Screening and Biopsy Prostate Cancer Antigen 3: Diagnostic Outcomes in Men Presenting With Urinary Prostate Cancer Antigen 3 Scores ‡100 €der, Lionne D. F. Venderbos, Roderick C. N. van den Bergh, Daphne Hessels, Fritz H. Schro Geert J. L. H. van Leenders, Pim J. van Leeuwen, Tineke Wolters, Jelle O. Barentsz, and Monique J. Roobol OBJECTIVE METHODS
RESULTS
CONCLUSION
To describe the results of 3 rounds of diagnostic testing and linkage to the Dutch Cancer Registry for men with an initial prostate cancer antigen 3 (PCA3) score 100. Within an earlier reported comparative study of PCA3 vs prostate-specific antigen in a prescreened population, 90 men with a PCA3 score 100 were identified and underwent biopsy, 28 prostate cancers (PCs) were found, 62 men remained at risk of a diagnosis of PC. All men were offered repeat testing; 6 PCs were found in 20 men at rebiopsy. Men with at least 1 negative biopsy (n ¼ 56) were invited to undergo magnetic resonance imaging (MRI) studies and MRIguided biopsies if indicated. Linkage to the Dutch Cancer Registry after 2.8 years of follow-up was performed for men with negative biopsies. Of the 56 men at risk, 28 agreed to participate in further testing. They were offered MRI studies; only 7 men agreed, and in 2, suspicious lesions were found and biopsies carried out. Only 1 PC was diagnosed and classified as T1c, Gleason 3 þ 3 ¼ 6. The overall findings of 3 rounds of testing and of linkage to the cancer registry show that eventually 35 PCs were detected in 90 men with PCA3 scores 100 (positive predictive value 38.9%). Finding no PC despite extended diagnostic efforts in many men with initial PCA3 scores 100 is unexpected and might be clinically relevant. UROLOGY 83: 613e616, 2014. 2014 Elsevier Inc.
A
lthough prostate cancer antigen 3 (PCA3) has been shown in multiple publications to be a promising marker, its value in the primary diagnosis, the identification of aggressive disease and the possibility to avoid unnecessary prostate biopsies, is still under discussion. The available information has been summarized in recent reviews, including the studies by Auprich et al and Hessels and Schalken.1,2 The value of PCA3 with a cutoff score of 35 in men who had been biopsied in the past, however, is considered to be firmly established by several studies, including the studies by Hessels et al and Wu et al.3,4 The present report is a final step of a study of 965 men aged 63-75 years, participants in the European Randomized Study of Screening for
Financial Disclosure: The authors declare that they have no relevant financial interests. Funding Support: This study was funded in part by an unconditional grant of Gen Probe Corp, San Diego, CA. From the Department of Urology, Erasmus Medical Center, Rotterdam, the Netherlands; the Department of Urology, Radboud University Nijmegen Medical Centre, Nijmegen, the Netherlands; the Department of Pathology, Erasmus Medical Center, Rotterdam, the Netherlands; and the Department of Radiology, Radboud University Nijmegen Medical Centre, Nijmegen, the Netherlands Reprint requests: Fritz H. Schr€oder, M.D., Ph.D., Department of Urology, Erasmus Medical Center, University Medical Center, P.O. Box 2040, 3000 CA Rotterdam, the Netherlands. E-mail: [email protected] Submitted: July 11, 2013, accepted (with revisions): December 3, 2013
ª 2014 Elsevier Inc. All Rights Reserved
Prostate Cancer (ERSPC) section Rotterdam screening arm, who were evaluated by prostate-specific antigen (PSA) and PCA3 during the period from September 2007 to February 2008.5,6 All participants have been screened once or twice for respectively 8 years earlier in previous screening rounds using PSA as a biopsy indication. The purpose of the study was to assess the value of PCA3 as a first-line diagnostic test in comparison with PSA with a cutoff of 3 ng/mL as a biopsy indication. To establish performance characteristics of PCA3 in a prescreened population, a low PSA score of 10 was used as a biopsy indication in addition to a PSA cutoff value of 3.0 ng/mL. After 2 rounds of screening, 56 men still remained at risk of a possible later diagnosis of prostate cancer (PC). These men are the subject of the present investigation and of this report.
