Interobserver reproducibility in the diagnosis of prostatic intraepithelial neoplasia

To assess interobserver reproducibility in the categorization of prostatic intraepithelial neoplasia (PIN) seven pathologists reviewed 25 lesions. Rather than classic or consecutive examples of PIN, cases were selected to represent the full spectrum of diagnostic issues in this field. Lesions were classified into one of six categories: (a) benign prostate tissue, (b) PIN1, (c) PIN2, (d) PIN3, (e) PIN3 cannot rule out associated cancer, and (f) PIN3 plus cancer. Following evaluation of the slides, data were also analyzed by combining several of the groups into three categories: (a) benign/PIN1; (b) PIN2/PIN3/PIN cannot rule out cancer; and (c) PIN plus cancer. The level of agreement was fair (Kappa = 0.33) for the six categories and substantial (Kappa = 0.61) for the three groups. In no case was there a uniform diagnosis of PIN1; in all cases at least some pathologists considered the biopsies to be normal. This finding provides support for not commenting on PIN1 in biopsy material. In general, there was good distinction between low-grade PIN (PIN1) and high-grade PIN (PIN2-3). Among the seven cases for which there was a consensus that the lesion represented high-grade PIN, there was no case in which there was uniform agreement as to whether the lesion represented PIN2 or PIN3. This finding supports combining PIN2 and PIN3 into high-grade PIN. Cases classified as low-grade PIN by some and as high-grade PIN by others were those with pleomorphism but without prominent nucleoli. Difficulties in distinguishing "high-grade PIN" from "high grade PIN cannot rule out cancer" were those with cribriform glands, glands with necrosis, and where high-grade PIN was associated with only a few adjacent small atypical glands. These same histologies caused the participating pathologists difficulty in distinguishing "high-grade PIN cannot rule cancer" from "high grade PIN plus cancer."