The Prostate 75:103^111 (2015)
Accuracy of the Prostate Health Index Versus the Urinary Prostate Cancer Antigen 3 Score to Predict Overall and Signif|cant Prostate Cancer at Initial Biopsy Thomas Seisen,1,2 Morgan Rouprêt,1,2* Didier Brault,2,3 Priscilla Leon,1,2 Geraldine Cancel-Tassin,2,4 Eva Comperat,5 Raphae€le Renard-Penna,6 Pierre Mozer,1 Jerome Guechot,3 and Olivier Cussenot1,2,4,7 1
2
Academic Department of Urology, Pitie-Salpetriere Hospital, AP-HP, Paris, France UPMCUniv Paris 06,GRC5,ONCOTYPE-URO, Institut Universitaire de Cancerologie, Paris, France 3 Academic Department of Biochemistry,GH Paris Est, AP-HP, Paris, France 4 CeRePP,Tenon Hospital, Paris, France 5 Academic Department of Pathology, Pitie-Salpetriere Hospital, AP-HP, Paris, France 6 Academic Department of Radiology, Pitie-Salpetriere Hospital, AP-HP, Paris, France 7 Academic Department of Urology,GH Paris Est, AP-HP, Paris, France BACKGROUND. It remains unclear whether the Prostate Health Index (PHI) or the urinary Prostate-Cancer Antigen 3 (PCA-3) score is more accurate at screening for prostate cancer (PCa). The aim of this study was to prospectively compare the accuracy of PHI and PCA-3 scores to predict overall and significant PCa in men undergoing an initial prostate biopsy. METHODS. Double-blind assessments of PHI and PCA-3 were conducted by referent physicians in 138 patients who subsequently underwent trans-rectal ultrasound-guided prostate biopsy according to a 12-core scheme. Predictive accuracies of PHI and PCA-3 were assessed using AUC and compared according to the DeLong method. Diagnostic performances with usual cut-off values for positivity (i.e., PHI >40 and PCA-3 >35) were calculated, and odds ratios associated with predicting PCa overall and significant PCa as defined by pathological updated Epstein criteria (i.e., Gleason score 7, more than three positive cores, or >50% cancer involvement in any core) were estimated using logistic regression. RESULTS. Prevalences of overall and significant PCa were 44.9% and 28.3%, respectively. PCA-3 (AUC ¼ 0.71) was the most accurate predictor of PCa overall, and significantly outperformed PHI (AUC ¼ 0.65; P ¼ 0.03). However, PHI (AUC ¼ 0.80) remained the most accurate predictor when screening exclusively for significant PCa and significantly outperformed PCA-3 (AUC ¼ 0.55; P ¼ 0.03). Furthermore, PCA-3 >35 had the best accuracy, and positive or negative predictive values when screening for PCa overall whereas these diagnostic performances were greater for PHI >40 when exclusively screening for significant PCa. PHI > 40 combined with PCA-3 > 35 was more specific in both cases. In multivariate analyses, PCA-3 >35 (OR ¼ 5.68; 95%CI ¼ [2.21–14.59]; P < 0.001) was significantly correlated with the presence of PCa overall, but PHI >40 (OR ¼ 9.60; 95%CI ¼ [1.72–91.32]; P ¼ 0.001) was the only independent predictor for detecting significant PCa. CONCLUSIONS. Although PCA-3 score is the best predictor for PCa overall at initial biopsy, our findings strongly indicate that PHI should be used for population-based screening to avoid overdiagnosis of indolent tumors that are unlikely to cause death. Prostate 75:103–111, 2015. # 2014 Wiley Periodicals, Inc.
Grant sponsor: Access Hybritech p2PSA ([-2]proPSA) reagents were provided by Beckman Coulter Inc. Correspondence to: Morgan Rouprêt, MD, PhD, Department of Urology, Pitie-Salpêtriere Hospital, 47-83 boulevard de l’hopital, 75013 Paris, France. E-mail: [email protected]
ß 2014 Wiley Periodicals, Inc.
Received 29 June 2014; Accepted 21 August 2014 DOI 10.1002/pros.22898 Published online 18 October 2014 in Wiley Online Library (wileyonlinelibrary.com).
