Free PSA and prostate volume on the diagnosis of prostate carcinoma 1

Objective: To analyse the influence of prostate volume on the performance of total prostate specific antigen (tPSA) and free PSA (fPSA) on the diagnosis of prostate adenocarcinoma. Methods: A total of 188 patients underwent transrectal ultrasound guided biopsies (10-12 cores) due to prostate nodes detected by digital rectal examination and/or tPSA range of 2.5-10ng/ml. Mean age was 65.7±8.7 years. 19/100 (19%)(GI) patients with prostate volume >40ml had prostate cancer while the corresponding figure for patients with prostate <40ml was 26/88 (29.5%)(GII). We analyzed the sensitivity and specificity of tPSA at cut-off points of 2.5 and 4ng/ml as well as the influence of the ratio f/tPSA in both groups of patients. Results: In the group GI tPSA sensitivity and specificity were 94.4% and 19.5% at the cut-off level of 4ng/ml and 100% and 6% at 2.5ng/ml. The corresponding values for GII were 76.5% and 62.9%, and 100% and 19.3%. In group GI a cut-off of 19% for the ratio f/tPSA kept tPSA sensitivity over 90% while the specificity increased to 46.2% at cut-off level of 4ng/ml and to 32.9% at 2.5ng/ml. In the group GII the ratio f/tPSA was not able to increase the specificity of tPSA at a cut-off level of 4ng/ml without an expressive reduction of sensitivity. On the other side, for this group a cutoff of 16% for the f/tPSA ratio rose the specificity to 46.7% for a sensitivity over 90%. Conclusion: We recommend stratification of patients according to prostate volume to define tPSA cut-off point. The cut-off level of 2.5ng/ml for tPSA combined with f/tPSA ratio of 19% in prostates >40ml and 16% in prostates <40ml was a better option for prostate biopsy indication than tPSA at a cutoff of 4ng/ml associated or not with f/tPSA ratio.


INTRODUCTION
The prostate specific antigen (PSA) is a kallicrein controlled by a gene on cromossome 19.It is a glicoprotein produced by the prostate epithelium which fuction is to promote semen liquefation 1,2 .In the plasma PSA circulates free or complexed with proteins: a1antichemotrypsin (ACT) and a2-macroglobulin (MG).There are several types of assay to determine PSA level in the serum and the most common are able to measure the total PSA (tPSA) and the free PSA (fPSA).
Approximately 30% of the patients with tPSA between 4-10ng/ml bear prostate adenocarcinoma while 20% of tumors occur in patients with tPSA <4ng/ml 2,3,4,5 .In 1997, Catalona et al. 6 reported that 22% of patients with tPSA between 2.6 to 4ng/ml have prostate adenocarcinoma and that 80% of these tumors are confined into the gland.The digital rectal examination reveals no suspiction of prostate carcinoma in 96% of patients with tPSA between 2.5-4ng/ml 6,7 .Thus, there is a concern about the better tPSA cut-off level for biopsy indication that should be the one with the highest sensitivy and specificity.The use of a cut-off level of 2.5ng/ml is associated with a low specificity which could be increased by the combination of tPSA with f/tPSA ratio 7 .
As benign prostate hyperplasia (BPH) increases tPSA level the prostate volume may affect the performance of the test, and for such reason the tPSA density might improve the specificity of the test without affecting its sensitivity 8 .On the other hand, there are reports suggesting that f/ tPSA ratio is also influenced by prostate volume 9,10 .
The aim of our study is to investigate the influence of prostate volume on tPSA performance and explore whether f/tPSA can enhance the test specificity.

