [4%]) associated pulmonary consolidation and tracheal exudate (2 [8%]) was uncommon. One (4%) dog had tracheal and mainstem bronchial collapse. Bronchiectasis was observed in 4 (15%) dogs. Lymphadenomegaly was identified in 5 (19%) dogs; the tracheobronchial lymph nodes were affected in 2 (8%) dogs, and the axillary, cranial mediastinal, and sternal lymph node were affected in 1 (4%) dog each. For 4 of the 5 dogs with lymphadenomegaly, the largest lymph node height in transverse section was < 2 cm.
The BWPA ratio was ≥ 0.6 for at least 1 cranial lung lobe in 4 (15%) dogs with chronic bronchitis and 1 (5%) control dog according to one observer and in only 1 (4%) dog with chronic bronchitis and none of the control dogs according to the other observer.
Performance of subjective CT evaluation and BWPA ratio for the diagnosis of chronic bronchitis
Subjective CT evaluation had an accuracy of 57% for the diagnosis of chronic bronchitis, with a sensitivity of 46% (95% CI, 27% to 66%) and specificity of 90% (95% CI, 67% to 98%). Positive predictive value was 86% (95% CI, 56% to 98%), and negative predictive value was 56% (95% CI, 38% to 73%).
Median BWPA ratio for the left cranial lung lobe was 0.36 (IQR, 0.32 to 0.44), which was slightly but nonsignificantly (P = 0.72) larger than the median value for the right cranial lung lobe (0.35 [IQR, 0.31 to 0.43]). A small but significant difference in median BWPA ratio was identified between dogs with chronic bronchitis and control dogs for the right cranial lung lobe (0.37 [IQR, 0.31 to 0.47] and 0.32 [0.28 to 0.36], respectively; P = 0.02) but not for the left cranial lung lobe (0.38 [IQR, 0.32 to 0.48] and 0.33 [0.31 to 0.39], respectively; P = 0.08).
Area under the ROC curve for the BWPA ratio used to diagnose chronic bronchitis, with the final diagnosis made by the attending internist used as the reference standard, was 0.66 for the left cranial lung lobe and 0.70 for the right; only the area for the right side was significantly (P = 0.02) different from 0.5. The ROC curve analysis revealed that a cutoff value of 0.35 for the BWPA ratio provided an optimum balance of sensitivity and specificity for both the left and right BWPA ratios, with sensitivities of 65% (left lung) and 67% (right lung) and specificities of 60% (left lung) and 65% (right lung). Increasing the cutoff value to ≥ 0.6 resulted in a sensitivity of 8% and specificity of 100% for the left cranial lung lobe and of 4% and 100%, respectively, for the right cranial lung lobe.
Interobserver agreement for the subjective CT diagnosis of chronic bronchitis was moderate (κ = 0.50 [95% CI, 0.22 to 0.78]). The calculated intraclass correlation coefficients for the 2 observers' BPWA ratio measurements for the right (0.59 [95% CI, 0.37 to 0.76]) and left (0.41 [95% CI, 0.13 to 0.63]) cranial lung lobes were all < 0.6, indicating slight to poor agreement.
Univariable linear regression revealed no association between dog sex and BWPA ratios for both the left and right cranial lung lobes. The P values for associations between age and BWPA ratios (P = 0.06 for the right and P = 0.11 for the left side) and between body weight and BWPA ratios (P = 0.001 for both sides) met the criteria for consideration of these variables for inclusion in a multivariable model. In the final multivariable model, after controlling for body weight, age was no longer significant, meaning that only body weight was retained. In the final models of the BWPA ratio for the left and right cranial lung lobes, as body weight increased, the BWPA ratio decreased slightly (correlation coefficients, 0.49 and 0.50, respectively; P = 0.001 for both).
Discussion
The present study was conducted to characterize thoracic CT findings for dogs with a presumptive diagnosis of chronic bronchitis and assess the accuracy of and interobserver agreement for thoracic CT to diagnose chronic bronchitis in dogs. Findings indicated that more than half of the affected dogs had no evidence of bronchial wall thickening on CT examination and that bronchial plugging, tracheal exudate, bronchiectasis and lymphadenopathy were uncommon. Interestingly, a few dogs with chronic bronchitis had CT evidence of multifocal ground-glass attenuation or small areas of pulmonary consolidation. These dogs had the most severe bronchial changes, and the noted pulmonary parenchymal lesions could have represented secondary bronchopneumonia, pneumonitis, atelectasis due to bronchial plugging and obstruction, or a combination of these conditions.
Thoracic CT had limited accuracy and moderate interobserver agreement for the diagnosis of chronic bronchitis in dogs. The results obtained were comparable with those of another study,5 in which the accuracy of thoracic radiography was evaluated for the diagnosis of chronic bronchitis in dogs, yielding a similarly poor sensitivity (52% to 65%) and good specificity (91%). One might expect that the diagnostic performance of CT would be superior to that of radiography.