MATERIALS AND METHODS The present report aims at determining the diagnostic outcomes in 56 men who were still at risk of being diagnosed with PC because of very high initial PCA3 scores, PSA values 3.0 ng/mL, and negative biopsies after 2 rounds of testing.5,6 Of these men, 28 have refused to participate in further testing. To the remaining 28, multiparametric magnetic resonance imaging (MRI) and if indicated MRI-guided biopsy were offered, but only 0090-4295/14/$36.00 http://dx.doi.org/10.1016/j.urology.2013.12.005
613
7 agreed to undergo this procedure. MRI was performed at a field strength of 3 T without endorectal coil according to the European Society of Urological Radiolog (ESUR) 2012 guidelines and MRI-guided biopsy as described by Hoeks et al.7,8 All remaining men at risk were linked to the Dutch Cancer Registry, for which very high accuracy was documented in the past,9 to show or exclude a diagnosis of PC after a 2.8-year follow-up. This prospective study is based on the screen arm of the European Randomized Study of Screening for Prostate Cancer (ERSPC) trial, section Rotterdam described in detail in ref.10 and is the final third step in a side study reported earlier.5,6 Permission to conduct the present study (ISBN 978-90-5549-653-2) was given by the Ministry of Health of the Netherlands on July 24, 2007. Written informed consent was obtained from all men who confirmed their participation.
965 pre-screened men age 63-74 underwent a PCA3 urine test [1]. Re-testing was offered to 62 men with normal biopsies
First round PSA and PCA3 testing
N
%
Study population (all biopsied) PCA3 ≥10 PC PCA3 ≥ 100 PC (PPV 31.1%) Still at risk
721 689 118 90 28 62
100 95.6 17.1 12.5
Second round PSA and PCA3 testing*
N
%
Offered a second round Accepted re-testing PCA3 ≥ 100 Biopsy indicated Biopsied PC (PPV 30.0)
62 34 21 33 20 6
100 61.8 97.1 58.8
Third round PCA3 testing**
N
Offered a third round Refused further testing mpMRI offered mpMRI study done PC found at MRI biopsy Linked to cancer registry PC found by linkage (2.8 yrs fu)
56 28 28 7 1 55 0
8.6
RESULTS Our study population consists of 56 men with PCA3 scores of >100 at previous screens, of whom 28 decided not to participate in further diagnostic testing. To understand how our study population is derived, it is necessary to consider the first and second rounds of PSA and PCA3 testing, which are described in Figure 1. Of 965 men, 721 men had at least one of the biopsy indications and were biopsied, 122 PCs were detected (16.9%), 4 with elevated PSA values and PCA3 scores 35, and 100 in all the 721 patients who were biopsied as part of the initial study population.5 The distribution of PCA3 scores in men without PC is also shown together with the numbers of biopsies and PPV values of PC cases per range of PCA3 scores. Both the PPVs of diagnosed PC cases and of high-grade PINs seem to increase with increasing PCA3 scores. However, even if both PPVs are added up, the resulting figure of 19.2% leaves 80.8% of the high PCA3 scores unexplained.
COMMENT In our study, of 721 men who were all biopsied, the performance characteristics of PCA3 were evaluated with special consideration of very high PA3 scores. The PPV of PCA3 for 698 patients presenting with scores 10 amounted to only 17.1%. This is based on 118 cancers detected at the first round of evaluation. For 90 men with an initial PCA3 score 100, all biopsied during the first round, the PPV was 31.1% (28 PC/90 men). Six additional cancers were detected in the second round of biopsies in 20 men who agreed to participate. If we UROLOGY 83 (3), 2014
extrapolate this figure to all 62 remaining men and assume an identical detection rate, a total of 19 PCs can be assumed to be found. This would have resulted in a PPV of 52.2% (28 þ 19 divided by 90 men). One more cancer was found in 7 men who agreed to undergo MRIguided biopsies. Adding the 7 cancers detected in rounds 2 and 3 in those who consented to be biopsied, the PPV for PCA3 100 rose to 38.9% (35PC/90 men). When we compensated for missing data by linkage to the Dutch Cancer Registry after an average follow-up of 2.8 years for all the 55 men who were still at risk, no PCs were clinically diagnosed during this period. One might expect the detection rate of 38.9% to be low considering the fact that our population of men has been prescreened. This, however, does not seem to be the case at least if the PPV based on extrapolation is considered. Current literature reports detection rates with one or several biopsies for men with PCA3 scores 100 ranging from 45.6% to 88.9%.6 The fact that 61.1% of PCA3 scores 100 remain unexplained remains surprising on the background of the fact that PCA3 might be considered the only PC-specific marker on the basis of the impressive research reported by Bussemakers et al.12 This publication describes overexpression of the PCA3 gene, then called DD3, in more than 95% of PC tissue specimens and metastases. No overexpression was found in normal prostate tissue or benign prostatic hyperplasia. In the absence of detectable PC, other mechanisms that might explain the low detection rate and low PPV must be considered. In the first place, PC might be missed by biopsy. It is for this reason that we decided to maximize diagnostic efforts. With 2 further rounds of clinical investigations and linkage to the cancer registry with a follow-up of 2.8 years, no more than 35 cancers were found in 90 men with initial PCA3 scores 100. Furthermore, the finding of lower PCA3 scores at the second round of testing has been disturbing. This is counterintuitive, considering the 19-month interval during which one would expect that a cancer-specific marker would rise rather than decrease. Still, even in second round PCA3 testing,6 only one of the 90 patients showed a PCA3 score
