Urinary PCA3 to predict prostate cancer in a cohort of

Progrès en urologie (2015) 25, e1—e8
Disponible en ligne sur
ScienceDirect
www.sciencedirect.com
ORIGINAL ARTICLE
Urinary PCA3 to predict prostate cancer
in a cohort of 1015 patients
Test urinaire PCA3 et diagnostic du cancer prostatique : étude à partir de
1015 patients
V. Vlaeminck-Guillem a,∗,b, M. Devonec c,d,
D. Champetier c, M. Decaussin-Petrucci b,d,e,
P. Paparel c,d, P. Perrin c,d, A. Ruffion b,c,d
a
Unité d’oncologie moléculaire et transfert, service de biochimie et biologie moléculaire
sud, centre hospitalier Lyon-Sud, hospices civils de Lyon, chemin du Grand-Revoyet, 69495
Pierre-Bénite, France
b
Inserm 1052 CNRS 5286, centre de recherche en cancérologie de Lyon — centre Léon-Bérard,
université Lyon 1, 69373 Lyon cedex 08, France
c
Service d’urologie, centre hospitalier Lyon-Sud, hospices civils de Lyon, chemin du
Grand-Revoyet, 69495 Pierre-Bénite, France
d
Faculté de médecine Lyon-Sud, université Lyon 1, chemin du Petit-Revoyet, 69921 Oullins
cedex, France
e
Laboratoire d’anatomie et cytologie pathologiques, centre hospitalier Lyon-Sud, hospices
civils de Lyon, chemin du Grand-Revoyet, 69495 Pierre-Bénite, France
Received 2 July 2015; accepted 10 October 2015
Available online 14 November 2015
KEYWORDS
Prostate cancer;
Diagnosis;
Prostate biopsies;
Prediction;
PCA3;
Urinary test
∗
Summary
Aim. — To evaluate the performance of urinary PCA3 test to predict prostate biopsy outcome
in a large French cohort.
Patients and methods. — A urine sample was prospectively obtained from 1015 patients undergoing prostate biopsies to determine the PCA3 score. The predictive value of PCA3 was explored
using receiver operating characteristic curve analysis (ROC), multivariable logistic regression
analysis and decision curve analysis.
DOI of original article:http://dx.doi.org/10.1016/j.purol.2015.08.005.
Corresponding author.
E-mail address: [email protected] (V. Vlaeminck-Guillem).
http://dx.doi.org/10.1016/j.purol.2015.10.006
1166-7087/© 2015 Elsevier Masson SAS. All rights reserved.
e2
V. Vlaeminck-Guillem et al.
Results. — The median PCA3 score was significantly higher in patients with positive biopsies. The
PCA3 score AUC was 0.76 (0.73—0.79), significantly higher than that of PSA (0.55; 0.51—0.58). At
the cutoff of 35, sensitivity was 68%, specificity 71%, positive and negative predictive values 67%
and 71%, and accuracy 69%. Using multivariate analysis, PCA3 score appeared as an independent
predictor of biopsy outcome and its addition to a base model including usual clinicobiological
parameters resulted in a significant increase in predictive accuracy. At the cutoff of 20, about
1/2 of the ultimately unnecessary biopsies would have been avoided while ignoring 7% of cancers
with Gleason score ≥ 7. PCA3 score did not correlate with Gleason score but did correlate with
tumor volume (proportion of positive cores).
Conclusion. — Urinary PCA3 is a useful test with high diagnostic performances for early prostate
cancer diagnosis. Its correlation with cancer aggressiveness seems rather represented by a link
to prostate volume than Gleason score.
Level of evidence.— 5.
© 2015 Elsevier Masson SAS. All rights reserved.
MOTS CLÉS
Cancer de la
prostate ;
Diagnostic ;
Biopsies
prostatiques ;
Prédiction ;
PCA3 ;
Test urinaire
Résumé
But. — Évaluer la capacité du test urinaire PCA3 à prédire le résultat des biopsies prostatiques
dans une large cohorte issue du centre hospitalier Lyon Sud.
Patients et méthodes. — Le score PCA3 a été déterminé prospectivement chez 1015 patients
adressés pour biopsies prostatiques. La capacité prédictive du score PCA3 a été évaluée par la
comparaison des aires sous les courbes ROC, de modèles de régression logistique et une analyse
par « decision curve analysis » (DCA).