104
Seisen et al. KEY WORDS: prostatic neoplasm; biopsy; early detection of cancer; biological markers; prostate-specific antigen; antigens
INTRODUCTION Recent debate on screening for prostate cancer (PCa) has focused on over-diagnosis and overtreatment of insignificant tumors known for their low propensity to cause death [1]. To evaluate the efficacy of population-based screening, two large randomized trials have been conducted but their conflicting results emphasized the rationale of the current controversy [2,3]. The European Randomized Study of Screening for Prostate Cancer (ERSPC) reported that screening for PCa was associated with a significant reduction in specific mortality after a median follow-up of 11 years. In contrast, the Prostate, Lung, Colorectal and Ovary (PLCO) project concluded that PCa-related mortality was very low when PCa was detected by screening and did not significantly differ from that of the standard-care population. However, a relative risk reduction of 42% for developing metastatic disease has increased interest in improving early detection using opportunistic prostate-specific antigen (PSA)-based screening [4]. Unfortunately, PSA testing has become a victim of its own success and no longer fulfills all its commitments. Over the last few years, this cornerstone of PCa screening has been often cited as producing too many negative prostate biopsies and leading to diagnosis of tumors that are unlikely to cause future health problems, even if left untreated. Consequently, the limited ability of PSA to detect significant PCa has prompted the development of several biomarkers that share the common objective of avoiding over-diagnosis and over-treatment. In this context, urinary molecular biomarkers have recently been proposed and, among these, the Prostate Cancer Antigen 3 (PCA-3) score is now used to guide indications for a repeat biopsy in cases where there is persistent suspicion of PCa after a previous negative biopsy [5]. Moreover, the new serum biomarkers, such as [2]proPSA and its derivative, the Prostate Health Index (PHI), have significantly increased the ability to detect PCa, especially at an initial biopsy [6]. The secretion of [2]proPSA appears to be more specific to tumor cells, and the combination of this PSA isoform with pertinent serum biomarkers, including total PSA (tPSA) and free PSA (fPSA), has resulted in the PHI, which has greatly improved discrimination between men with and without PCa and between those with significant and insignificant disease [7–9]. Despite several direct comparisons of the diagnostic effectiveness of these biomarkers, it remains unclear whether the PHI or the PCA-3 score should be used
The Prostate
for PCa screening. Although they are both more accurate than classic PSA testing, no significant difference has been reported between them in the setting of initial biopsies [10–14]. Thus, the aim of our study was to prospectively compare the diagnostic accuracy of PHI and PCA-3 scores to screen for overall and significant PCa in a cohort of men undergoing an initial prostate biopsy. MATERIALS AND METHODS Population Between June 2013 and December 2013, 177 consecutive patients undergoing a first prostate biopsy at our academic institution for suspected PCa, based on a tPSA ranging from 4 to 20 ng/ml and/or an abnormal digital rectal examination, were screened for inclusion in this prospective, observational, cohort study. Men receiving 5-a-reductase inhibitors (dutasteride or finasteride), largely known to affect tPSA level, and those with a history of previous invasive treatment for benign prostatic hyperplasia or acute bacterial prostatitis within the previous 3 months were excluded from the analyses. In addition, subjects with chronic renal failure, marked blood-protein alterations, hemophilia, or those previously polytransfused were excluded from the study, as these conditions may alter the concentrations of [2]proPSA. Overall, 138 men, aged >50 years, met the inclusion criteria and were enrolled in the current study after providing their written informed consent. This study received approval from the local hospital ethics committee and was performed according to the Standards for Reporting of Diagnostic Accuracy (STARD). Collection of Biomarkers Blood samples were drawn at the time of the first biopsy, just before any prostatic manipulations, to avoid any transient increase in pre-biopsy tPSA, fPSA, or [2]proPSA levels. After clotting the blood, serum aliquots were extracted by centrifugation within 3 hr and frozen at 80°C, according to the method of Semjonow et al. [15]. Specimens were processed using a UniCel1 DxI800 Immunoassay System analyzer, and serum biomarkers were assessed in a blinded fashion using a calibrated Hybritech1 Access2 Immunoassay System analyzer. To determine PHI, the following combination of the three previously assessed serum biomarker values was used: (([2]proPSA/fPSA) HtPSA).
PHI vs. PCA-3 to Detect Overall and Significant Prostate Cancer Urine samples were collected after a digital rectal examination using three strokes per lobe and immediately before a first prostate biopsy, as described by Groskopf et al. [16]. Specimens were processed and tested in a blinded fashion to quantify PCA-3 mRNA and PSA mRNA urinary concentrations using the Progensa1 PCA-3 assay. The PCA-3 score was calculated using the following formula: (PCA-3 mRNA/PSA mRNA) 1000. Prostate Biopsies After assessing for biomarkers, trans-rectal ultrasound-guided prostate biopsies were performed by a single senior urologist according to a standardized classic 12-core scheme. Trans-rectal ultrasound was also used to assess prostate volume and to target any significant abnormalities in the prostate gland during the procedure. All cores were mapped for location and submitted separately in specific single-core specimen containers filled with 10% buffered formalin for pathological evaluation. All specimens were processed and evaluated by our dedicated senior uro-pathologist who was blinded to the results from the urine and serum biomarkers. PCa was identified and graded according to definitions of the 2005 consensus conference of the International Society of Urological Pathology [17]. Significant PCa was defined using pathological updated Epstein criteria as follows: a Gleason score 7, more than three positive cores, or >50% cancer involvement in any core [18]. High-grade prostate intra-epithelial neoplasia or atypical small acinar proliferations were not considered as PCa. Study Endpoints The primary endpoint of the study was to directly compare the accuracy of PHI and PCA-3 scores to determine the most accurate biomarker to predict overall and significant PCa at an initial biopsy. We also compared the accuracy of these biomarkers with the PSAD to confirm that PHI and PCA-3 outperformed classic PSA testing when screening for overall and significant PCa. Statistical Analyses All statistical analyses were performed using R software (version 2.12.1; R Foundation for Statistical Computing, Vienna, Austria). Results were expressed as medians for continuous variables, and categorical variables were reported as percentages. The Shapiro– Wilk test assessed the normal distribution of variables. Student’s t-test and the Mann–Whitney U-test were used to compare normally and non-normally distribut-
105
ed continuous variables, respectively. The Chi-squared test was used to compare non-normally distributed categorical variables. All tests were two-sided with a P-value of 0.15, PHI >40, PCA-3 >35). Multivariate logistic regression analyses were performed to assess whether PCA-3 >35 and PHI >40 were independent predictors for the presence of overall and significant PCa at an initial biopsy.
