BACKGROUND. Prostate-specific antigen (PSA) cannot differentiate benig
n prostatic hyperplasia (BPH), from prostatitis, or prostate cancer in
the range of 4.0-10 ng/ml. An accurate cytologic or histologic assess
ment is necessary to confirm the proper diagnosis. The nature of a bio
psy tends to make it a selective test not frequently repeated. We are
reporting a technique employing semen as a source for the differential
diagnosis of prostate epithelial cells. METHODS. Eleven vasectomized
and nonvasectomized prostate cancer patients provided semen samples (s
tage T1 to T2). Two patients provided repeat samples. In addition, 15
vasectomized or nonvasectomized individuals without evidence of diseas
e provided semen samples. Three million cells fixed with 50% ethanol w
ere stained by an antibody (7E11.C5) to prostate-specific membrane ant
igen (PSMA), Hybritech Antibody (399) to PSA, and cytokeratin 8 and 18
. In addition to the antibodies described, a DNA stain To-Pro 3 was us
ed to identify 2n-4n DNA containing cells. A dual laser, Becton Dickin
son FACSCaliber cytometer, was used to analyze the samples. RESULTS. A
ll semen specimens contained diploid, cytokeratin 18-positive epitheli
al cells regardless of disease status. A clear difference between pros
tate cancer and normal prostate cell samples was observed using staini
ng with 7E11.C5. The ratio of prostatic cells in the total epithelial
cell population (PSMA:cytokeratin ratios) was calculated for each spec
imen. A retrospective study of sixteen semen samples from 11 prostate
cancer patients had a mean PSMA:cytokeratin ratio of 0.57, whereas the
samples from 15 patients without evidence of cancer had a mean PSMA:c
ytokeratin ratio of 0.11. This difference was significant. PSA stainin
g was variable and inconsistent. CONCLUSIONS. This report demonstrates
that human semen contains prostate cells that can be characterized an
d used in the clinical diagnosis of prostate cancer. Prostate 36:181-1
88, 1998. (C) 1998 Wiley-Liss, Inc.