Résultats. — Le score PCA3 médian était significativement plus élevé chez les patients avec
biopsies positives. L’AUC était de 0,76, significativement plus élevée que celle du PSA à 0,55.
Au seuil de 35, la sensibilité était de 68 %, la spécificité de 71 %, les valeurs prédictives positive
et négative de 67 % et 71 %, et l’efficience de 69 %. En analyse multivariée, le score PCA3 était
prédicteur indépendant du résultat des biopsies et son addition à un modèle de base comportant
les données clinicobiologiques classiques apportait un gain diagnostique significatif. Au seuil de
20, près de la moitié des biopsies a posteriori inutiles auraient été évitées, tout en ayant ignoré
7 % des cancers avec score de Gleason ≥ 7. Le score PCA3 n’apparaissait pas corrélé au score
de Gleason, mais était bien corrélé au volume tumoral (évalué par la proportion de carottes
envahies).
Conclusion. — Le test urinaire PCA3 présente des performances diagnostiques élevées pour le
diagnostic précoce du CaP. Sa corrélation avec l’agressivité du cancer s’exprime à travers le
volume tumoral plus que par le score de Gleason.
Niveau de preuve.— 5.
© 2015 Elsevier Masson SAS. Tous droits réservés.
Introduction
The first urine test for the quantitative assessment of
prostate cancer gene 3 (PCA3) RNA transcripts in patients
with suspected prostate cancer (PCa) was published more
than 10 years ago [1]. The PCA3 RNA product is almost
exclusively expressed in cancerous prostate tissue. It is not
expressed in other tissues and only very weakly expressed
in healthy prostate tissue or in non-malignant prostatic diseases [1—3]. The discovery and diagnostic use of PCA3 came
about in response to the lack of specificity of serum prostatespecific antigen (PSA) testing for the diagnosis of PCa.
To date, approximately 100 studies, involving cumulatively close to 30,000 patients, have been conducted
to evaluate urinary PCA3 RNA measurement (PCA3 test)
in clinical practice [4—6]. The PCA3 test was approved
for diagnostic purposes by the United States Food and
Drug Administration in 2012, as a decisional aid for repeat
prostate biopsy in patients aged at least 50 years who have
had one or more previous negative biopsies. In France,
as of this writing, the PCA3 test is only available for
clinical research studies or in the setting of a voluntary,
patient-payed diagnostic process. Published clinical studies, variable in scientific quality, tend to show that the
urinary PCA3 test can be of significant value when deciding
whether or not to perform prostate biopsy. In contrast to the
serum PSA, which is affected by prostate volume, the PCA3
score does not increase in the presence of a non-malignant
prostatic disease [7]. It does however correlate with the
risk of positive biopsy and its use may reduce the number
Urinary PCA3 to predict prostate cancer in a cohort of 1015 patients
of ultimately-negative (and thus, a posteriori unnecessary)
biopsies by half to two-third [4—6]. Correlations are assumed
to exist between the PCA3 score and histoprognostic criteria
identified on biopsy or prostatectomy specimens.
The purpose of the present work is to report our experience with the diagnostic and prognostic performance of the
urinary PCA3 test in a large cohort of patients referred for
prostate biopsy following a suspicion of prostate cancer.
e3
scores to biopsy histological results. The contribution of
the PCA3 score in relation to existing clinical and biological
data was evaluated in multivariate (nested logistic regression models) and decision curve analyses (DCA). Calculations
were performed using Stata Statistical Software Release 11
(StataCorp, College Station, Texas, USA). A P-value < 0.05
was considered statistically significant.
Results
Patients and methods
Baseline characteristics
Patients
In the study period, 1029 patients had a urinary PCA3 test
before undergoing prostate biopsy for suspicion of PCa.
Urine samples were informative for 1015 (98.6%) of them.
The characteristics of the 14 patients with non-informative
samples did not differ from those of the other patients
except for a higher rate of 5␣-reductase inhibitor use: 21%
vs 2% (P = 0.003). Further analyses were thus performed in
the 1015 patients with informative samples (Table 1). Significant differences between the 480 (47%) patients with
positive biopsies and the 535 (53%) with negative biopsies
were, for the former: older age, higher serum PSA levels,
lower prostate volume, less frequent history of negative
biopsies, DRE more frequently suspicious (Table 1).