RESULTS Participants The STARD flowchart is shown in Figure 1. The demographic characteristics of the study population are listed in Table I. Overall, PCa was diagnosed in 62 of the 138 men (44.9%) undergoing an initial biopsy, with a median PSAD of 0.17 (range: 0.02–0.55), median PHI of 42.4 (range: 5.9–171.7) and median PCA-3 of 49.2 (range: 1.0–304.0). When comparing patients with and without PCa, no significant difference was found in terms of clinical and biochemical characteristics except for PSAD, PHI, and PCA-3. Indeed, the proportions of patients presenting with PSAD >0.15 (67.7% vs. 57.9%; P ¼ 0.02), PHI >40 (43.5% vs. 32.9%; P ¼ 0.01), PCA-3 >35 (66.1% vs. 31.6%; P ¼ 0.01) and PHI > 40 combined with PCA-3 > 35 (27.4% vs. 14.5%; P ¼ 0.04) were significantly higher in the PCa group. Similarly, median PSAD (0.19 vs. 0.08; P ¼ 0.04), PHI (50.6 vs. 31.1; P ¼ 0.02), and PCA-3 (67.3 vs. 30.8; P < 0.01) significantly differed between men with and without PCa. Significant PCa was diagnosed in 39 of the 138 men (28.3%) undergoing an initial biopsy. When comparing patients with and without significant PCa, no significant difference was found in terms of clinical and biochemical characteristics except for PSAD, PHI, and PCA-3. Indeed, greater proportions of patients presented with PSAD >0.15 (79.5% vs. 55.6%; P < 0.01), PHI >40 (66.7% vs. 26.3%; P ¼ 0.01), PCA-3 The Prostate
106
Seisen et al.
Fig. 1. Flowchart showing patients inclusion process according to the Standards for Reporting of Diagnostic Accuracy (STARD). BPH, Benign prostatic hyperplasia; PCa, prostate cancer; S-PCa, significant PCa.
>35 (61.5% vs. 41.4%; P ¼ 0.03) and PHI > 40 combined with PCA-3 > 35 (28.2% vs. 17.2%; P ¼ 0.02) in the significant PCa group. Furthermore, patients with significant PCa had significantly higher median PSAD (0.26 vs. 0.13; P ¼ 0.02), PHI (65.7 vs. 36.5; P < 0.01), and PCA-3 (53.1 vs. 45.5; P ¼ 0.01). ROC-Curves Analysis In univariate accuracy analysis, PCA-3 (AUC ¼ 0.71) was the most accurate predictor of PCa overall at the initial biopsy compared to PHI (AUC ¼ 0.65; P ¼ 0.03), age (AUC ¼ 0.59; P ¼ 0.04) and PSAD (AUC ¼ 0.57; P ¼ 0.04). In contrast, PHI (AUC ¼ 0.80) remained the most accurate predictor when screening exclusively for significant PCa, and significantly outperformed PSAD (AUC ¼ 0.68; P ¼ 0.02), age (AUC ¼ 0.60; P ¼ 0.01) and PCA-3 (AUC ¼ 0.55; P ¼ 0.03). ROC curves depicting individual predictors for overall and significant PCa are shown in Figure 2. The Prostate
Among the PHI components, [2]proPSA was the most accurate at detecting PCa overall (AUC ¼ 0.64) and significant PCa (AUC ¼ 0.77) at an initial biopsy. Indeed, this PSA isoform significantly outperformed tPSA (AUC ¼ 0.53; P ¼ 0.003 and AUC ¼ 0.67; P ¼ 0.01, respectively) and fPSA (AUC ¼ 0.53; P ¼ 0.001 and AUC ¼ 0.61; P ¼ 0.02, respectively) in both cases. ROC curves depicting the ability of PHI to predict overall and significant PCa are shown in Figure 3. Diagnostic Performances of PSAD >0.15, PHI > 40, PCA-3 >35, and PHI > 40 Combined With PCA-3 >35 A detailed analysis of diagnostic performances of PSAD >0.15, PHI > 40, PCA-3 >35 and PHI > 40 combined with PCA-3 >35 for screening for overall and significant PCa is given in Table II. Sensitivity of PSAD >0.15 and specificity of PHI > 40 combined
76 (55.1) 62.5 [42.0–81.0] 13 (17.1) 47.2 [19.0–107.0] 3.9 [0.6–15.9] 0.08 [0.03–0.53] 32 (42.1) 44 (57.9) 1.6 [0.04–11.6] 20.8 [5.6–78.9] 17.6 [1.1–84.1] 31.1 [4.2–93.2] 51 (67.1) 25 (32.9) 30.8 [0.6–155.4] 52 (68.4) 24 (31.6) 65 (85.5) 11 (14.5)
110 (79.7) 28 (20.3)
Absence of PCa
138 (100) 63.4 [44.0–83.0] 26 (18.8) 46.4 [18.0–107.0] 7.8 [0.7–19.8] 0.17 [0.04–0.55] 52 (37.7) 86 (62.3) 1.7 [0.1–12.9] 22.2 [5.8–81.5] 21.9 [1.2–86.0] 42.4 [5.9–171.7] 86 (62.3) 52 (37.7) 49.2 [1.0–304.0] 73 (52.9) 65 (47.1)
Overall population
45 (72.6) 17 (27.4)
62 (44.9) 65.6 [45.0–83.0] 13 (20.9) 45.4 [17.0–105.0] 8.7 [0.7–19.8] 0.19 [0.03–0.56] 20 (32.3) 42 (67.7) 1.8 [0.3–13.1] 24.7 [6.4–83.6] 25.9 [1.6–89.4] 50.6 [8.7–187.2] 35 (56.5) 27 (43.5) 67.3 [18.1–311.2] 21 (33.9) 41 (66.1)
Presence of PCa
Screening for PCa overall
0.04
0.15 Median [fPSA], ng/ml [IQR] Median %[fPSA] [IQR] Median [2]proPSA, pg/ml [IQR] Median PHI [IQR] No PHI 40 (%) No PHI >40 (%) Median PCA-3 [IQR] No PCA-3 35(%) No PCA-3 >35(%) No PHI >40 þ PCA-3 >35(%) Absent Present
Characteristics
TABLE I. Descriptive Characteristics of the Study Population
0.02
0.01 0.03
0.18 0.07 0.15 35 had the best accuracy and positive or negative predictive values when screening for PCa overall whereas these diagnostic performances were greater for PHI >40 when exclusively screening for significant PCa. Predictors for PCa Overall and for Clinically Signif|cant PCa The biomarkers (i.e., PSAD >0.15, PHI >40, PCA-3 >35, PHI > 40 combined with PCA-3 >35) were tested in multivariate analyses for their effectiveness in overall and significant PCa screening. Only PCA-3 >35 (OR ¼ 5.68; 95%CI ¼ [2.21–14.59]; P < 0.001) The Prostate
Fig. 3. ROC curves depicting the ability of the Prostate Health Index (PHI) components to predict the presence of overall (A) and significant (B) prostate cancer (PCa) at aninitialbiopsy.
