Computed Tomography Accuracy for Prediction of Adrenal Tumor Adhesion With Organs and Vessels in Dogs and Cats

The presence of adhesion created by an adrenal tumor (AT) with adjacent structures is important information for surgical planning. We aimed to establish the accuracy of computed tomography (CT) in predicting adrenal tumor adhesion (ATA) with adjacent organs and vessels. This prospective multicenter study enrolled 30 dogs and 2 cats with AT, pre- and post-contrast CT, and adrenalectomy. Images were reviewed by five blinded radiologists using consensus-based CT criteria for adhesions (five criteria for vessels and eight for organs). During surgery, ATA was observed in 27/32 patients (84.4%). A single criterion alone could not predict the presence of adhesion. Using a probabilistic classification tree, combinations of two or more criteria with other variables (side, image quality, and tumor rupture) showed good accuracy in predicting ATA for some organs and vessels. For vessels, the combination of some criteria could predict the presence of adhesion: for example, for the caudal vena cava, a combination of criteria 1 + 3 + 4 + 5 in a right AT had an accuracy of 86.2%; for the renal vein, the combined 1 + 2+ 3 + 4 + 5 criteria had an accuracy of 84.8% and 93.8% in a CT study of excellent quality. For organs, a combination of specific criteria could predict adhesion: for the kidney, the combination 1 + 2+ 3 + 4 + 5 had an accuracy of 90%. Radiologist's agreement for detecting ATA was excellent (>80%). A combination of repeatable CT criteria could provide probabilistic information for the presence of ATA with adjacent organs and vessels.