Silver-stained structures in prostatic carcinoma: evaluation of diagnostic and prognostic relevance by automated image analysis.

The comparison of the diagnostic and prognostic significance of histology, immunohistochemical parameters (PSA, PSP), and silver-stained nucleolar organizer regions (AgNORs) was estimated in paraffin sections taken of 63 prostatic carcinomas prior to therapy. AgNORs were visualized with a one-step silver staining technique with the appropiate staining time determined by preliminary staining-time series. The mean AgNOR number per cell (n) and the mean AgNOR area per silver-stained dot (A) were determined by means of an automatic image analysis system. Thereby prostatic carcinomas exhibited multiple small AgNORs within their nuclei (n = 4.7, A = 0.09 micron 2), whereas benign prostatic epithelium showed few but large silver-stained particles (n = 1.8, A = 0.27 micron 2; p less than 0.001). This relationship was then calculated as a quotient of AgNOR number and area (NQ = n/A) which provided additional information for the diagnosis of malignancy as well as survival. Univariate survival analysis disclosed a set of four variables predicting death from prostatic cancer; cribriform growth pattern, AgNOR quotient, histological grade, and PSA immunoreactivity. Of these parameters, immunoreactivity of PSA failed to prove its prognostic significance in multivariate survival analysis (Cox model). No relation to prognosis was found for the number as well as the area of AgNORs alone. Therefore, image analysis proved to be a prerequisit for the feasibility of this promising technique by providing objective and reproducible results.