Use of computed radiology as a screening test for the identification of congenital heart disease in dogs

Abstract

The objectives of this study were to evaluate the accuracy of thoracic radiology as a screening test for congenital heart diseases in dogs, to identify the main contributions and limitations of this modality, and to verify the reproducibility of the evaluations by three observers with different levels of training. An interobserver, observational, retrospective and prospective study was carried out, including ninety dogs: thirty healthy animals, thirty with acquired heart diseases and thirty with congenital heart diseases, which all had thoracic radiographs and a confirmed echocardiographic diagnosis. The cases were separated and randomized by a mediator who did not participate in the reading of the radiographic examinations, and no evaluator had access to the patients' data. Sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and accuracy of each observer were calculated in relation to the correct classification of dogs to groups of normal or acquired and congenital heart diseases, as well as identification of enlargement of the cardiac silhouette and large vessels of dogs with congenital heart diseases. Finally, the Kappa coefficient was obtained between the observers to verify the reproducibility of the radiological evaluations performed. In general, sensitivity, PPV and accuracy were unsatisfactory ( 70%), and the agreement ranged from poor to reasonable (between 0 and 0.39). Although greater accuracy was achieved in the differentiation of healthy dogs from those with acquired and congenital heart diseases by thoracic radiography, when compared to the other studies, the modality was able only to identify healthy patients, and could not differentiate the individuals with different forms of heart disease or define the cardiac malformations. In addition, there was low reproducibility between observers, therefore, this technique should not be used as a sole screening method in dogs with suspected congenital heart diseases.