All patients referred to our urology department between 20
December 2007 and 5 May 2014 for prostate biopsy were
consecutively included in this prospective, single-center
study. Reasons for referral were serum PSA ≥ 4 ng/mL
and/or an abnormal digital rectal examination (DRE) and/or
a family history of PCa. Patients with a personal history of
PCa were excluded, but those with a history of one or more
negative prostate biopsies were included. One hundred and
sixty of the included patients had been included in an earlier
study [8]. All patients provided written informed consent.
After the collection of urine samples for the PCA3
test, the patients underwent a prostate biopsy series in
accordance with the recommendations of the European
Association of Urology, i.e., at least six cores taken from
each lobe under transrectal ultrasound guidance, this latter
permitting also an evaluation of prostate volume. The following anatomopathological data were recorded: Gleason
score, percentage of positive cores, percentage of tumorous
prostatic tissue, unilateral or bilateral involvement, presence or absence of perineural invasion, presence or absence
of extracapsular extension.
Urine samples and urinary PCA3 test
First-catch urine samples (25—30 mL) were collected after
standardized DRE [9]. The samples were immediately transferred to specific specimen tubes (Progensa Urine Specimen
Transport Kit, Gen-Probe), and stored at −20 ◦ C until the test
could be performed. The PCA3 test (Progensa PCA3 assay,
DTS400 system, Hologic Gen-Probe) was performed at the
hospital’s biochemistry service as per the manufacturer’s
recommendations. The PCA3 score was calculated as the
ratio of PCA3 to PSA RNA copies, multiplied by 1000. When
less than 10,000 copies of PSA mRNA were detected, the
urine sample was considered as non-informative.
Diagnostic performance of the urinary PCA3
test
The median PCA3 score was significantly higher in patients
with positive biopsies, i.e., 53 vs 20 (P < 0.0001) in patients
with negative biopsies (Table 1), and the risk of positive
biopsy increased with the PCA3 score (Fig. 1). The AUC of
the PCA3 score was significantly higher than that of the
serum PSA, respectively 0.76 (95% CI: 0.73—0.79) and 0.55
(0.51—0.58) (Fig. 2). The cutoff of 35, usually used in the literature, was indeed a good compromise between sensitivity
(68%) and specificity (71%) for an accuracy of 69% (Fig. 2).
The risk of positive biopsy (67%) in the 482 (47%) patients
with PCA3 score ≥ 35 was more than twice that of the 533
patients with PCA3 score < 35 (P < 0.001).
The diagnostic performance of the PCA3 test was similar
in patients with or without a history of negative biopsies,
Statistical analyses
The normal distribution of quantitative variables was verified to permit their comparison using the Student’s t-test.
Proportions for qualitative variables were compared using
the Chi2 test. Correlations between quantitative variables
were identified using linear regression.
ROC curve was established for PCA3 score to calculate
the area under the curve (AUC). For a given threshold, test
performance was evaluated in terms of sensitivity, specificity, positive and negative predictive values, and accuracy
(rate of correctly-classified patients) by comparing PCA3
Figure 1.
score.
Increasing risk of positive biopsies with the urinary PCA3
e4
V. Vlaeminck-Guillem et al.
Table 1 Baseline characteristics of the 1015 patients with informative urine samples, and correlations with presence
of cancer on biopsies. For quantitative variables, median, IQR (interquartile range), mean and standard deviation are
indicated.