In the present study, dogs with positive results of bacterial culture of BAL fluid were excluded on the basis that recovered organisms could be of hematogenous origin or represent aspiration pneumonia. Dogs with a positive culture result possibly also had chronic bronchitis with secondary infection. Similarly, we decided to exclude dogs with a diagnosis of pulmonary eosinophilic infiltrate. Although the aim of CT was never to differentiate between types of bronchial inflammatory infiltrate, the distinction between eosinophilic bronchitis and eosinophilic bronchopneumopathy can be unclear; therefore, all dogs with pulmonary eosinophilic infiltrate were excluded to prevent any confusion. Results of BAL fluid analysis were not considered as exclusion criteria in the previous radiographic study,5 and dogs with eosinophilic bronchitis may have dramatic bronchial thickening.14 Likewise, dogs with chronic bronchitis and secondary infection might have more severe lesions identified on CT than dogs without secondary infection. Excluding these patients from our sample of affected dogs could represent a selection bias in the present study, leading to poorer sensitivity and underestimation of the lesions visible on CT.
The BWPA ratio in the study reported here had poor accuracy for the diagnosis of chronic bronchitis in dogs and slight to poor interobserver agreement. In a previous study,13 the accuracy of this ratio for assessing bronchial wall thickening in dogs with chronic bronchitis was evaluated, revealing a high accuracy, sensitivity of 77%, and specificity of 100% when a cutoff ratio value for the cranial lung lobes of ≥ 0.6 was used to indicate bronchial wall thickening. These results differ markedly from ours insofar as only 15% of dogs with chronic bronchitis in the present study had a ratio ≥ 0.6, and use of this cutoff value resulted in poor sensitivity (8% and 4% for BWPA ratios for the left and right cranial lung lobes, respectively) and a specificity of 100% for ratios for both cranial lung lobes. In the previous study,13 observers were not blinded to the final diagnosis, bronchoscopy results were available for only 2 of 16 affected dogs, dogs with eosinophilic bronchitis were not excluded, and there was no mention of microbial culture of BAL fluid.
It was also possible that affected dogs of the previous study had more severe bronchial changes than the dogs of the present study. In the previous study,13 a significant difference existed between affected and control dogs in mean body weight (11.5 kg [25.3 lb] and 23.5 kg [51.7 lb], respectively), which could have contributed to the observed difference in the BWPA ratio between affected and control dogs. Indeed, a negative correlation was identified in the present study between dog body weight and BWPA ratio, which was not expected. A possible explanation is that as the size of a dog increases, the pulmonary artery diameter increases proportionally more that the bronchial wall thickness.
The present study had several limitations. Because of the retrospective nature of the data collection, CT and anesthetic protocols varied among dogs, and such differences might have influenced BWPA ratio measurements and the subjective assessment of bronchial thickness. However, all CT images were considered of diagnostic quality. In addition, most dogs had no fecal analysis performed to detect respiratory parasites. The Baermann technique is considered the gold standard for the diagnosis of angiostrongylosis during the patent period of the disease in dogs.15 However, no lar vae are shed during the pre-patent period and intermittent shedding may occur during the patent period15; therefore, even if all dogs had been tested with this technique, we could not have excluded the possibility of angiostrongylosis. A study16 involving dogs with experimentally induced angiostrongylosis showed that all infected dogs had persistent visible larvae and a high eosinophil count on BAL fluid cytologic analysis.16 Despite the exclusion of dogs with such findings from the present study, again the possibility of angiostrongylosis could not be completely ruled out.
No cardiac ultrasonographic examination was performed for dogs included in the study reported here. Although no dogs had evidence of pulmonary vessel enlargement on subjective evaluation, pulmonary hypertension was not excluded, and this condition, if present, could have artifactually decreased the BWPA ratio. Two dogs had bronchomalacia identified on bronchoscopic evaluation. One of those dogs had CT evidence of tracheal and bronchial collapse, and this finding alone could have been responsible for the chronic cough; however, patient history and clinical and biological findings were also consistent with chronic bronchitis.
Other limitations of the study reported here pertained to the intrinsic subjectivity of bronchoscopic assessment, which is influenced by the experience of the performing clinician, and the variability in BAL fluid findings. Neither of these limitations could have been avoided. In addition, no attempt was made to exclude dogs with subclinical chronic bronchitis from the control group by performing bronchoscopy or BAL; however, such testing was not indicated for those dogs. Comparison between the severity or extent of CT findings noted with chronic bronchitis and the prognosis of the disease or severity of clinical signs was beyond the scope of the study but would be interesting to explore.
