Observer variability in the histopathological reporting of needle biopsy specimens of the prostate.

ABSTRACT

The Scottish Pathology Consistency Group has in previous studies examined the consistency of histopathological reporting of biopsies from the cervix, bladder, bronchus, and rectum. In the current study, consisting of 100 needle biopsy specimens of the prostate, a single hematoxylin-eosin (H&E) slide from each case was circulated in batches of 10 to the 12 pathologists, who filled in a simple proforma. This had two sections a diagnostic category (benign; suspicious or malignant) along with a standard Gleason score for those regarded as malignant. The majority diagnosis of the 100 cases was benign, 53; suspicious, 1; and malignant, 46. The Kappa value for benign cases versus others was 0.86 and for malignant cases versus others was 0.91. Analysis of the data on Gleason scores showed a value of 0.54 when cases were divided into two categories (2 to 6 v 7 to 10) and 0.41 when three categories were used (2 to 4; 5 to 6; 7 to 10). Although not initially part of the design of the study, the majority diagnosis was compared with the original reported diagnosis. In a small subset, examination of further levels, basal cell antibody staining, along with further clinical information, was obtained. With this added information, it appears that there were probably 52 benign and 48 malignant cases. Of the 48 malignant cases, the group majority diagnosis was malignant, 46; suspicious, 1; and benign, 1. The original reported diagnosis was 56 benign, 1 suspicious, and 43 malignant. The group therefore appeared to perform better than the original reporting pathologists. When compared with the results of our previous studies, this study has shown that the diagnosis of carcinoma of the prostate on a needle biopsy is robust.