RESULTS
CONCLUSION
To describe the results of 3 rounds of diagnostic testing and linkage to the Dutch Cancer Registry for men with an initial prostate cancer antigen 3 (PCA3) score 100. Within an earlier reported comparative study of PCA3 vs prostate-specific antigen in a prescreened population, 90 men with a PCA3 score 100 were identified and underwent biopsy, 28 prostate cancers (PCs) were found, 62 men remained at risk of a diagnosis of PC. All men were offered repeat testing; 6 PCs were found in 20 men at rebiopsy. Men with at least 1 negative biopsy (n ¼ 56) were invited to undergo magnetic resonance imaging (MRI) studies and MRIguided biopsies if indicated. Linkage to the Dutch Cancer Registry after 2.8 years of follow-up was performed for men with negative biopsies. Of the 56 men at risk, 28 agreed to participate in further testing. They were offered MRI studies; only 7 men agreed, and in 2, suspicious lesions were found and biopsies carried out. Only 1 PC was diagnosed and classified as T1c, Gleason 3 þ 3 ¼ 6. The overall findings of 3 rounds of testing and of linkage to the cancer registry show that eventually 35 PCs were detected in 90 men with PCA3 scores 100 (positive predictive value 38.9%). Finding no PC despite extended diagnostic efforts in many men with initial PCA3 scores 100 is unexpected and might be clinically relevant. UROLOGY 83: 613e616, 2014. 2014 Elsevier Inc.
A
lthough prostate cancer antigen 3 (PCA3) has been shown in multiple publications to be a promising marker, its value in the primary diagnosis, the identification of aggressive disease and the possibility to avoid unnecessary prostate biopsies, is still under discussion. The available information has been summarized in recent reviews, including the studies by Auprich et al and Hessels and Schalken.1,2 The value of PCA3 with a cutoff score of 35 in men who had been biopsied in the past, however, is considered to be firmly established by several studies, including the studies by Hessels et al and Wu et al.3,4 The present report is a final step of a study of 965 men aged 63-75 years, participants in the European Randomized Study of Screening for
Financial Disclosure: The authors declare that they have no relevant financial interests. Funding Support: This study was funded in part by an unconditional grant of Gen Probe Corp, San Diego, CA. From the Department of Urology, Erasmus Medical Center, Rotterdam, the Netherlands; the Department of Urology, Radboud University Nijmegen Medical Centre, Nijmegen, the Netherlands; the Department of Pathology, Erasmus Medical Center, Rotterdam, the Netherlands; and the Department of Radiology, Radboud University Nijmegen Medical Centre, Nijmegen, the Netherlands Reprint requests: Fritz H. Schr€oder, M.D., Ph.D., Department of Urology, Erasmus Medical Center, University Medical Center, P.O. Box 2040, 3000 CA Rotterdam, the Netherlands. E-mail: [email protected] Submitted: July 11, 2013, accepted (with revisions): December 3, 2013
ª 2014 Elsevier Inc. All Rights Reserved
Prostate Cancer (ERSPC) section Rotterdam screening arm, who were evaluated by prostate-specific antigen (PSA) and PCA3 during the period from September 2007 to February 2008.5,6 All participants have been screened once or twice for respectively 8 years earlier in previous screening rounds using PSA as a biopsy indication. The purpose of the study was to assess the value of PCA3 as a first-line diagnostic test in comparison with PSA with a cutoff of 3 ng/mL as a biopsy indication. To establish performance characteristics of PCA3 in a prescreened population, a low PSA score of 10 was used as a biopsy indication in addition to a PSA cutoff value of 3.0 ng/mL. After 2 rounds of screening, 56 men still remained at risk of a possible later diagnosis of prostate cancer (PC). These men are the subject of the present investigation and of this report.
MATERIALS AND METHODS The present report aims at determining the diagnostic outcomes in 56 men who were still at risk of being diagnosed with PC because of very high initial PCA3 scores, PSA values 3.0 ng/mL, and negative biopsies after 2 rounds of testing.5,6 Of these men, 28 have refused to participate in further testing. To the remaining 28, multiparametric magnetic resonance imaging (MRI) and if indicated MRI-guided biopsy were offered, but only 0090-4295/14/$36.00 http://dx.doi.org/10.1016/j.urology.2013.12.005
613
7 agreed to undergo this procedure. MRI was performed at a field strength of 3 T without endorectal coil according to the European Society of Urological Radiolog (ESUR) 2012 guidelines and MRI-guided biopsy as described by Hoeks et al.7,8 All remaining men at risk were linked to the Dutch Cancer Registry, for which very high accuracy was documented in the past,9 to show or exclude a diagnosis of PC after a 2.8-year follow-up. This prospective study is based on the screen arm of the European Randomized Study of Screening for Prostate Cancer (ERSPC) trial, section Rotterdam described in detail in ref.10 and is the final third step in a side study reported earlier.5,6 Permission to conduct the present study (ISBN 978-90-5549-653-2) was given by the Ministry of Health of the Netherlands on July 24, 2007. Written informed consent was obtained from all men who confirmed their participation.