was significantly correlated with the presence of PCa overall at an initial biopsy, but PHI >40 (OR ¼ 9.60; 95%CI ¼ [1.72–91.32]; P ¼ 0.001) was the only independent predictor for detecting significant PCa (Table III). DISCUSSION The low accuracy of opportunistic PSA-based screening has led to the development of more specific plasma- and urine-based biomarkers, including PHI and PCA-3 scores, which have been widely described to improve detection of PCa compared to classic PSA testing [20–22]. To our knowledge, no significant
PHI vs. PCA-3 to Detect Overall and Significant Prostate Cancer
109
TABLE II. Sensitivity, Specif|city, and PredictiveValues of PSAD >0.15, PHI >40, and PCA-3 >35 for Screening for Overall and Signif|cant Prostate Cancer PCa overall Diagnostic performances Accuracy Sensitivity Specificity PPV NPV
Significant PCa
PSAD >0.15
PHI >40
PCA-3 >35
PHI >40 þ PCA-3 >35
PSAD >0.15
PHI >40
PCA-3 >35
PHI >40 þ PCA-3 >35
53.6 67.7 42.1 48.8 61.5
56.5 43.5 67.1 51.9 59.3
67.4 66.1 68.4 63.1 71.2
59.4 27.4 85.5 60.7 59.1
54.3 79.5 44.4 36.0 84.6
71.7 66.7 73.7 50.0 84.9
59.4 61.5 58.6 36.9 79.5
67.4 28.2 82.8 39.3 74.5
PCa, prostate cancer; PSAD, PSA density; PHI, Prostate Health Index; PCA-3, Prostate-Cancer Antigen-3 score; PPV, positivity predictive value; NPV, negative predictive value. Bold values are the greatest ones for each diagnostic performance when screening for PCa overall and significant PCa.
TABLE III. Multivariate Analyses of the Biomarkers to Predict Overall and Signif|cant Prostate Cancer at an Initial Biopsy Presence of PCa overall Biomarkers PSAD >0.15 PHI >40 PCA-3 >35 PHI >40 þ PCA-3 >35
Presence of significant PCa
OR [95% CI]
P
OR [95% CI]
P
1.35 [0.54–2.57] 2.07 [0.71–6.01] 5.68 [2.21–14.59] 0.62 [0.14–2.74]
0.53 0.21 35 remained the only independent predictor after adjusting for PSAD >0.15 and PHI >40 in multivariate analysis. However, we believe that the increased propensity of PCA-3 to detect PCa overall in our study was largely because of its potential to predict insignificant tumors [23]. This agrees with the recent concerns that have been raised regarding the interest of the PCA-3 score for a populationbased screening considering the actual trend to avoid over-diagnosis of low stage and grade PCa. The quest to find relevant biomarkers in PCa screening no longer consists of conducting a simple blood or urine test leading to a positive biopsy core. Indeed, there is now growing interest in screening only for significant
PCa. Interestingly, several authors have reported that PHI can be highly predictive at detecting the more aggressive PCa phenotype, and particularly high-grade disease [6,9,24]. Accordingly, we have demonstrated in our study that PHI was the most accurate predictor for significant PCa at an initial biopsy, and significantly outperformed the PCA-3 score. This finding was consistent in multivariate analysis, which found that a PHI >40 remained the only independent predictor after adjusting for PSAD >0.15 and PCA-3 >35. Consequently, these results suggested that instead of searching for completely different biomarkers, we should simply refine the way we use PSA testing by combining different accurate PSA-based measurements. Indeed, each individual component of PHI makes sense. In particular, [2]proPSA has been reported to be the best marker in terms of detecting significant PCa while also maintaining relevant sensitivity in population-based screening [25]. Similarly, we found that [2]proPSA was the most accurate biomarker amongst the PHI components when screening for PCa overall and for significant PCa but, logically, the combination of PHI components performed better than [2]proPSA, fPSA, or tPSA individually. The Prostate
110
Seisen et al.
