| Japanese Journal of Clinical Oncology | Pages |
Significance of Free to Total PSA Ratio in Men with Slightly Elevated Serum PSA Levels: A Cooperative Study
Introduction
Materials And Methods
Results
Stability of Total and Free PSA Determination
The Suitable Cut-off Value for Total PSA Values Less than 20 ng/ml
Relationship Between Total PSA and f/t
Diagnostic Accuracy of Total PSA and f/t Using Various PSA Ranges
Suitable Cut-off f/t Values
PPV of f/t in the Total PSA Range of 5.1-10 ng/ml (Table 5)
Discussion
References
Significance of Free to Total PSA Ratio in Men with Slightly Elevated Serum PSA Levels: A Cooperative Study
Background: The ratio of free PSA in total PSA (f/t) has been reported to improve the diagnostic accuracy of prostate cancer in the group with slightly elevated serum PSA values. In Japanese cases, the clinical significance of f/t is still controversial.
Methods: The diagnostic significance of f/t in serum for prostate cancer was evaluated in a cooperative study. A total of 77 cases with prostate cancer and 224 with non-prostate cancer showing less than 20 ng/ml of total PSA were evaluated.
Results: Serum total and free PSA values were not affected by storage at 25°C for 2 days. The determination of f/t was useful in the cases with a serum total PSA of 5.1-10 ng/ml; the specificity was 60% with a sensitivity of 90% at an f/t of 0.148. The positive predictive value for diagnosis of prostate cancer also increased to 54% from 34% of that in total PSA alone. In the range of 4.1-10 ng/ml, the cut-off value of f/t was 0.155 for obtaining relatively high specificity; sensitivity was 85% and specificity was 56.5%.
Conclusions: Thus, the determination of f/t was considered to be an effective tool for discriminating the non-prostate cancer cases from those of prostate cancer.
INTRODUCTION
The determination of serum prostate-specific antigen (PSA) has become an essential tumor marker for diagnosis, evaluation of the treatment and follow-up of the patients with prostate cancer (CaP) (1,2). Due to the biological nature of PSA, its serum values may be elevated even in benign prostatic diseases such as acute prostatitis and benign prostatic hyperplasia (BPH). Therefore, the diagnosis of CaP is not accurate for the cases showing slightly elevated serum PSA values. Currently, clinicians select the patients for core-needle biopsy after a combination of digital rectal examination (DRE) and transrectal ultrasonography (TRUS) with PSA determination.
Recently the ratio of free PSA in total PSA (f/t) has been reported to improve the diagnostic accuracy in the group with slightly elevated serum PSA values (3-7). However, especially in Japanese cases, the clinical significance of f/t is still controversial. We studied the clinical significance of f/t in the diagnosis of CaP in a cooperative research study on a relatively large number of cases.
MATERIALS AND METHODS
The stability of total and free PSA determination in various stock conditions, time of serum separation, the effect of temperature and time of freezing and thawing was studied first.
The cases studied had a serum total PSA value of less than 20 ng/ml and histologically confirmed the diagnosis of CaP and non-prostate cancer (non-CaP) were chosen. A total of 301 cases including 77 CaP and 224 non-CaP were enrolled in this study. The clinical stages of CaP cases were stage A in 3, B in 53, C in 11, D in 6, and unclear stage in 4.
Serum samples were obtained before treatment and stored at -80°C until assay. Total PSA values were determined with Immulyze HS PSA® (Nippon DPC Co. Ltd, Chiba, Japan, detection sensitivity 0.003 ng/ml) and free PSA values were also detected by Immulyze Free PSA® (Nippon DPC, detection sensitivity 0.05 ng/ml). The Immulyze HS PSA-kit® used in this study gave similar results to those reported by Tandem-R PSA. The linear regression line and correlation coefficient reported for 245 cases with PSA values below 20 ng/ml are y = -0.111+1.063x and r = 0.991 (y, Immulyze HS PSA: x, Tandem-R PSA) (8). The upper cut-off level was considered to be 4.0 ng/ml as to that in Tandem-R PSA (8). Furthermore, this Immulyze HS PSA is said to be an equimolar assay for detection of both free PSA and PSA-[alpha]1antichymotrypsin (PSA-ACT) complex (9).
The clinical utilities were evaluated by sensitivity (a/a+c×100), specificity (d/b+d×100), accuracy (a+d/a+b+c+d×100), and positive predictive value (PPV, a/a+b×100), where a is CaP in positive test, b is non-CaP in positive test, c is CaP in negative test and d is non-CaP in negative test. The U-test of the Mann-Whitney and Dunn method (10) was used for statistical analysis. The detection of area under curve (AUC) and its statistics on receiver operating characteristics (ROC) analysis were obtained according to the methods of Hanley and McNeil (11,12). Difference with a P value of less than 0.05 was regarded as significant.