Whole cohort
Number of patients
n = 1015
Median age
64 years (59—69)
(n = 1015)
(mean: 64 ± 7)
Prior negative biopsies
0
n = 825 (81%)
1
n = 144 (14%)
2
n = 31 (3%)
≥ 3
n = 15 (1%)
Previous administration of 5˛-reductase inhibitors
Yes
n = 20 (2%)
No
n = 995 (98%)
Digital rectal examination
Suspicious
n = 145 (14%)
Not suspicious
n = 870 (86%)
Median prostate volume
40 mL (29—54)
(n = 1003)
(mean: 44 ± 21)
Median number of biopsy
12 (12—12)
cores
(mean: 12 ± 1)
Total serum PSA
Median
6.6 ng/mL
(5—9.4)
(mean 6.2 ± 4.3)
< 2.5 ng/mL
2.5—3.99 ng/mL
4—9.99 ng/mL
≥ 10 ng/mL
Median urinary PCA3 score
n = 25 (2%)
n = 67 (7%)
n = 697 (69%)
n = 226 (22%)
32 (16—66)
(mean: 57 ± 67)
Patients with
negative biopsies
Patients with
positive biopsies
n = 535 (53%)
63 years (58—67)
(mean: 63 ± 6)
n = 480 (47%)
66 years (61—70)
(mean: 66 ± 7)
n = 406 (76%)
n = 95 (18%)
n = 22 (4%)
n = 12 (2%)
n = 419 (87%)
n = 49 (10%)
n = 9 (2%)
n = 3 (1%)
< 0.001
n = 11 (2%)
n = 524 (98%)
n = 9 (2%)
n = 471 (98%)
= 0.836
n = 41 (8%)
n = 494 (92%)
45 mL (32—60)
(mean: 49 ± 23)
12 (12—12)
(mean: 12 ± 1)
n = 104 (22%)
n = 376 (78%)
35 mL (26—48)
(mean: 38 ± 17)
12 (12—12)
(mean: 12 ± 1)
< 0.001
6.4 ng/mL
(4.9—9)
(mean: 7.6 ± 4.8)
6.8 ng/mL
(5.1—10)
(mean:
16.4 ± 57.2)
n = 8 (2%)
n = 376 (6%)
n = 315 (66%)
n = 126 (26%)
53 (28—100)
(mean: 81 ± 78)
= 0.0004
n = 17 (8%)
n = 36 (7%)
n = 382 (71%)
n = 100 (19%)
20 (11—41)
(mean: 35 ± 47)
i.e., AUCs of 0.74 (0.67—0.82) and 0.76 (0.73—0.79) respectively, P = 0.643. It differed however as a function of PSA
values. Whether the PSA values were between 2.5 and 4,
between 4 and 10, or greater than 10 ng/mL, the diagnostic
performance of the PCA3 test was homogenous, with AUCs
between 0.73 and 0.76. However, in the group of patients
with PSA < 2.5 ng/mL, performance was particularly high
with an AUC of 0.95 (0.85—1.00).
Integration of the PCA3 score in a nomogram
Age, DRE findings (suspicious vs non suspicious), history of
negative prostate biopsies (no history vs at least one negative biopsy series), prostate volume, history of physical
treatment for benign prostate hyperplasia (BPH), serum PSA
and PCA3 score were all predictive factors of biopsy results
in univariate analysis (Table 2). They were all independently
predictive in multivariate analysis also, except for physical
treatment for BPH. All the clinical and biological variables
predictive of biopsy results in multivariate analysis (except
the PCA3 score) were integrated into a base predictive
model, the AUC of which was 0.75 (0.72—0.78) (Table 2).
When the PCA3 score was added to the base model, the
AUC increased significantly to 0.80 (0.77—0.83) (P < 0.001)
(Table 2). To compare the net predictive benefit of the PCA3
Significativity (P)
< 0.0001
< 0.0001
= 1.000
= 0.018
< 0.0001
test, a decision curve analysis (DCA) was performed. The
model incorporating the PCA3 score produced a predictive
benefit greater than that of the base model (Fig. 3).
Prediction of prostate cancer aggressiveness
In the 480 patients with positive biopsies, the median PCA3
score did not differ significantly between the 225 (47%)
patients with Gleason 6 score and the 255 (53%) patients
with Gleason score ≥ 7 (Table 3). Equally, there were no differences according to the presence or absence of perineural
invasion or extracapsular extension (Table 3). Conversely,
PCA3 score was significantly higher when lesions were bilateral, when ≥ 33% of cores were positive or when ≥ 10% of
prostate tissue was tumorous (Table 3).
An issue in PCa diagnosis is to only identify significant
cancers. Using a cut-off of 20 for the PCA3 score to perform a
biopsy, 332 (33%) of the patients in this study would have not
undergone the biopsy. A third of the ultimately unnecessary
biopsies would have been avoided, but 67 of the 480 cancers
(14%) would have missed. Among those 67 cancers with PCA3
score < 20, 17 were Gleason 7 and 2 were Gleason 8. Among
the 48 unidentified Gleason 6 cases, only one had a rate of
positive cores ≥ 33%. Among the 67 missed cancer cases,
14 had abnormal DRE and/or PSA level ≥ 10 ng/mL.
