Purpose: Serum prostate specific antigen (PSA) in patients with prostate carcinoma is influenced by prostate size, transition zone volume, and tumor differentiation and volume. Immunohistochemistry studies have demonstrated an inverse correlation between PSA staining intensity and tumor grade, yet to our knowledge tissue expression of PSA has never been correlated with serum PSA.
Materials and methods: In 47 radical prostatectomy cases serum PSA was corrected for gland size and tumor volume. Standard immunohistochemistry staining techniques were applied to specimens using monoclonal antibodies to PSA and cytokeratin CAM5.2. Color images of PSA and CAM5.2 immunohistochemistry stained slides were digitally acquired and analyzed using a standard image analysis system. Representative tumor foci in each slide were imaged with a 20x objective and 10x eyepiece. Staining extent and intensity of the tumor epithelium were measured, and stromal elements and luminal areas were excluded from analysis. For each case quantitative PSA staining intensity was expressed relative to keratin staining in adjacent benign epithelium.
Results: Gland volume and tumor volume independently correlated with serum PSA. Furthermore, tissue PSA intensity inversely correlated with histological grade of the tumor (p <0.00001). After gland size, tumor volume and grade were considered, corrected quantitative tissue PSA intensity did not significantly correlate with corrected serum PSA.
Conclusions: Immunohistochemistry expression of tissue PSA in prostate carcinoma cannot be used to explain variations in serum PSA. This discrepancy may relate to differences between the amount of PSA produced by prostatic tumors and the amount secreted, and/or the sensitivity of detecting various tissue isoforms of PSA with immunohistochemistry.