RESULTS
Stability of Total and Free PSA Determination
To study the stability of total and free PSA, whole blood was kept at 25°C for 8 h, then serum was separated and total and free PSA were detected according to the manufacturer's procedures (8,9). As shown in Table 1, no marked changes were observed in total and free PSA determination on the time of serum separation. There were no marked differences in values either in total or free PSA after storage of sera at 25°C, 4°C and -20°C except for total and free PSA values after 5 and 9 days at 25°C. However, f/t values were not affected by storage because both values decreased in parallel (Table 1). Moreover, repeated freezing and thawing had no effect up to 10 times (Table 1).
Table 1. Stabilities of total PSA and free PSA determination in various conditions
(a) Whole blood after blood-taking at 25°C
Time (h)
Sample 1
Sample 2
Total PSA
(ng/ml)Free PSA
(ng/ml)f/t
Total PSA
(ng/ml)Free PSA
(ng/ml)f/t
0
0.920
0.35
0.380
0.711
0.37
0.520
1
0.909
0.33
0.363
0.703
0.36
0.512
3
0.958
0.32
0.334
0.666
0.34
0.511
5
0.911
0.33
0.362
0.677
0.35
0.517
8
0.951
0.32
0.336
0.720
0.33
0.458
(b) Serum and stocked at various conditions
Condition
Time
(days)Sample 1
Sample 2
Total PSA
(ng/ml)Free PSA
(ng/ml)f/t
Total PSA
(ng/ml)Free PSA
(ng/ml)f/t
at 4°C
0
9.92
0.40
0.040
4.35
0.51
0.117
2
10.9
0.40
0.037
4.43
0.43
0.097
5
10.7
0.40
0.037
4.42
0.46
0.104
9
10.3
0.38
0.037
4.43
0.50
0.113
at 25°C
0
11.9
3.6
0.303
4.35
0.51
0.117
2
12.5
3.6
0.288
4.43
0.55
0.124
5
10.1
2.8
0.277
3.99
0.45
0.113
9
8.00
2.4
0.300
3.76
0.41
0.109
at -20°C
0
10.8
3.7
0.343
5.60
0.25
0.045
2
11.9
3.6
0.303
5.39
0.24
0.045
4
11.2
3.9
0.348
5.36
0.26
0.049
(c) The effect of freezing and thawing
Time
Sample 1
Sample 2
Total PSA
(ng/ml)Free PSA
(ng/ml)f/t
Total PSA
(ng/ml)Free PSA
(ng/ml)f/t
1
11.9
3.6
0.303
5.39
0.24
0.045
10
11.8
3.4
0.288
5.55
0.23
0.041
The Suitable Cut-off Value for Total PSA Values Less than 20 ng/ml
ROC analysis was performed using the cases showing a total PSA of less than 20 ng/ml. At the cut-off value of 5.14 ng/ml, sensitivity, specificity, accuracy, and AUC were 77.9% (60/77), 61.6% (138/224), 65.6% (198/301), and 0.7413, respectively.
Relationship Between Total PSA and f/t
Figure 1 shows a scattergram of f/t with total PSA levels. As serum PSA values increased, f/t showed a tendency to decrease both in CaP and non-CaP groups. In general, f/t was lower in the CaP group than in the non-CaP group. However, f/t values were higher in cases with a total PSA of less than 5.0 ng/ml in group CaP than in those with a higher total PSA value.
Figure 1. Scattergram of free to total PSA ratio with total PSA levels for patients with prostate cancer and non-prostate cancer. Open circle means non-prostate cancer and closed symbols and x-mark express prostate cancer and each stage. In the prostate cancer group, there was a tendency of f/t increasing their values in less than 5 ng/ml of total PSA. Table 2. 
Disease and stages
No. of cases
Total PSA (ng/ml)
f/t ratio
mean±SD
median
mean±SD
median
non-prostate cancer
178
6.51±3.92
5.01
0.18±0.083
0.18
prostate cancer
76
9.26±4.64
8.29
0.12±0.084
0.088
stage A
3
4.01±1.08
3.75
0.23±0.15
0.17
stage B
53
8.81±4.13
7.90
0.11±0.076
0.074
stage C
10
9.71±5.83
10.2
0.13±0.096
0.093
stage D
6
14.8±4.96
16.7
0.12±0.064
0.080
Serum total PSA values and f/t showed significant differences between group CaP and group non-CaP (P<0.001 in each). Total PSA values increased according to progress in clinical stage of CaP and even in the range of 2.1-20 ng/ml. Total PSA showed a statistically significant difference between stages A and D (P<0.01). There was no correlation between clinical stages of CaP and f/t (Table 2).