Furthermore, the recent development of a PHI based-nomogram, which includes age, digital rectal examination, repeat-biopsy status, prostate volume, and PHI value, fits with the general idea that we should combine all relevant predictors of PCa during the biopsy process [26]. A rational approach to avoid over-diagnosis of indolent disease could be to only perform biopsies when there is high predicted probability of significant PCa based on the patient’s clinical, biochemical, but also morphological characteristics. Indeed, multiparametric magnetic-resonance imaging (mp-MRI) has previously demonstrated a highly negative predictive value for ruling out significant disease, especially Gleason patterns 4 or 5 [27]. However, we believe that candidates for mp-MRI should be initially selected using more adequate biomarkers than PSA testing, and our results suggest that it should be PHI. Two strengths of this study include its prospective design and the use of STARD guidelines to improve the accuracy and generalizability of the results. We limited selection bias by screening all prospective consecutive patients undergoing an initial prostate biopsy at our institution over a specific time period; also, the diagnostic procedures in this study adhered rigorously to welldefined protocols that included double-blind assessment of biomarkers and prostate-biopsy cores by a single referent physician, which avoided observer bias. However, we acknowledge that there are several limitations, such as its monocentric design in which only a small sample size is represented. Nevertheless, we chose to use standardized and widespread methods to perform biomarker assessment and prostate biopsies to guarantee reproducibility of the results.
CONCLUSION Although PCA-3 was the best predictor for PCa overall at initial biopsy, our findings strongly indicate that PHI should be used instead in population-based screening to avoid over-diagnosis of indolent tumors that are unlikely to cause death. We also consider that only patients with highly suspected significant PCa defined according PHI measurement, should be screened for a prostatic target using mp-MRI before undergoing any initial biopsies. Further prospective studies with a larger sample size are needed to confirm these results and to determine the optimal diagnostic sequence.
ACKNOWLEDGMENT The authors acknowledge Cecile Gaffory and Valerie Ondet for excellent technical assistance.
The Prostate
REFERENCES 1. Bill-Axelson A, Holmberg L, Ruutu M, Garmo H, Stark JR, Busch C, Nordling S, Haggman M, Andersson SO, Bratell S, Spangberg A, Palmgren J, Steineck G, Adami HO, Johansson JE, SPCG-4 Investigators. Radical prostatectomy versus watchful waiting in early prostate cancer. N Engl J Med 2011;364:1708–1717. 2. Schr€ oder FH, Hugosson J, Roobol MJ, Tammela TLJ, Ciatto S, Nelen V, Kwiatkowski M, Lujan M, Lilja H, Zappa M, Denis LJ, Recker F, Paez A, Maattanen L, Bangma CH, Aus G, Carlsson S, Villers A, Rebillard X, van der Kwast T, Kujala PM, Blijenberg BG, Stenman UH, Huber A, Taari K, Hakama M, Moss SM, de Koning HJ, Auvinen A, ERSPC Investigators. Prostate-cancer mortality at 11 years of follow-up. N Engl J Med 2012;366:981–990. 3. Andriole GL, Crawford ED, Grubb RL III, Buys SS, Chia D, Church TR, Fouad MN, Isaacs C, Kvale PA, Reding DJ, Weissfeld JL, Yokochi LA, O’Brien B, Ragard LR, CLapp JD, Rathmell JM, Riley TL, Hsing AW, Izmirlian G, Pinsky PF, Kramer BS, Miller AB, Gohagan JK, Prorock PC, PLCO Project Team. Prostate cancer screening in the randomized Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial: Mortality results after 13 years of follow-up. J Natl Cancer Inst 2012;104:125–132. €tt€ 4. Schr€ oder FH, Hugosson J, Carlsson S, Tammela T, M€ aa anen L, Auvinen A, Kwiatkowski M, Recker F, Roobol MJ. Screening for prostate cancer decreases the risk of developing metastatic disease: Findings from the European Randomized Study of Screening for Prostate Cancer (ERSPC). Eur Urol 2012;62:745–752. 5. Gittelman MC, Hertzman B, Bailen J, Williams T, Koziol I, Henderson RJ, Efros M, Bidair M, Ward JF. PCA3 molecular urine test as a predictor of repeat prostate biopsy outcome in men with previous negative biopsies: A prospective multicenter clinical study. J Urol 2013;190:64–69. 6. Lazzeri M, Haese A, de la Taille A, Palou Redorta J, McNicholas T, Lughezzani G, Scattoni V, Bini V, Freschi M, Sussman A, Ghaleh B, Le Corvoisier P, Alberola Bou J, Esquena Fernandez S, Graefen M, Guazzoni G. Serum isoform [2]proPSA derivatives significantly improve prediction of prostate cancer at initial biopsy in a total PSA range of 2–10 ng/ml: A multicentric European study. Eur Urol 2013;63:986–994. 7. Guazzoni G, Nava L, Lazzeri M, Scattoni V, Lughezzani G, Maccagnano C, Dorigatti F, Ceriotti F, Pontillo M, Bini V, Freschi M, Montorsi F, Rigatti P. Prostate-specific antigen (PSA) isoform p2PSA significantly improves the prediction of prostate cancer at initial extended prostate biopsies in patients with total PSA between 2.0 and 10 ng/ml: Results of a prospective study in a clinical setting. Eur Urol 2011;60:214–222. 8. Stephan C, Vincendeau S, Houlgatte A, Cammann H, Jung K, Semjonow A. Multicenter evaluation of [2]proprostate-specific antigen and the prostate health index for detecting prostate cancer. Clin Chem 2013;59:306–314. 9. Guazzoni G, Lazzeri M, Nava L, Lughezzani G, Larcher A, Scattoni V, Guadda GM, Bini V, Cestari A, Buffi NM, Freschi M, Rigatti P, Montorsi F. Preoperative prostate-specific antigen isoform p2PSA and its derivatives, %p2PSA and prostate health index, predict pathologic outcomes in patients undergoing radical prostatectomy for prostate cancer. Eur Urol 2012;61:455–466. 10. Ferro M, Bruzzese D, Perdonà S, Mazzarella C, Marino A, Sorrentino A, Di Carlo A, Autorino R, Di Lorenzo G, Buonerba C, Altieri V, Mariano A, Macchia V, Terracciano D. Predicting prostate biopsy outcome: Prostate health index (phi) and prostate cancer antigen 3 (PCA3) are useful biomarkers. Clin Chim Acta Int J Clin Chem 2012;413:1274–1278.
