Intestinal Biopsies. Substantial variation in opinion was seen regarding the intensity of intestinal mucosal infiltrates, although the variation was unpredictable. This likely occurred because there are no universally accepted criteria for evaluating intestinal mucosal specimens. Further, there is no gold standard for the histologic diagnosis of inflammatory bowel disease. The pathologists worked under conditions similar to those routinely encountered by diagnostic pathologists, namely that little or no information was provided about the animal and intestinal source of tissues. Also, a substantial number of biopsies submitted to laboratories may be from intestines that turn out not to be responsible for the patient’s clinical signs. For 7 of the 14 slides examined, there was nonuniformity of opinion among pathologists, and so accurate comparisons cannot be made between pathologists. Interobserver variation was of concern especially for 1 full&345;thickness surgical biopsy of the ileum of a clinically normal dog. Of the 5 descriptions made, results varied from normal to neoplastic. In defense of these pathologists, the range of appearance of clinically normal intestines may not have been appreciated and they may have been seeing histologic changes that are real but not extensive enough to cause clinical disease. This “compensating” subclinical disease phenomenon also is reported in humans.

Liver Biopsies. Median surface area of the needle biopsy specimens was 1/4 and 1/3 of those from wedge biopsies for dogs and cats, respectively. This may have contributed to the discordancy obtained between morphologic diagnoses of the paired needle and wedge biopsies. Such sampling errors are common in humans as well, especially when there is uneven distribution of lesions in patients with generally diffuse liver disorders. In humans, and in the present study in dogs and cats, important pathologic lesions could have been overlooked or misjudged if only one or a few needle biopsy specimens were examined.

Percutaneous needle biopsy has become popular in veterinary medicine because it is easy and minimally invasive, can be performed with ultrasound guidance, and have a low incidence of complications. However, results of the present study point out the inherent error rate of this technique (52%). Accurate pathologic interpretation depends upon adequate tissue representation and the size of the biopsy needle, and probably needs more than one needle biopsy specimen, especially for certain types of hepatic disease and its lobular distribution.

On a positive note, the 3 independent evaluators had a high rate of agreement for morphologic diagnoses assigned to individual cases. This was attributed to the use of objective diagnostic criteria, and the importance of reviewing adequate tissue samples. However, there was considerable variability in tissue involvement with certain disease processes, ultrasonographic positioning could overlook areas with substantial histologic changes, and biopsying a single liver lobe could misrepresent the overall disease process. Especially disturbing were the reports of hepatic disease from hepatic needle biopsies in 10 of 33 healthy animals. These were false-positive results, as wedge biopsies of these livers were normal and the animals had no clinical evidence of hepatic disease.