965 pre-screened men age 63-74 underwent a PCA3 urine test [1]. Re-testing was offered to 62 men with normal biopsies
First round PSA and PCA3 testing
N
%
Study population (all biopsied) PCA3 ≥10 PC PCA3 ≥ 100 PC (PPV 31.1%) Still at risk
721 689 118 90 28 62
100 95.6 17.1 12.5
Second round PSA and PCA3 testing*
N
%
Offered a second round Accepted re-testing PCA3 ≥ 100 Biopsy indicated Biopsied PC (PPV 30.0)
62 34 21 33 20 6
100 61.8 97.1 58.8
Third round PCA3 testing**
N
Offered a third round Refused further testing mpMRI offered mpMRI study done PC found at MRI biopsy Linked to cancer registry PC found by linkage (2.8 yrs fu)
56 28 28 7 1 55 0
8.6
RESULTS Our study population consists of 56 men with PCA3 scores of >100 at previous screens, of whom 28 decided not to participate in further diagnostic testing. To understand how our study population is derived, it is necessary to consider the first and second rounds of PSA and PCA3 testing, which are described in Figure 1. Of 965 men, 721 men had at least one of the biopsy indications and were biopsied, 122 PCs were detected (16.9%), 4 with elevated PSA values and PCA3 scores 35, and 100 in all the 721 patients who were biopsied as part of the initial study population.5 The distribution of PCA3 scores in men without PC is also shown together with the numbers of biopsies and PPV values of PC cases per range of PCA3 scores. Both the PPVs of diagnosed PC cases and of high-grade PINs seem to increase with increasing PCA3 scores. However, even if both PPVs are added up, the resulting figure of 19.2% leaves 80.8% of the high PCA3 scores unexplained.
COMMENT In our study, of 721 men who were all biopsied, the performance characteristics of PCA3 were evaluated with special consideration of very high PA3 scores. The PPV of PCA3 for 698 patients presenting with scores 10 amounted to only 17.1%. This is based on 118 cancers detected at the first round of evaluation. For 90 men with an initial PCA3 score 100, all biopsied during the first round, the PPV was 31.1% (28 PC/90 men). Six additional cancers were detected in the second round of biopsies in 20 men who agreed to participate. If we UROLOGY 83 (3), 2014
extrapolate this figure to all 62 remaining men and assume an identical detection rate, a total of 19 PCs can be assumed to be found. This would have resulted in a PPV of 52.2% (28 þ 19 divided by 90 men). One more cancer was found in 7 men who agreed to undergo MRIguided biopsies. Adding the 7 cancers detected in rounds 2 and 3 in those who consented to be biopsied, the PPV for PCA3 100 rose to 38.9% (35PC/90 men). When we compensated for missing data by linkage to the Dutch Cancer Registry after an average follow-up of 2.8 years for all the 55 men who were still at risk, no PCs were clinically diagnosed during this period. One might expect the detection rate of 38.9% to be low considering the fact that our population of men has been prescreened. This, however, does not seem to be the case at least if the PPV based on extrapolation is considered. Current literature reports detection rates with one or several biopsies for men with PCA3 scores 100 ranging from 45.6% to 88.9%.6 The fact that 61.1% of PCA3 scores 100 remain unexplained remains surprising on the background of the fact that PCA3 might be considered the only PC-specific marker on the basis of the impressive research reported by Bussemakers et al.12 This publication describes overexpression of the PCA3 gene, then called DD3, in more than 95% of PC tissue specimens and metastases. No overexpression was found in normal prostate tissue or benign prostatic hyperplasia. In the absence of detectable PC, other mechanisms that might explain the low detection rate and low PPV must be considered. In the first place, PC might be missed by biopsy. It is for this reason that we decided to maximize diagnostic efforts. With 2 further rounds of clinical investigations and linkage to the cancer registry with a follow-up of 2.8 years, no more than 35 cancers were found in 90 men with initial PCA3 scores 100. Furthermore, the finding of lower PCA3 scores at the second round of testing has been disturbing. This is counterintuitive, considering the 19-month interval during which one would expect that a cancer-specific marker would rise rather than decrease. Still, even in second round PCA3 testing,6 only one of the 90 patients showed a PCA3 score