PHI vs. PCA-3 to Detect Overall and Significant Prostate Cancer 11. Stephan C, Jung K, Semjonow A, Schulze-Forster K, Cammann H, Hu X, Meyer HA, Bogemann M, Miller K, Friedersdorff F. Comparative assessment of urinary prostate cancer antigen 3 and TMPRSS2: ERG gene fusion with the serum [2]proprostatespecific antigen-based prostate health index for detection of prostate cancer. Clin Chem 2013;59:280–288. 12. Perdonà S, Bruzzese D, Ferro M, Autorino R, Marino A, Mazzarella C, Perruolo G, Longo M, Spinelli R, Di Lorenzo G, Oliva A, De Sio M, Damiano R, Altieri V, Terracciano D. Prostate health index (phi) and prostate cancer antigen 3 (PCA3) significantly improve diagnostic accuracy in patients undergoing prostate biopsy. Prostate 2013;73:227–235. 13. Scattoni V, Lazzeri M, Lughezzani G, De Luca S, Passera R, Bollito E, Randone D, Abdollah F, Capitanio U, Larcher A, Lista G, Gadda GM, Bini V, Montorsi F, Guazzoni G. Head-tohead comparison of prostate health index and urinary PCA3 for predicting cancer at initial or repeat biopsy. J Urol 2013;190: 496–501. 14. Ferro M, Bruzzese D, Perdonà S, Marino A, Mazzarella C, Perruolo G, D’Esposito V, Cosimato V, Buonerba C, Di Lorenzo G, Musi G, De Cobelli O, Chun FK, Terracciano D. Prostate Health Index (Phi) and Prostate Cancer Antigen 3 (PCA3) significantly improve prostate cancer detection at initial biopsy in a total PSA range of 2–10 ng/ml. PloS ONE 2013;8:e67687.
111
20. Jansen FH, van Schaik RHN, Kurstjens J, Horninger W, Klocker H, Bektic J, Wildhagen MF, Roobol MJ, Bangma CH, Bartsch G. Prostate-specific antigen (PSA) isoform p2PSA in combination with total PSA and free PSA improves diagnostic accuracy in prostate cancer detection. Eur Urol 2010;57:921–927. 21. Leyten GHJM, Hessels D, Jannink SA, Smit FP, de Jong H, Cornel EB, de Reijke TM, Vergunst H, Kil P, Knipscheer BC, van Oort IM, Mulders PF, Hulsbergen-van de Kaa CA, Schalken JA. Prospective multicentre evaluation of PCA3 and TMPRSS2-ERG gene fusions as diagnostic and prognostic urinary biomarkers for prostate cancer. Eur Urol 2014;65:534–542. 22. Chevli KK, Duff M, Walter P, Yu C, Capuder B, Elshafei A, Malczewski S, Kattan MW, Jones JS. Urinary PCA3 as a predictor of prostate cancer in a cohort of 3,073 men undergoing initial prostate biopsy. J Urol. In press. doi: 10.1016/j.juro.2013.12.005. 23. Auprich M, Chun FK-H, Ward JF, Pummer K, Babaian R, Augustin H, Luger F, Gutschi S, Budaus L, Fish M, Huland H, Graefen M, Haese A. Critical assessment of preoperative urinary prostate cancer antigen 3 on the accuracy of prostate cancer staging. Eur Urol 2011;59:96–105.
15. Semjonow A, K€ opke T, Eltze E, Pepping-Schefers B, Bürgel H, Darte C. Pre-analytical in-vitro stability of [2]proPSA in blood and serum. Clin Biochem 2010;43:926–928.
24. Catalona WJ, Partin AW, Sanda MG, Wei JT, Klee GG, Bangma CH, Slawin KM, Marks LS, Loeb S, Broyles DL, Shin SS, Cruz AB, Chan DW, Sokoll LJ, Roberts WL, va Schaik RH, Mizrahi IA. A multicenter study of [2]pro-prostate specific antigen combined with prostate specific antigen and free prostate specific antigen for prostate cancer detection in the 2.0 to 10.0 ng/ml prostate specific antigen range. J Urol 2011;185:1650–1655.
16. Groskopf J, Aubin SMJ, Deras IL, Blase A, Bodrug S, Clark C, Brentano S, Mathis J, Pham J, Meyer T, Cass M, Hodge P, Macairan ML, Marks LS, Rittenhouse H. APTIMA PCA3 molecular urine test: Development of a method to aid in the diagnosis of prostate cancer. Clin Chem 2006;52:1089–1095.
