Application of gray-scale ultrasonography to identify hepatic parenchymal disease is currently limited. No information exists on the accuracy of ultrasonographic interpretation for differentiating between normal liver and diffuse liver disease or among various types of diffuse liver infiltrates. A retrospective study was performed to assess whether objectively applied ultrasonographic interpretive criteria were statistically useful in differentiating among 7 defined categories of liver condition: vacuolar hepatopathy, lipidosis, inflammation, round cell neoplasia, non-round cell neoplasia, normal liver, and other. Static, hard-copy ultrasonographic images from 299 dogs and 104 cats were analyzed by 3 board-certified veterinary radiologists using a defined set of 51 extra- and intrahepatic criteria. The radiologists had no knowledge of patient status or results of hematologic or serum biochemical analyses. Cytologic, histologic, or necropsy findings were screened by a board-certified veterinary internist and classified into 1 of the 7 categories. Statistical analysis revealed that applying defined ultrasonographic criteria was insufficiently accurate to be clinically useful in categorizing diffuse liver disease in dogs and cats.
COMMENTARY: Of particular interest in this article were the statistical comparisons between the accuracy of predictions of liver status category in the following cases: when biochemical and/or hematologic variables were used alone, when ultrasonographic criteria were used alone, and when both hematologic or biochemical variables were used with ultrasonographic criteria. Hematologic or biochemical variables seemed to predict the category of liver status only nominally better than random assignment. When ultrasonographic criteria alone were used, the accuracy of prediction was much greater. Interestingly, adding hematologic or biochemical variables to the ultrasonographic criteria did not greatly improve this accuracy. Regardless, individual ultrasonographic criteria were not found to be sufficiently accurate to be clinically useful, emphasizing the importance of cytologic or histologic analysis for assessment of diffuse liver disease.
Statistical relevance of ultrasonographic criteria in the assessment of diffuse liver disease in dogs and cats. Feeney DA, Anderson KL, Ziegler LE, et al. AM J VET RES 69:212-221, 2008.