Using AUC on ROC analysis, the diagnostic accuracy of total PSA and f/t was compared in various total PSA ranges (Table 3). In the total PSA ranges of 2.1-20, 2.1-10, and 4.1-10 ng/ml, the AUCs for f/t showed the tendency to be higher than those for total PSA. The range of 5.1-10 ng/ml showed the best statistical data (P=0.003). Whereas, in the total PSA ranges of 2.1-4.0 and 2.1-5.0 ng/ml, the AUCs for total PSA were better results than those for f/t. The AUCs for free PSA showed the lowest value among the three parameters in all the ranges examined. Table 3. Diagnostic Accuracy of Total PSA and f/t Using Various PSA Ranges
Total PSA (ng/ml)
2.1-4.0
2.1-5.0
2.1-10
2.1-20
4.1-10
5.1-10
No. of cases
non-prostate cancer
60
89
145
178
85
56
prostate cancer
10
16
46
76
36
29
AUC
f/t
0.5311
0.5329
0.7194
0.7518
0.7453
0.8251
total PSA
0.6150
0.5575
0.6523
0.6864
0.5985
0.5413
free PSA
0.4642
0.5046
0.3630
0.4276
0.2820
0.2004
P value (f/t vs total)
0.274
0.409
0.088
0.065
0.020
0.003
Table 4.
| Sensitivity (%) |
Total PSA (ng/ml) | ||||||||||||||
| 2.1-5.0 | 2.1-10 | 2.1-20 | 4.1-10 | 5.1-1.0 | |||||||||||
| f/t | specificity | accuracy | f/t | specificity | accuracy | f/t | specificity | accuracy | f/t | specificity | accuracy | f/t | specificity | accuracy | |
| 100 | 0.408 | 2.2% | 51.1% | 0.408 | 1.3% | 50.7% | 0.471 | 1.1% | 50.6% | 0.312 | 4.7% | 52.4% | 0.312 | 5.4% | 52.7% |
| 95 | 0.402 | 2.2 | 48.6 | 0.274 | 11.0 | 53.0 | 0.280 | 10.1 | 52.6 | 0.255 | 5.9 | 50.5 | 0.259 | 7.1 | 51.5 |
| 90 | 0.278 | 13.5 | 51.8 | 0.255 | 13.1 | 51.6 | 0.252 | 12.4 | 51.2 | 0.228 | 18.8 | 54.4 | 0.148* | 60.7 | 75.5 |
| 85 | 0.259 | 16.9 | 51.0 | 0.243 | 15.2 | 50.1 | 0.193 | 37.4 | 61.1 | 0.155* | 56.5 | 70.1 | 0.144 | 60.7 | 72.4 |
| 80 | 0.252 | 16.9 | 48.5 | 0.191 | 40.1 | 60.1 | 0.155* | 59.6 | 69.8 | 0.149 | 62.4 | 71.2 | 0.136 | 62.5 | 71.3 |
| 75 | 0.247 | 18.0 | 46.5 | 0.155* | 62.1 | 68.6 | 0.144 | 66.9 | 71.0 | 0.146 | 63.5 | 69.3 | 0.102 | 85.7 | 80.4 |
Suitable Cut-off f/t Values
When the sensitivities are stepwisely set from 100 to 75%, f/t values, specificity, and accuracy were studied in various total PSA ranges (Table 4). The specificity and accuracy leveled off at an f/t of 0.15. The cut-off value for obtaining 90% sensitivity, in the total PSA range of 2.1-20 ng/ml, was 0.252 and specificity was only 12.4%. In the same manner, the total PSA range of 4.1-10 ng/ml, at a cut-off f/t value of 0.228, specificity was 18.8%. Whereas, in the range of 5.1-10 ng/ml, at a cut-off f/t value of 0.148 of f/t, specificity was 60.7% and accuracy was 75.5%. To obtain the same specificity in the range of 4.1-10 ng/ml, sensitivity was reduced to 85%, the cut-off value was 0.155 with a specificity of 56.5% and accuracy of 70.1%. The ROC curves for f/t, total PSA, and free PSA in the total PSA range of 5.1-10 ng/ml are presented in Fig. 2.
Figure 2. ROC curves of total PSA, free PSA and free to total PSA ratio in the cases showing total PSA ranging from 5.1 to 10 ng/ml. The symbols of -O-, -(- and -X- mean the ROC curve of f/t, total PSA and free PSA, respectively. At 0.148 of f/t, 60.7% of specificity was obtained with maintaining 93.1% of detection sensitivity.