25. Catalona WJ, Bartsch G, Rittenhouse HG, Evans CL, Linton HJ, Horninger W, Klocker H, Mikolajczyk SD. Serum pro-prostate specific antigen preferentially detects aggressive prostate cancers in men with 2–4 ng/ml prostate specific antigen. J Urol 2004; 171:2239–2244.
17. Epstein JI, Allsbrook WC, Jr., Amin MB, Egevad LL, ISUP Grading Committee. The 2005 International Society of Urological Pathology (ISUP) Consensus Conference on Gleason Grading of Prostatic Carcinoma. Am J Surg Pathol 2005;29:1228–1242. 18. Ploussard G, Epstein JI, Montironi R, Carroll PR, Wirth M, Grimm M-O, Bjartell AS, Montorsi F, Freedland SJ, Erbersdobler A, van der Kwast TH. The contemporary concept of significant versus insignificant prostate cancer. Eur Urol 2011;60:291–303. 19. DeLong ER, DeLong DM, Clarke-Pearson DL. Comparing the areas under two or more correlated receiver operating characteristic curves: A nonparametric approach. Biometrics 1988;44: 837–845.
26. Lughezzani G, Lazzeri M, Larcher A, Lista G, Scattoni V, Cestari A, Buffi NM, Bini V, Guazzoni G. Development and internal validation of a Prostate Health Index based nomogram for predicting prostate cancer at extended biopsy. J Urol 2012;188: 1144–1150. 27. Hoeks CMA, Somford DM, van Oort IM, Vergunst H, Oddens JR, Smits GA, Roobol MJ, Bul M, Hambrock T, Witjes JA, Futterer JJ, Hulsbergen-van de Kaa CA, Barentsz JO. Value of 3-T multiparametric magnetic resonance imaging and magnetic resonance-guided biopsy for early risk restratification in active surveillance of low-risk prostate cancer: A prospective multicenter cohort study. Invest Radiol 2014;49:165–172.
The Prostate
Accuracy of the Prostate Health Index Versus the Urinary Prostate Cancer Antigen 3 Score to Predict Overall and Signif|cant Prostate Cancer at Initial Biopsy Thomas Seisen,1,2 Morgan Rouprêt,1,2* Didier Brault,2,3 Priscilla Leon,1,2 Geraldine Cancel-Tassin,2,4 Eva Comperat,5 Raphae€le Renard-Penna,6 Pierre Mozer,1 Jerome Guechot,3 and Olivier Cussenot1,2,4,7 1
2
Academic Department of Urology, Pitie-Salpetriere Hospital, AP-HP, Paris, France UPMCUniv Paris 06,GRC5,ONCOTYPE-URO, Institut Universitaire de Cancerologie, Paris, France 3 Academic Department of Biochemistry,GH Paris Est, AP-HP, Paris, France 4 CeRePP,Tenon Hospital, Paris, France 5 Academic Department of Pathology, Pitie-Salpetriere Hospital, AP-HP, Paris, France 6 Academic Department of Radiology, Pitie-Salpetriere Hospital, AP-HP, Paris, France 7 Academic Department of Urology,GH Paris Est, AP-HP, Paris, France BACKGROUND. It remains unclear whether the Prostate Health Index (PHI) or the urinary Prostate-Cancer Antigen 3 (PCA-3) score is more accurate at screening for prostate cancer (PCa). The aim of this study was to prospectively compare the accuracy of PHI and PCA-3 scores to predict overall and significant PCa in men undergoing an initial prostate biopsy. METHODS. Double-blind assessments of PHI and PCA-3 were conducted by referent physicians in 138 patients who subsequently underwent trans-rectal ultrasound-guided prostate biopsy according to a 12-core scheme. Predictive accuracies of PHI and PCA-3 were assessed using AUC and compared according to the DeLong method. Diagnostic performances with usual cut-off values for positivity (i.e., PHI >40 and PCA-3 >35) were calculated, and odds ratios associated with predicting PCa overall and significant PCa as defined by pathological updated Epstein criteria (i.e., Gleason score 7, more than three positive cores, or >50% cancer involvement in any core) were estimated using logistic regression. RESULTS. Prevalences of overall and significant PCa were 44.9% and 28.3%, respectively. PCA-3 (AUC ¼ 0.71) was the most accurate predictor of PCa overall, and significantly outperformed PHI (AUC ¼ 0.65; P ¼ 0.03). However, PHI (AUC ¼ 0.80) remained the most accurate predictor when screening exclusively for significant PCa and significantly outperformed PCA-3 (AUC ¼ 0.55; P ¼ 0.03). Furthermore, PCA-3 >35 had the best accuracy, and positive or negative predictive values when screening for PCa overall whereas these diagnostic performances were greater for PHI >40 when exclusively screening for significant PCa. PHI > 40 combined with PCA-3 > 35 was more specific in both cases. In multivariate analyses, PCA-3 >35 (OR ¼ 5.68; 95%CI ¼ [2.21–14.59]; P < 0.001) was significantly correlated with the presence of PCa overall, but PHI >40 (OR ¼ 9.60; 95%CI ¼ [1.72–91.32]; P ¼ 0.001) was the only independent predictor for detecting significant PCa. CONCLUSIONS. Although PCA-3 score is the best predictor for PCa overall at initial biopsy, our findings strongly indicate that PHI should be used for population-based screening to avoid overdiagnosis of indolent tumors that are unlikely to cause death. Prostate 75:103–111, 2015. # 2014 Wiley Periodicals, Inc.