e5
0.30
0.20
0.10
0.00
Net benefit
0.40
0.50
Urinary PCA3 to predict prostate cancer in a cohort of 1015 patients
0
20
40
60
80
Seuil
probabilité
biopsies
en %
Probabili
tydethresho
lddefor
posipositives
ve biopsies
(%)
anone
ucun
Base model
Figure 2. Diagnostic performances of the urinary PCA3 score and
comparison with serum PSA. AUC: area under curve ROC; 95% CI:
95% confidence interval; Se: sensitivity; Spe: specificity; PPV: positive predictive value; NPV: negative predictive value; Acc: accuracy
(proportion of correctly-classified patients).
tou
alls
Base model + ¨PCA3 score
Figure 3. Evaluation of the ability of the PCA3 score to predict
positive biopsies (net benefit) using decision curve analysis (DCA);
base model: age, digital rectal examination, previous negative biopsies, prostate volume, serum PSA. DCA examined the theoretical
relationship between the threshold probability of the outcome of
positive prostate biopsies and the relative value of false-positive
and false-negative results. Here, we estimated the magnitude of
benefit resulting from altering clinical management in patients with
different threshold probabilities of positive biopsies.
Table 2 Multivariate analyses evaluating the ability of clinicobiological variables to predict prostate biopsy outcome
and comparison with the addition of urinary PCA3 score (n = 1015 patients).
Univariate analysis
Multivariate analysis
Base model
Age
Previous biopsies
Prostate volume
Digital rectal
examination
Serum PSA
PCA3 score
Base model
+ PCA3 score
OR
(IC 95%)
P
AUC
OR
(IC 95%)
P
OR
(IC 95%)
P
1.06
(1.04—1.08)
0.46
(0.33—0.64)
0.97
(0.96—0.98)
3.33
(2.27—4.90)
1.04
(1.02—1.06)
1.02
(1.01—1.02)
< 0.001
61.6%
(58—65)
55.7%
(53—58)
65.5%
(62—69)
57.0%
(55—59)
54.9%
(51—58)
75.8%
(73—79)
1.07
(1.05—1.10)
0.45
(0.30—0.67)
0.97
(0.96—0.97)
2.15
(1.39—3.34)
1.07
(1.02—1.12)
—
< 0.001
1.05
(1.02—1.07)
0.48
(0.31—0.75)
0.97
(0.96—0.97)
2.03
(1.31—3.15)
1.07
(1.02—1.12)
1.01
(1.01—1.02)
80.1%
(77—82)
< 0.001
Prediction
accuracya
Increase in
performances
OR: odds ratio.
a Estimation using the AUC (%).
< 0.001
< 0.001
< 0.001
< 0.001
< 0.001
74.6%
(72—78)
+5.5%
P < 0.001
< 0.001
< 0.001
0.001
0.005
—
0.001
< 0.001
0.002
0.006
< 0.001
e6
Table 3
V. Vlaeminck-Guillem et al.
Pathological characteristics of the prostate biopsies and correlations with the urinary PCA3 score.
Pathological criteria
Gleason score
6
7
8
9
Proportion of invaded cores
Median (IQR)
< 33%
≥ 33%
Proportion of invaded tissue
Median (IQR)
< 10%
≥ 10%
Laterality
Unilateral
Bilateral
Perineural involvement
No
Yes
Extracapsular extension
No
Yes
PCA3 Score
Median (IQR) (P)
= 0.674
n = 225 (47%)
n = 208 (43%)
n = 35 (7%)
n = 12 (3%)
n = 225 (47%)
n = 255 (53%)
48 (26—96)
60 (32—108)
= 0.0001
25% (14—48)
n = 264 (55%)
n = 216 (45%)
45 (23—85)
68 (39—132)
= 0.002
6% (2—14)
n = 307 (64%)
n = 171 (36%)
49 (24—93)
65 (36—129)
n = 263 (55%)
n = 217 (45%)
42 (23—82)
72 (42—132)
n = 389 (81%)
n = 91 (19%)
51 (28—97)
61 (29—110)
n = 464 (97%)
n = 16 (3%)
52 (28—99)
80 (25—141)
< 0.0001
= 0.863
= 0.200
IQR: interquartile range.