PPV of f/t in the Total PSA Range of 5.1-10 ng/ml (Table 5)
A total of 85 cases showing a total PSA of 5.1-10 ng/ml was evaluated. Judging from the total PSA values, biopsy would be considered necessary in all 85 cases and 29 cases with CaP would be diagnosed, PPV being 34.1%. If we used the f/t values, the number of cases requiring a biopsy would be reduced to 50 cases (reduction rate 41%) and 27 patients would be diagnosed with CaP, PPV being 54.0%, and the resting 2 CaP cases would be misdiagnosed. However, 33 from 56 cases with non-prostate cancer could be avoided from unnecessary biopsy.
DISCUSSION
Screening for CaP with the use of serum PSA determination has shown some limitations (1,2). Because of the biological properties of PSA such as its prostate-tissue specific protease, serum PSA values in benign prostatic diseases may be slightly elevated (2). As one of the tools for discriminating between CaP and benign prostatic diseases in patients showing a gray zone PSA level, the detection of free PSA and estimation of f/t have been widely used for improving the diagnostic accuracy especially for increasing specificity (3-7). Development of an equimolar type of PSA assay has also been an achievement for obtaining true total PSA values.
There have been a few reports about the ratio of f/t changes at various stock conditions. The assay methods for total and free PSA used in this study gave stable results except when the serum samples had been stored at 25°C for more than 2 days. Repeated freezing and thawing also had no effect on the serum values of total and free PSA up to 10 times. Furthermore, this Immulyze HS PSA® is reported to be an equimolar assay of PSA (9). Thus these systems for total and free PSA showed high reliability and rationality for evaluation of f/t in CaP diagnosis.
Table 5.
| PPV (%) | ||
| Total PSA | 34.1 | (29/85 cases) |
| f/t ratio | 54.0 | (27/50 cases) |
The f/t values both in CaP and non-CaP were not distributed uniformly when the total PSA was less than 20 ng/ml. Although the total PSA values increased according to the progression of the clinical stage of CaP, such phenomenon was not observed in f/t. Thus the f/t value may be useful for distinguishing CaP from non-CaP cases, but not for prediction of clinical stage.
Our studies clarified that the f/t value gives a higher specificity than the total PSA. In the range of 5.1-10 ng/ml, 90% sensitivity and 60% specificity were obtained, whereas, in the range of 4.1-10 ng/ml, sensitivity decreased to 85% while maintaining 56.5% specificity. This difference occurs mainly because of low specificity of f/t in the range 4.1-5.0 ng/ml of total PSA. The attempts for increasing specificity in a group of the lower levels of PSA will be necessary in the future.
The optimal cut-off values for diagnosis of CaP vary with the reports, mainly due to the differences in the number of cases examined, the objective total PSA range, total and free PSA assay systems. Catalona et al. recommended an f/t of 0.234 as a cut-off value in the total PSA range of 4.1-10 ng/ml (4). With the use of Hybritech systems, they obtained 90% sensitivity and 31.3% specificity. In another study (13), they changed the cut-off f/t value to 0.27 for the total PSA range 2.6-4.0 ng/ml and obtained 18% specificity with 90% sensitivity. They considered this 18% reduction meaningful for avoiding unnecessary biopsy for the total PSA lower than the gray zone. The specificity of 60% and sensitivity of 90% may be sufficient for application obtained at a total PSA above 5.0 ng/ml. The cut-off f/t value of 0.15 is similar to that established in Europe (14).
Conclusively, the detection of free PSA and calculation of f/t were useful for distinguishing prostate cancer from non-prostate cancer. Furthermore, this method which is simple to use, objective, and requires no special equipment is recommended for mass screening of CaP.
References
Received May 8, 1998; accepted July 29, 1998
This page is run by Oxford University Press, Great Clarendon Street, Oxford OX2 6DP, as part of the OUP Journals
Comments and feedback: www-admin{at}oup.co.uk
Last modification: 24 Nov 1998
Copyright©Japanese Journal of Clinical Oncology, 1998.
CiteULike 
Connotea 
Del.icio.us What's this?
This Article 
Abstract
Alert me when this article is cited
Alert me if a correction is posted
Services
Email this article to a friend
Similar articles in this journal
Similar articles in ISI Web of Science
Similar articles in PubMed
Alert me to new issues of the journal
Add to My Personal Archive
Download to citation manager
Search for citing articles in:
ISI Web of Science (12)
Request Permissions
Google Scholar
Articles by Kuriyama, M
Articles by Yamanaka, H
Search for Related Content
PubMed
PubMed Citation
Articles by Kuriyama, M
Articles by Yamanaka, H
Social Bookmarking
What's this?