The effect of percentage free prostate-specific antigen (PSA) level on the prostate cancer detection rate in a screening population with low PSA levels

Authors


Alexandre E. Pelzer, Department of Urology, Innsbruck Medical University, Anichstrasse 35, A-6020 Innsbruck, Austria.
e-mail: alexandre.pelzer@uibk.ac.at

Abstract

OBJECTIVE

To evaluate the prostate cancer detection rate at low total prostate-specific antigen (tPSA) ranges of 2.6–4 and 4.1–10 ng/mL, according to different percentage free (f/t) PSA levels in a screening population.

SUBJECTS AND METHODS

In all, 1809 consecutive screening volunteers with a tPSA level of 2.6–10.0 ng/mL were assessed. Ten systematic ultrasonography-guided prostate biopsies and, since 2000, an additional five Doppler-enhanced targeted biopsies were taken on the basis of age-specific tPSA reference ranges. We analysed the detection rate of prostate cancer according to f/tPSA ranges of 0–9%, 10–14%, 15–18% and >18%.

RESULTS

The detection rates for the subgroups with tPSA levels of 2.6–4.0 and 4.1–10.0 ng/mL were 20.2% and 27.0%, respectively. The cancer detection rate in the first group (2.6–4.0 ng/mL) at 0–10% fPSA was 22.9%, and that in the second group (4.1–10.0 ng/mL) at 0–10% was 36.9%. There were significant differences between these groups. If the f/tPSA was 10–15%, the cancer detection rate for the two groups were 22.6% and 32.5%, respectively (P < 0.05). There was no statistically significant difference in the cancer detecting rates at an f/tPSA of 15–18% or >18%.

CONCLUSION

There is a statistically significantly higher cancer detection rate when the f/tPSA is <15% than in groups of men with a f/tPSA of >15% in screening population assessed primarily using tPSA level.

Abbreviations
f

t, f/tPSA, free, total, free/total PSA

DET

Doppler-enhanced targeted (biopsies).

INTRODUCTION

The prostate remains one of the leading sites of internal malignancy and the second most common cause of cancer death in men. Prostate cancer is the fourth most common male malignancy worldwide. The measurement of PSA in serum is the most widely used and efficient method for the early detection of prostate cancer [1]. While PSA levels of >10 ng/mL provide sufficient sensitivity to detect cancer, most cancers detected in this range are pathologically advanced [2]. Most men (>80%) with a high PSA value have serum levels of 4–10 ng/mL but only 25% of men with an unremarkable DRE have prostate cancer [3]. In these men the most likely reason for the high PSA value is prostate enlargement, not cancer, because of the high prevalence of BPH in the population. Before PSA screening was available most prostate cancers were identified at a stage where the possibility of cure was limited. Currently, with the widespread use of PSA screening, more cancers are identified at an earlier stage, when they can be treated effectively with surgical or nonsurgical approaches. As in the USA, the incidence of prostate cancer has increased in many countries since the early 1990s. Although much of this increase can be correlated with the introduction of widespread PSA testing, some of the increase pre-dates cancer screening. At present, most patients with prostate cancer and mildly elevated PSA levels are diagnosed with early-stage, organ-confined disease, and more than half of men with PSA levels of >10 ng/mL have advanced, extraprostatic disease. With the increasing use of PSA as a screening tool to detect cancer more recent research has focused on men with a PSA level of <4 ng/mL and a negative DRE. However, data reported so far indicate that biopsy-detected cancer, including high-grade cancers, is not rare among men with PSA levels of ≤ 4.0 ng/mL, and the incidence and histological grading of cancer at a PSA level of 2.5–4.0 ng/mL seems to be comparable to that at 4–10 ng/mL [4]. Other investigations indicate that the measurement of the ratio between free (fPSA) and total PSA (tPSA) in serum may serve to differentiate more clearly between BPH and cancer in some patients [5–7].

The percentage fPSA/tPSA (f/tPSA) is the proportion of fPSA present in serum tPSA. Most of the circulating PSA is complexed with the serine protease inhibitor α1-antichymotrypsin or α2-macroglobulin. This proportion of fPSA is significantly lower in patients with prostate cancer than in those with BPH or prostatitis [8]. Therefore, f/tPSA has also been applied to the intermediate tPSA range, and enhances the specificity of cancer detection in the low PSA range. When a threshold of 25% was used as a criterion for prostate biopsy, 90% of cancers would have been detected while avoiding 18% of biopsies, and Catalona et al.[9] concluded that f/tPSA is an independent predictor of cancer within a tPSA level of 4–10 ng/mL. In another analysis, using a threshold of ≤ 18% saved 45.5% of biopsies with a cancer detection rate of 100%[10]. With the studies from Catalona et al. in mind, the aim of the present study was to evaluate the cancer detection rate at low tPSA ranges of 2.6–4.0 and 4.1–10 ng/mL at different f/tPSA levels.