Discussion
In the present study, we evaluated the ability of the PCA3
urine test to predict prostate biopsy results in more than
1000 patients. To our knowledge, this is the largest study
of its kind ever performed in France. Our results confirm
previously published data concerning the urinary PCA3 test,
particularly its function as a predictor of biopsy outcome,
independent of other clinical and biological variables routinely used in clinical practice to decide whether or not
to perform a biopsy. We demonstrated here, both in multivariate logistic regression and in DCA, that the addition of
the PCA3 score to the usual clinical and biological variables
provides a significant diagnostic gain. The main limit of our
study is its single-center nature; all of the patients used in
this study were treated in the same urology department.
We do emphasize however that the referrals for the 1015
patients were made by 20 different urologists with different practices. For example, the rate of positive biopsy for
the six physicians who referred at least 75 patients each
ranged from 36 to 65%. We think therefore that our cohort
was sufficiently diverse and representative of a large range
of sensibilities in urology.
The overall diagnostic performance of the PCA3 test that
we report here, with an AUC of 0.76, is perfectly coherent
with that reported in the literature [4—6]. In our experience, the cut-off of 35 has indeed proven to be a good
compromise between sensitivity (68%) and specificity (71%).
Two-thirds of our patients with PCA3 score ≥ 35 had positive biopsies compared to less than a third for those with
score < 35. However, our findings also clearly indicate that
a high PCA3 score is not necessarily synonymous with cancer (false positives) and that low or even very low scores
are not necessarily synonymous with the absence of cancer (false negatives). It underlines the importance to use
the PCA3 score in conjunction with the usual clinical and
biological findings, and even consider its use as part of a
specific nomogram [10,11]. Five to ten percent of PCa do
not express the PCA3 RNA product [2,12], which explains
a certain number of false negatives. In contrast, there is
currently no clearly identified mechanism to explain false
positives, but patients should be closely monitored [13—15].
We also confirmed that the risk of positive biopsies increases
as the PCA3 score increases (Fig. 1). The use of a threshold is thus an arbitrary approach that must find a balance
between two contradictory imperatives: avoiding unnecessary biopsies whenever possible (to reduce public healthcare
costs, morbidity, psychological impacts for the patient) and
not missing too many cancers, particularly those that are
life-threatening. The PCA3 threshold of 20 may be a better
compromise than 35; with its a-third of our biopsies (i.e.,
half of the ultimately unnecessary biopsies) could have been
avoided. However, the absence of correlation between the
PCA3 score and the Gleason score on biopsied tissue, at least
in our study (see infra), is an obstacle to improve the ratio
of avoided biopsies to missed significant cancers. Cost effectiveness studies are needed to determine if the generalized
use of the PCA3 test can effectively reduce healthcare costs
by lowering the number of unnecessary biopsies and their
associated complications.
Urinary PCA3 to predict prostate cancer in a cohort of 1015 patients
In our large cohort, the PCA3 test was robust, particularly
with more than 98% of informative samples. This high rate is
usual with this commercial kit, which includes a target RNA
selection step before amplification [9]. In our study, only the
use of 5␣-reductase inhibitors appeared to reduce the rate
of informative samples, which is logical since these agents
induce a reduction in the expression of PSA RNA. However, when informative samples are retrieved from patients
taking 5␣-reductase inhibitors, the diagnostic performance
of the PCA3 test does not appear to be affected, as was
demonstrated in the REDUCE trial on dutasteride for PCa
prevention [16].
Also illustrating the robustness of the PCA3 test was the
lack of influence of the usual clinical variables on its performance, which remained good regardless of age, prostate
volume, DRE results, or the presence or absence of a history
of negative biopsies. Considering these findings, it appears
to us that restricting the availability of the PCA3 test to
only those patients with a history of negative biopsies is not
justified; this position is additionally strengthened by the
results of recent studies on patients undergoing an initial
prostate biopsy [11,17]. In contrast to PSA levels, the PCA3
score does not increase with prostate volume. Furthermore,
in our study, we detected an inverse correlation wherein
the PCA3 score increases as prostate volume decreases.
This reflects essentially the higher frequency of cancer in
patients with low prostate volume (PCA3 being a cancer
marker) in opposition to the higher frequency of BPH in
patients with high prostate volume. The only factor that had
an influence on PCA3 test performance in our study was the
PSA level itself. More precisely, the diagnostic performance
of the PCA3 test was stable across all PSA levels greater
than 2.5 ng/mL (AUCs between 0.73 and 0.76), but under
that threshold, it improved (AUC of 0.95). These patients
with PSA < 2.5 ng/mL were mainly referred for biopsy based
solely on a confirmed family history of prostate cancer. Thus,
the PCA3 test appears to be particularly advantageous in this
specific population, where its accuracy has the potential to
limit repeated and unnecessary biopsies in frequently young
patients.
