Objective To determine whether the free/total prostate-specific antige
n (PSA) ratio can discriminate between patients with prostate cancer o
r benign prostatic hyperplasia (BPH). Patients and methods A prospecti
ve study was conducted using free and total PSA assays in patients who
underwent transrectal-ultrasound guided biopsies indicated by a total
serum PSA level of >4 ng/mL and/or a positive digital rectal examinat
ion. Sixty-nine men (median age 68 years, range 57-86) who presented t
o our out-patient department with symptoms of prostatism were included
in the study. Blood samples were drawn from all patients before biops
y. Results Histopathological examination detected prostate cancer in 1
7 of 69 (25%) patients and 13 of these 17 patients had a free/total PS
A ratio of <0.15; only 12 of 52 (23%) patients with BPH had a ratio of
<0.15. Receiver operating characteristic analysis indicated a thresho
ld free/total PSA ratio of less than or equal to 0.15 was the optimum
discriminatory level. In the whole study group, this threshold had sen
sitivity, specificity, positive- and negative-predictive values of 76%
, 77%, 52% and 91%, respectively, There were 40 patients with serum PS
A levels of 4-10 ng/mL and 17.5% (7/40) of these were diagnosed with c
ancer. Using a free/total PSA ratio of 0.15 would have failed to diagn
ose two patients of seven with prostate cancer but 30 patients would h
ave avoided a biopsy. In this subgroup, the threshold ratio of 0.15 ha
d sensitivity, specificity, positive- and negative-predictive values o
f 71%, 85%, 50% and 93%, respectively. The rates for a PSA density (PS
AD) at a threshold of greater than or equal to 0.15 were 71%, 76%, 38%
, 93%, respectively. Conclusion These results indicate that using the
free/total PSA ratio gives a significant improvement over PSAD and tot
al PSA values alone in the diagnosis of prostate cancer; its use may a
lso enhance the diagnostic accuracy in patients with intermediate PSA
levels.