Histopathology slides from sequential cases undergoing endoscopic biopsy were examined by 4 pathologists by pictorial templates.
Changes for 9 microscopic features were recorded as normal, mild, moderate or severe, and 2- and 4-point scales were tested for consistency of interpretation.
Logistic regression models determined odds ratios (OR) of histologic lesions being associated with hypoalbuminemia while ê statistics determined agreement between pathologists on histologic lesions.
Results: There was poor agreement (ê=−0.013 to 0.3) between pathologists, and institution of origin of slides had effect (ê= 1.0 for 3 of 4 lesions on slides from Institution 5) on agreement between pathologists on selected histologic features.
Using 2 point as opposed to 4-point grading scale increased agreement between pathologists (maximum ê= 0.69 using 4-point scale versus maximum ê= 1.0 using 2-point scale).
Significant association (P= .019 .04; 95% OR = 3.1410.84) between lacteal dilation and hypoalbuminemia was found by 3 pathologists.
Conclusions and Clinical Importance: Substantial inconsistency between pathologists remains despite use of pictorial template because of differences in slide processing.
Distinguishing between mild and moderate lesions might be important source of the disagreement among pathologists.
Source: M.D. Willardet al (2010): Effect of Tissue Processing on Assessment of Endoscopic Intestinal Biopsies in Dogs and Cats. In: Journal of Veterinary Internal Medicine
Volume 24 Issue 1, Pages 84 - 89
Published Online: 30 Nov 2009
|
