Studies on the agreement of histologic diagnoses of biopsy specimens among pathologists are common in human medicine but rare in veterinary medicine. Histologic diagnosis has not been subjected to validation; the reliability of a diagnostic test depends upon reproducibility of interpretation. Histologic diagnosis is based on a visual image and inherently involves subjectivity. However, this test is particularly amenable to retrospective agreement studies because decisions and descriptions by pathologists are recorded and slides and blocks are available for review. In this study, 513 surgical biopsy specimens of amputated digits were read in a blinded manner by 2 pathologists. The original diagnosis was compared to the one made by the study pathologists. If there was a discrepancy between the diagnoses, the slide was reviewed by a second board-certified pathologist and, if necessary, a third pathologist. Of the 513 cases reviewed, there was agreement in 409 cases (79.9%). Of the 104 cases for which there was disagreement, 77 were determined to be clinically significant. Examples of the most common instances of disagreement between the first and second diagnosis included keratoacanthoma (first diagnosis) and squamous cell carcinoma (second diagnosis), and melanoma (first diagnosis) and inflammation (second diagnosis). The diagnosis of keratoacanthoma was subject to disagreement in 19 of 21 cases (94%).

COMMENTARY: The topic of diagnostic agreement (and disagreement) brings home a reality that most pathologists accept-that is, time-restricted examination of complex visual images followed by subjective interpretation may result in variance of opinion. Experienced clinicians at veterinary schools who have one-on-one access to pathologists and the capacity for the pathologist to actually show and demonstrate the entity diagnosed often understand diagnostic agreement discrepancies. However, for practicing clinicians, a 20% diagnostic disagreement that is clinically significant three fourths of the time may seem alarming. Obviously, pathologists would like for diagnostic disagreement to be 0%. However, the reality of disagreement points to the necessity of good feedback and interaction between clinicians and the pathologists that they choose to use. Additional agreement studies such as this are warranted to point to problem diagnoses in other tissues and organs that may be flags signifying the need for greater vigilance or review within laboratories to provide the most accurate diagnosis.

Agreement among surgical pathologists evaluating routine histologic sections of digits amputated from cats and dogs. Wobeser BK, Kidney BA, Powers BE, et al.  J VET DIAGN INVEST 19:439-443, 2007.