SUBJECTS AND METHODS

Between 1995 and 2001, 1809 consecutive screening volunteers with a tPSA level of 2.6–10.0 ng/mL were assessed. A DRE was not part of the screening protocol, but was used before biopsy. Exclusion criteria were clinical prostatitis within a month of biopsy, active UTI or contraindications for the ultrasonography contrast agent SonoVue (Bracco, Milan, Italy), including after acute myocardial infarction and cardiac insufficiency (grade IV, New York criteria). After obtaining informed consent, the night before biopsy all participants began a 5-day course of therapy with a fluoroquinolone antibiotic or an appropriate alternative antibiotic if there was a fluoroquinolone allergy. A cleansing enema was administered on the morning of the biopsy. The men were instructed not to ingest aspirin or other anticoagulating agents for ≥ 10 days before biopsy. Ten systematic TRUS-guided prostate biopsies in a standard spatial distribution, and since 2000 an additional five Doppler-enhanced targeted (DET) biopsies, were taken on the basis of age-specific tPSA reference ranges. These reference ranges were defined as a half the age-reference PSA levels reported by Oesterling et al.[11] combined with a f/tPSA of <18% (Table 1). Men with PSA levels of >10 ng/mL were advised to undergo biopsy irrespective of their f/tPSA levels. Men with f/tPSA levels of >18% were mostly were those with abnormal DRE findings.

Table 1. 
Bisected, age-specific reference ranges for tPSA levels. Subjects with tPSA levels of >10 ng/mL were advised to undergo biopsy irrespective of their f/tPSA level
Age range, yearsPSA range, ng/mL
40–490–1.25
50–590–1.75
60–690–2.25
70–790–3.25

DET biopsies were used since 2001 and taken as described previously [12]; subsequently another investigator, unaware of the DET findings, took 10 biopsies in the standard spatial distribution. Two biopsy cores were obtained per side from the apex area, including one medial and one lateral; another biopsy core was obtained on each side from the lateral aspect of the mid prostate, one on each side from the posterolateral area at the base, most likely in the central zone, and a final core from each side of the transition zone (anterior biopsies in the mid prostate). Biopsies were obtained transrectally using an 18 G biopsy needle.

The standard biopsy was guided by TRUS, using a Combison 530MT unit (Kretztechnik, Zipf, Austria) fitted with a biplanar probe operating at a grey-scale frequency of 10 MHz and at locations as described previously [12]. Biopsy cores were reviewed by one pathologist and assessed as being cancerous, with an assigned Gleason score.

The tPSA and f/tPSA levels were assessed before any prostatic manipulation and were evaluated in our laboratory on the same day. tPSA levels were determined with the Immuno 1 PSA assay (Bayer Diagnostics, Tarrytown, NY, USA), and the f/tPSA levels with the IMx Immunoassay (Abbott Laboratories, Abbott Park, IL, USA).

The Mann–Whitney U-test was used for statistical analyses, with P < 0.05 considered to indicate statistical significance. We analysed the cancer detection rate at a tPSA of 2.6–4.0 and 4.1–10.0 ng/mL, according to f/tPSA ranges of 0–9%, 10–14%, 15–18% and >18%.

RESULTS

In all, 1809 consecutive screening volunteers with PSA levels of 2.6–10 ng/mL were assessed. All screening volunteers were Caucasian; the study population characteristics of the different tPSA and f/tPSA subgroups are shown in Table 2. The mean age of the men with tPSA level of 2.6–4.0 ng/mL was 57 years, mean tPSA 2.9 ng/mL and the mean prostate volume 30.8 mL; the mean age of men with tPSA levels of 4.1–10 ng/mL was 62.7 years, the mean tPSA level 6.2 ng/mL and the mean prostate volume 39.6 mL. The cancer detection rates for the subgroups with tPSA levels of 2.6–4.0 ng/mL (559 men) and 4.1–10.0 ng/mL (1250 men) were 20.2% and 27.0%, respectively. The cancer detection rates for the different tPSA groups according to the various f/tPSA ranges are shown in Table 3. Differences in tPSA for the first two ranges were significant (Mann-Whitney U-test, P < 0.05) but for the higher ranges of f/tPSA the differences in detection rate between the tPSA groups were not significant.

