SHOULD WE USE DIFFERENT CUT OFF VALUES OF PSA DERIVATES FOR DIFFERENT RANGES OF PSA?

Introduction & Objectives: PSA ratio (PSAr) and PSA density (PSAd) are derivates of PSA
which have been introduced to improve sensibility and specifity of total PSA for values between
4 and 10 ng/ml. In this work we analyzed the predictive accuracy of PSAr and PSAd for different
ranges of PSA, and we assessed the corresponding values of their cut offs.
Material & Methods: A cohort of 1658 of not selected patients (pts) underwent transrectal
prostatic biopsy; we performed a sextant scheme biopsy plus 2/3 further cores laterally on
each lobe; in high volume prostates, number of cores was increased according to Vienna
nomogram; moreover in case of re-biopsy at least 2 cores were added in the transition zone.
Pts were stratified in 3 groups: PSA ≤4 ng/ml (group 1, n= 183), PSA 4 - 10 ng/ml (group 2,
n= 1043), PSA >10 ng/ml (group 3, n= 432). Indications for biopsy were: in group 1 familiarity
for prostate cancer, a suspicious clinical lesion (digital rectal exploration and/or hypoechoic
lesion on ultrasound) or PSA velocity > 0.75 ng/ml/year ; in group 2, a suspicious clinical lesion
(digital rectal exploration and/or hypoechoic lesion on ultrasound) or a PSAr < 20%; in group
3 the value of total PSA for itself was considered sufficient for biopsy. Differences of means of
prostate volumes were valuated with ANOVA test. A univariate analysis was constructed for
correlating PSAr and PSAd with the outcome of biopsy; receiving operating curve (ROC) was
used to test the predictive accuracy of derivates of PSA and to identify the optimal cut offs.
Results: The detection rates were 31.5%, 31% and 40% respectively in group 1, 2 and 3. Mean
prostate volume on ultrasound was significantly different among the 3 groups (40cc vs 53cc vs
64cc; p:0.001). At univariate analysis PSAd was significantly associated to a positive biopsy
among all the groups (respectively in group 1,2 and 3: p: 0.005, 0.001 and 0.001); PSAr was
significantly associated to a positive biopsy in group 2 and 3 (p: 0.001 e 0.006), but not in group
1 (p: 0.335). Area Under ROC (AUC) of PSAd was higher than that of PSAr on all the 3 groups
(respectively AUC: 0.70 vs 0.57; 0,71 vs 0.60; 0.78 vs 0.64); cut off values of PSAd were 0.10,
0.14 and 0.33 ng/ml/gr, and of PSAr were 14%, 13% e 10%, respectively in group 1,2 and 3.
Sensibility rates at the different cut off values of PSAd were higher than that of PSAr (group 1:
70 vs 52%; group 2: 74 vs 51%; group 3: 66 vs 53%), whereas specificity rates were similar (68
vs 75%; 62 vs 68%; 80 vs 77%)
Conclusions: PSAd has a higher sensibility than PSAr; PSAd could be more useful than PSAr
to give indications for biopsy, above all using different cut off values according to total PSA
ranges: this clinical approach could be recommended for pts with PSA ≤4 ng/ml, where, for the
smaller volume gland of these pts, variability of PSAd is more sensitive in case of cancer. A
value of PASd >0.10 ng/ml/gr may be used to select patient for prostate biopsy.