METHODS
From February of 2002 to March of 2003, 188 men underwent transrectal ultrasound guided biopsies (10-12 cores) due to prostatic nodes detected by digital rectal examination and/or serum tPSA of 2.5 -10ng/ml.The prostate volume was determined by transrectal ultrasound during the biopsy procedure.The mean age of the patients was 65.7±8.7 years.Of 100 men with prostates >40ml (GI), 19 (19%) had cancer.The proportion of prostate adenocarcinoma in men with prostate <40ml (GII) was 26/88 (29.5%).We analyzed the tPSA sensitivity and specificity at cut-off levels of 2.5 and 4.0ng/ml in both groups of patients and studied the influence of f/tPSA ratio on such parameters.The dosage of PSA was undertaken with kits from DPC-Immulite TM .
The comparison of prevalence of carcinoma in prostates larger or smaller than 40ml according to tPSA range was performed by the two tail Fisher's exact test using a software Instat, version 3.0.The level of significance considered was 5%.
The results of tPSA sensitivity and specificity are exhibited in Table 1.The level of 19% for f/tPSA was determined through the receiver operating curve as the best cut-off point for all groups, except for the GII at a tPSA cut-off level of 2.5ng/ml in which the best cut-off point was 16%.

DISCUSSION
Our data show that the prevalence of tumor in patients with tPSA between 2.5-4ng/ml is 17.3% which is quite impressive even though a bit lower than reported elsewhere 8,11 .In fact in our series such a proportion was similar to that observed in cases with tPSA of 4-10ng/ml (p=0.13).
The sensitivity and specificity of tPSA at a cutoff of 4ng/ml for the entire sample are within the range published elsewhere 3,4,12 .However, it is important to stress that at this cut-off point a large proportion of tumors is missed inasmuch as if one associates the f/tPSA ratio under 19% to make a decision to indicate or not the biopsy.usually higher in benign prostates >40ml which affects tPSA density as well as the ratio of freeto-total PSA 9,12,13 .
The sensitivity of tPSA is high and the specificity is low in both cut-off levels for prostates >40ml.A cut-off point of 19% for f/tPSA ratio was able to enhance expressively the specificity of the test in both cut-off levels in this group of patients (GI).It is worth to mention, however, that the adoption of this additional parameter to indicate the biopsy increases the proportion of missed tumor to 5.6% or 11.1% respectively for the tPSA cut-off of 2.5 or 4.0ng/ml.Some authors proposed a cut-off point of tPSA that increases with age range to avoid unnecessary biopsies 14,15 .As a matter of fact, such proposal reflects indirectely the prostate growth with aging in consequence of BPH.But, it is known that nearly 15 to 20% of elderly men have prostates of normal size, which makes the proposal imprecise.On the other side, there are men in the fifties with larger prostates than the usual.In our series, adenocarcinoma occurred in 45.2% of the prostates smaller than 40ml and tPSA higher than 4ng/ml.It is unwise to avoid biopsies in such patients.Perhaps, patients should be stratified according to prostate volume to define tPSA cut-off point instead of age-range.For men with prostates <40ml a tPSA cut-off of 2.5ng/ml combined with a f/tPSA ratio of 16% represented a better option than the cut-off of 4ng/ml.It is worth to mention that the most convincing recent report, involving the screening and follow-up of 12,902 men, showed that the 95% percentile of tPSA was 2.45ng/ml for men without adenocarcinoma, and for this reason such value was proposed as the upper limit of normality 16 .

CONCLUSION
We recommend stratification of patients according to prostate volume to define tPSA cut-off point.The cut-off level of 2.5ng/ml for tPSA combined with f/tPSA ratio of 19% in prostates >40ml and 16% in prostates <40ml was a better option for prostate biopsy indication than tPSA at a cut-off of 4ng/ml associated or not with f/tPSA ratio.

PATIENTS
As reported previously 6,7,12 a tPSA cut-off of 2.5ng/ml for all patients exhibits a low specificity which means a low yield prostatic biopsies.It is also interesting to notice in Table 1 that prostate volume affect the test performance in both cut-off levels as described by others 12,13 .A possible reason to explain such influence is the lack of a regular pattern of tissue composition seen in BPH which is responsible for the increase of prostate volume.The proportion of glands is

TABLE 1 :
Performance of tPSA associated of not to f/tPSA ratio.Ss -sensitivity, Sp -specificity, * -cut-off f/tPSA at 16%, B/T -number of biopsies per tumor, MT -missed tumor