BackgroundUltrasonography to visualize adrenal gland lesions and evaluate incidentally discovered adrenal masses in dogs has become more reliable with advances in imaging techniques. However, correlations between sonographic and histopathological changes have been elusive. The goal of our study was to investigate which ultrasound features of adrenal gland abnormalities could aid in discriminating between benign and malignant lesions. To this end, we compared diagnosis based on ultrasound appearance and histological findings and evaluated ultrasound criteria for predicting malignancy.ResultsClinical records of 119 dogs that had undergone ultrasound adrenal gland and histological examination were reviewed. Of these, 50 dogs had normal adrenal glands whereas 69 showed pathological ones. Lesions based on histology were classified as cortical adrenal hyperplasia (n = 67), adenocarcinoma (n = 17), pheochromocytoma (n = 10), metastases (n = 7), adrenal adenoma (n = 4), and adrenalitis (n = 4). Ultrasonographic examination showed high specificity (100%) but low sensitivity (63.7%) for identifying the adrenal lesions, which improved with increasing lesion size. Analysis of ultrasonographic predictive parameters showed a significant association between lesion size and malignant tumors. All adrenal gland lesions >20 mm in diameter were histologically confirmed as malignant neoplasms (pheochromocytoma and adenocarcinoma). Vascular invasion was a specific but not sensitive predictor of malignancy. As nodular shape was associated with benign lesions and irregular enlargement with malignant ones, this parameter could be used as diagnostic tool. Bilaterality of adrenal lesions was a useful ultrasonographic criterion for predicting benign lesions, as cortical hyperplasia.ConclusionsAbnormal appearance of structural features on ultrasound images (e.g., adrenal gland lesion size, shape, laterality, and echotexture) may aid in diagnosis, but these features alone were not pathognomic. Lesion size was the most direct ultrasound predictive criterion. Large and irregular masses seemed to be better predictors of malignant neoplasia and lesions <20 mm in diameter and nodular in shape were often identified as cortical hyperplastic nodules or adenomas.