The diagnostic performance of the PCA3 test prompted
the clinicians to assess its interest for prognosis. Correlations between the PCA3 score and the Gleason score on
biopsies have been suggested [4], but the literature remains
contradictory on this question. In the present study, we
found no correlations between the PCA3 score and the
Gleason score; our negative result can be explained neither by a lack of statistical power nor by differences in
studied populations (similar proportions of patients with
Gleason scores < or ≥ 7). We note as well that in an earlier work, our team found no correlations between PCA3
scores and Gleason scores calculated on radical prostatectomy specimens (study on 154 patients from the same
source cohort as here) [18]. Additionally, early histological
studies did not demonstrate any links between PCA3 gene
expression and cancer differentiation [2,3,19]. In contrast,
studies on prostatectomy specimens have demonstrated
associations between the PCA3 score and tumor volume
[18,20], as have biopsy studies such as the present via the
proportions of positive cores and tumorous prostatic tissue.
e7
Conclusion
The present study involving more than 1000 patients confirmed performance data on the urinary PCA3 test for the
prediction of prostate biopsy results. The classic clinical
and biological variables such as age, DRE findings, history
of negative biopsies, prostate volume or serum PSA level
did not influence the performance of the test. Additional
studies are needed to evaluate the interest of pairing (simultaneously or sequentially) the PCA3 test with MRI of the
prostate, which is increasingly used for the diagnosis of PCa
and the assessment of its aggressiveness. The benefits of
avoiding unnecessary biopsies is a major issue. When evaluated as a part of decisional models for repeat biopsies on
a history of negative biopsies, the introduction of the PCA3
test in clinical practice could save between D 1.7 and 5 million according to the costs related to biopsy complications
[21]. The cost of cancers that may escape detection when
the PCA3 test is used to decide on a biopsy should be integrated into these calculations. Finally, the impact of the test
in the diagnostic process overall remains to be determined.
Toward this, an evaluation is underway as part of a support
program for costly innovative technologies funded by the
French National Cancer Institute (INCa).
Acknowledgments
We thank Doctors M. Vinet, E. Briant, N. Gobeaux, J.L.
Campos-Fernandez, E. Adam, S. Genevoix, C. De Vendin,
R. Lardon, G. Pic, F.X. Buttin, F. Lalloue, M. Goris and X.
Borgnat for their aid in enrolling patients, Ms. M. Cottancin
and Ms. B. Grangier for their technical assistance, and Ms.
M. Dupuis for her assistance with data collection.
Disclosure of interest
The authors declare that they have no competing interest.
References
[1] Hessels D, Klein Gunnewiek JM, van Oort I, Karthaus HF, van
Leenders GJ, van Balken B, et al. DD3(PCA3)-based molecular
urine analysis for the diagnosis of prostate cancer. Eur Urol
2003;44:8—15.
[2] Bussemakers MJ, van Bokhoven A, Verhaegh GW, Smit FP,
Karthaus HF, Schalken JA, et al. DD3: a new prostate-specific
gene, highly overexpressed in prostate cancer. Cancer Res
1999;59:5975—9.
[3] de Kok JB, Verhaegh GW, Roelofs RW, Hessels D, Kiemeney LA,
Aalders TW, et al. DD3(PCA3), a very sensitive and specific
marker to detect prostate tumors. Cancer Res 2002;62:2695—8.
[4] Auprich M, Bjartell A, Chun FK, de la Taille A, Freedland
SJ, Haese A, et al. Contemporary role of prostate cancer
antigen 3 in the management of prostate cancer. Eur Urol
2011;60:1045—54.
[5] Bradley LA, Palomaki G, Gutman S, Samson DJ, Aronson N. PCA3
testing for the diagnosis and management of prostate cancer. AHRQ comparative effectiveness reviews. Rockville (MD):
Agency for Healthcare Research and Quality (US); 2013.