Table 2. 
The characteristics of the tPSA and f/tPSA subgroups (1809 men)
VariableMean (sd, range)
Age, years62.4 (7.9, 41–97)
tPSA, ng/mL 5.4 (1.8, 2.6–10.0)
fPSA, ng/mL 0.7 (0.4, 0.08–3.9)
f/tPSA, %14.6 (6.1, 0–56.3)
Prostate volume, mL39.4 (17.4, 8–160)
Table 3. 
The cancer detection rate for the different PSA groups according to the various f/tPSA ranges
f/tPSA, %Detection rate, % (n), at tPSA, ng/mL P
2.6–4.04.1–10.0
  1. NS, not significant.

N5591250 
<10 22.9 (128)  36.9 (461)<0.05
10–15 22.6 (126)  32.5 (406)<0.05
15–18 17.0 (95)  21.6 (270)NS
>18 12.7 (70)  12.1 (151)NS

DISCUSSION

The use of serum tPSA levels and the DRE improved the early detection rate for prostate cancer, but the use of either of these tests can give a poor prediction of prostate cancer. In the present study we retrospectively enrolled 1809 men with tPSA levels of 2.6–4.0 and 4.1–10.0 ng/mL, and with or with no suspicious DRE. The aim of the study was to evaluate the cancer detection rate at low and intermediate tPSA levels and at different f/tPSA levels.

There are few reports on the effect of f/tPSA on detecting prostate cancer in men with a tPSA of <4 ng/mL. The utility of f/tPSA at a tPSA level of 4.0–10 ng/mL is widely accepted, considering its significant increase (20%) in specificity compared to using only tPSA. Previous studies showed that f/tPSA is an independent predictor for detecting prostate cancer in the initial biopsy [7,13]. When taking a prostate biopsy, regardless of patient age and prostate size, these authors recommended a threshold for f/tPSA of 25%, which is in the ‘grey zone’ of tPSA (4.0–10 ng/mL). The results showed that a lower f/tPSA was associated with a greater risk of cancer [5]. Although other studies suggested the use of several PSA-related variables, e.g. PSA density, PSA velocity, and age-specific PSA reference ranges, f/tPSA was the most useful indicator for taking a prostate biopsy [14–16]. Some studies recommend that in men with a tPSA level in the grey zone, then a f/tPSA of 20% should be the threshold to differentiate between those with prostate cancer and those with BPH. Carlson et al.[17] concluded that the lower limit for the tPSA range should be 4.0 ng/mL and that the lowest tPSA value at which f/tPSA both differed between benign and malignant cases and predicted prostate cancer was 4.0 ng/mL. In that study nonparametric tests showed that when the tPSA level was ≥ 3.9 ng/mL, the f/tPSA distributions for benign and malignant cases were significantly different. On the basis of logistic regression analysis, at a tPSA level of 3.9 ng/mL, the f/tPSA could not predict prostate cancer, but at 4.0 ng/mL it could. Some limitations of that comparable study were that the prostate cancer detection was not compared with an intermediate tPSA group, and that cancer was detected using a sextant biopsy only. With biopsy strategies with >10 cores recently becoming the standard of care in cancer detection, the sensitivity of the confirmatory test has increased, the characteristics of the cancer identified have changed, and the standard sextant biopsy protocol is currently considered insufficient [18–20]. Studies show that one set of six biopsies may miss clinically detectable prostate cancer in 15–34% of men [21,22]. In the present study we used a systematic 10-core TRUS-guided biopsy system with a standard spatial distribution, and since 2000 have taken a further five DET biopsies [12]. Especially when combined with a contrast-enhancing agent, Doppler ultrasonography is a reliable, sensitive and noninvasive method to show tumour blood flow, and therefore it has an important role in diagnostic ultrasonography.

With this biopsy regimen and in the present subjects the overall cancer detection rate in the low and high tPSA groups was comparable with results from other studies. In both groups there was a statistically significantly higher cancer detection rate when f/tPSA was <15% than in groups of men with a f/tPSA of >15%.

Thompson et al.[23] investigated the prevalence of prostate cancer among men who had a tPSA level of ≤ 4.0 ng/mL, concluding that biopsy-detected prostate cancer, including high-grade cancer, is not rare among men with such tPSA levels. The prevalence of high-grade cancers increased from 12.5% in men with a tPSA level of ≤ 0.5 ng/mL, to 25% of cancers at tPSA level of 3.1–4.0 ng/mL. However, the extensive use of tPSA for prostate cancer screening has confronted the clinician with many more men with a benign DRE and tPSA levels of 2–4 ng/mL, and there is a need to improve the detection rate in this population.

In conclusion, there was a clear effect of f/tPSA level on cancer detection in a population initially screened using tPSA levels, and that combining the f/tPSA level with tPSA, DRE and TRUS findings could help to reduce the number of unnecessary biopsies. There was a statistically significantly higher prostate cancer detection rate when the f/tPSA was <15% than in groups of men with a f/tPSA of >15%.

CONFLICT OF INTEREST

None declared.

Ancillary