Overall, the results of the present study indicated that thoracic CT and the BWPA ratio had limited diagnostic value and appeared to offer little advantage over thoracic radiography in the diagnosis of chronic bronchitis in dogs. However, thoracic CT could be useful for excluding other thoracic diseases such as pulmonary neoplasia or granulomatous disease that may be responsible for chronic cough. Chronic bronchitis appears to be largely a diagnosis of exclusion when it comes to CT, and unremarkable thoracic CT findings do not rule out this condition in dogs with a persistent cough.
Acknowledgments
No financial support was received in connection with this study or the writing of this manuscript.
Presented in part at the European Association of Veterinary Diagnostic Imaging (British and Irish Division) Pre–British Small Animal Veterinary Association Congress, Birmingham, England, April 2016.
ABBREVIATIONS
BAL | Bronchoalveolar lavage |
BWPA | Bronchial wall thickness to pulmonary artery diameter |
CI | Confidence interval |
IQR | Interquartile (25th to 75th percentile) range |
ROC | Receiver operating characteristic |
Footnotes
Siemens Somatom, Siemens Healthcare Diagnostics, Deerfield, Ill.
Aquilion Prime 80, Toshiba Medical Systems Corp, Tokyo, Japan.
Xenetix, Guerbet Laboratories Ltd, Roissy, France.
OsiriX Pixmeo, Geneva, Switzerland.
SPSS for Windows, version 22.0, SPSS Inc, Chicago, Ill.
References
Hawkins EC, Clay LD, Bradley JM, et al. Demographic and historical findings, including exposure to environmental tobacco smoke, in dogs with chronic cough. J Vet Intern Med 2010;24:825–831.
Padrid PA, Hornof WJ, Kurpershoek CJ, et al. Canine chronic bronchitis. A pathophysiologic evaluation of 18 cases. J Vet Intern Med 1990;4:172–180.
Hawkins EC, DeNicola DB, Kuehn NF. Bronchoalveolar lavage in the evaluation of pulmonary disease in the dog and cat. State of the art. J Vet Intern Med 1990;4:267–274.
Hawkins EC, DeNicola DB, Plier ML. Cytological analysis of bronchoalveolar lavage fluid in the diagnosis of spontaneous respiratory tract disease in dogs: a retrospective study. J Vet Intern Med 1995;9:386–392.
Mantis P, Lamb CR, Boswood A. Assessment of the accuracy of thoracic radiography in the diagnosis of canine chronic bronchitis. J Small Anim Pract 1998;39:518–520.
Brillet PY, Fetita CI, Saragaglia A, et al. Investigation of airways using MDCT for visual and quantitative assessment in COPD patients. Int J Chron Obstruct Pulmon Dis 2008;3:97–107.
Grydeland TB, Dirksen A, Coxson HO, et al. Quantitative computed tomography measures of emphysema and airway wall thickness are related to respiratory symptoms. Am J Respir Crit Care Med 2010;181:353–359.
Huertas A, Palange P. COPD: a multifactorial systemic disease. Ther Adv Respir Dis 2011;5:217–224.
Lee YK, Oh YM, Lee JH, et al. Quantitative assessment of emphysema, air trapping, and airway thickening on computed tomography. Lung 2008;186:157–165.
Niimi A, Matsumoto H, Amitani R, et al. Airway wall thickness in asthma assessed by computed tomography: relation to clinical indices. Am J Respir Crit Care Med 2000;162:1518–1523.
Matsuoka S, Kurihara Y, Nakajima Y, et al. Serial change in airway lumen and wall thickness at thin-section CT in asymptomatic subjects. Radiology 2005;234:595–603.
Diederich S, Jurriaans E, Flower C. Interobserver variation in the diagnosis of bronchiectasis on high-resolution computed tomography. Eur Radiol 1996;6:801–806.
Szabo D, Sutherland-Smith J, Barton B, et al. Accuracy of a computed tomography bronchial wall thickness to pulmonary artery diameter ratio for assessing bronchial wall thickening in dogs. Vet Radiol Ultrasound 2015;56:264–271.
Mesquita L, Lam R, Lamb CR, et al. Computed tomographic findings in 15 dogs with eosinophilic bronchopneumopathy. Vet Radiol Ultrasound 2015;56:33–39.
Traversa D, Guglielmini C. Feline aelurostrongylosis and canine angiostrongylosis: A challenging diagnosis for two emerging verminous pneumonia infections. Vet Parasitol 2008;157:163–174.
Barçante JMP, Barçante TA, Ribeiro VM, et al. Cytological and parasitological analysis of bronchoalveolar lavage fluid for the diagnosis of Angiostrongylus vasorum infection in dogs. Vet Parasitol 2008;158:93–102.