[6] Vlaeminck-Guillem V, Ruffion A, Andre J. Value of urinary PCA3
test for prostate cancer diagnosis. Prog Urol 2008;18:259—65.
e8
[7] Vlaeminck-Guillem V, Bandel M, Cottancin M, RodriguezLafrasse C, Bohbot JM, Sednaoui P. Chronic prostatitis does not
influence urinary PCA3 score. Prostate 2012;72:549—54.
[8] Vlaeminck-Guillem V, Devonec M, Paparel P, Champetier D,
Campos-Fernandes JL, Perrin P, et al. Value of PCA3 urinary test
for prostate biopsy decision: the Lyon-Sud University Hospital
experience. Ann Biol Clin 2011;69:31—9.
[9] Groskopf J, Aubin SM, Deras IL, Blase A, Bodrug S, Clark C,
et al. APTIMA PCA3 molecular urine test: development of a
method to aid in the diagnosis of prostate cancer. Clin Chem
2006;52:1089—95.
[10] Chun FK, de la Taille A, van Poppel H, Marberger M, Stenzl A,
Mulders PF, et al. Prostate Cancer Gene 3 (PCA3): development
and internal validation of a novel biopsy nomogram. Eur Urol
2009;56:659—68.
[11] Ruffion A, Devonec M, Champetier D, Decaussin-Petrucci M,
Rodriguez-Lafrasse C, Paparel P, et al. PCA3 and PCA3-based
nomograms improve diagnostic accuracy in patients undergoing
first prostate biopsy. Int J Mol Sci 2013;14:17767—80.
[12] Floriano-Sanchez E, Cardenas-Rodriguez N, Castro-Marin M,
Alvarez-Grave P, Lara-Padilla E. DD3(PCA3) gene expression in cancer and prostatic hyperplasia. Clin Invest Med
2009;32:E258.
[13] Remzi M, Haese A, Van Poppel H, De La Taille A, Stenzl A,
Hennenlotter J, et al. Follow-up of men with an elevated
PCA3 score and a negative biopsy: does an elevated PCA3
score indeed predict the presence of prostate cancer? BJU Int
2010;106:1138—42.
[14] Roobol MJ, Schroder FH, van Leenders GL, Hessels D, van
den Bergh RC, Wolters T, et al. Performance of Prostate
Cancer Antigen 3 (PCA3) and prostate-specific antigen in
V. Vlaeminck-Guillem et al.
[15]
[16]
[17]
[18]
[19]
[20]
[21]
prescreened men: reproducibility and detection characteristics for prostate cancer patients with high PCA3 Scores (≥ 100).
Eur Urol 2010;58:893—9.
Schroder FH, Venderbos LD, van den Bergh RC, Hessels D, van
Leenders GJ, van Leeuwen PJ, et al. Prostate cancer antigen
3: diagnostic outcomes in men presenting with urinary prostate
cancer antigen 3 scores ≥ 100. Urology 2014;83:613—6.
Aubin SM, Reid J, Sarno MJ, Blase A, Aussie J, Rittenhouse
H, et al. Prostate Cancer Gene 3 Score predicts prostate
biopsy outcome in men receiving dutasteride for prevention
of prostate cancer: results from the REDUCE Trial. Urology
2011;78:380—5.
Chevli KK, Duff M, Walter P, Yu C, Capuder B, Elshafei A,
et al. Urinary PCA3 as a predictor for prostate cancer in a
cohort of 3073 men undergoing initial prostate biopsy. J Urol
2014;191:1743—8.
Tallon L, Luangphakdy D, Ruffion A, Colombel M, Devonec M,
Champetier D, et al. Comparative evaluation of urinary PCA3
and TMPRSS2: ERG scores and serum PHI in predicting prostate
cancer aggressiveness. Int J Mol Sci 2014;15:13299—316.
Balcerczak E, Mirowski M, Sasor A, Wierzbicki R. Expression
of p65, DD3 and c-erbB2 genes in prostate cancer. Neoplasma
2003;50:97—101.
Nakanishi H, Groskopf J, Fritsche HA, Bhadkamkar V, Blase A,
Kumar SV, et al. PCA3 molecular urine assay correlates with
prostate cancer tumor volume: implication in selecting candidates for active surveillance. J Urol 2008;179:1804—9.
Malavaud B, Cussenot O, Mottet N, Rozet F, Ruffion A, Smets L,
et al. Impact of adoption of a decision algorithm including PCA3
for repeat biopsy on the costs for prostate cancer diagnosis in
France. J Med Econ 2013;16:358—63.