Grant sponsor: Access Hybritech p2PSA ([-2]proPSA) reagents were provided by Beckman Coulter Inc. Correspondence to: Morgan Rouprêt, MD, PhD, Department of Urology, Pitie-Salpêtriere Hospital, 47-83 boulevard de l’hopital, 75013 Paris, France. E-mail: [email protected]
ß 2014 Wiley Periodicals, Inc.
Received 29 June 2014; Accepted 21 August 2014 DOI 10.1002/pros.22898 Published online 18 October 2014 in Wiley Online Library (wileyonlinelibrary.com).
104
Seisen et al. KEY WORDS: prostatic neoplasm; biopsy; early detection of cancer; biological markers; prostate-specific antigen; antigens
INTRODUCTION Recent debate on screening for prostate cancer (PCa) has focused on over-diagnosis and overtreatment of insignificant tumors known for their low propensity to cause death [1]. To evaluate the efficacy of population-based screening, two large randomized trials have been conducted but their conflicting results emphasized the rationale of the current controversy [2,3]. The European Randomized Study of Screening for Prostate Cancer (ERSPC) reported that screening for PCa was associated with a significant reduction in specific mortality after a median follow-up of 11 years. In contrast, the Prostate, Lung, Colorectal and Ovary (PLCO) project concluded that PCa-related mortality was very low when PCa was detected by screening and did not significantly differ from that of the standard-care population. However, a relative risk reduction of 42% for developing metastatic disease has increased interest in improving early detection using opportunistic prostate-specific antigen (PSA)-based screening [4]. Unfortunately, PSA testing has become a victim of its own success and no longer fulfills all its commitments. Over the last few years, this cornerstone of PCa screening has been often cited as producing too many negative prostate biopsies and leading to diagnosis of tumors that are unlikely to cause future health problems, even if left untreated. Consequently, the limited ability of PSA to detect significant PCa has prompted the development of several biomarkers that share the common objective of avoiding over-diagnosis and over-treatment. In this context, urinary molecular biomarkers have recently been proposed and, among these, the Prostate Cancer Antigen 3 (PCA-3) score is now used to guide indications for a repeat biopsy in cases where there is persistent suspicion of PCa after a previous negative biopsy [5]. Moreover, the new serum biomarkers, such as [2]proPSA and its derivative, the Prostate Health Index (PHI), have significantly increased the ability to detect PCa, especially at an initial biopsy [6]. The secretion of [2]proPSA appears to be more specific to tumor cells, and the combination of this PSA isoform with pertinent serum biomarkers, including total PSA (tPSA) and free PSA (fPSA), has resulted in the PHI, which has greatly improved discrimination between men with and without PCa and between those with significant and insignificant disease [7–9]. Despite several direct comparisons of the diagnostic effectiveness of these biomarkers, it remains unclear whether the PHI or the PCA-3 score should be used
The Prostate
for PCa screening. Although they are both more accurate than classic PSA testing, no significant difference has been reported between them in the setting of initial biopsies [10–14]. Thus, the aim of our study was to prospectively compare the diagnostic accuracy of PHI and PCA-3 scores to screen for overall and significant PCa in a cohort of men undergoing an initial prostate biopsy. MATERIALS AND METHODS Population Between June 2013 and December 2013, 177 consecutive patients undergoing a first prostate biopsy at our academic institution for suspected PCa, based on a tPSA ranging from 4 to 20 ng/ml and/or an abnormal digital rectal examination, were screened for inclusion in this prospective, observational, cohort study. Men receiving 5-a-reductase inhibitors (dutasteride or finasteride), largely known to affect tPSA level, and those with a history of previous invasive treatment for benign prostatic hyperplasia or acute bacterial prostatitis within the previous 3 months were excluded from the analyses. In addition, subjects with chronic renal failure, marked blood-protein alterations, hemophilia, or those previously polytransfused were excluded from the study, as these conditions may alter the concentrations of [2]proPSA. Overall, 138 men, aged >50 years, met the inclusion criteria and were enrolled in the current study after providing their written informed consent. This study received approval from the local hospital ethics committee and was performed according to the Standards for Reporting of Diagnostic Accuracy (STARD). Collection of Biomarkers Blood samples were drawn at the time of the first biopsy, just before any prostatic manipulations, to avoid any transient increase in pre-biopsy tPSA, fPSA, or [2]proPSA levels. After clotting the blood, serum aliquots were extracted by centrifugation within 3 hr and frozen at 80°C, according to the method of Semjonow et al. [15]. Specimens were processed using a UniCel1 DxI800 Immunoassay System analyzer, and serum biomarkers were assessed in a blinded fashion using a calibrated Hybritech1 Access2 Immunoassay System analyzer. To determine PHI, the following combination of the three previously assessed serum biomarker values was used: (([2]proPSA/fPSA) HtPSA).
PHI vs. PCA-3 to Detect Overall and Significant Prostate Cancer Urine samples were collected after a digital rectal examination using three strokes per lobe and immediately before a first prostate biopsy, as described by Groskopf et al. [16]. Specimens were processed and tested in a blinded fashion to quantify PCA-3 mRNA and PSA mRNA urinary concentrations using the Progensa1 PCA-3 assay. The PCA-3 score was calculated using the following formula: (PCA-3 mRNA/PSA mRNA) 1000. Prostate Biopsies After assessing for biomarkers, trans-rectal ultrasound-guided prostate biopsies were performed by a single senior urologist according to a standardized classic 12-core scheme. Trans-rectal ultrasound was also used to assess prostate volume and to target any significant abnormalities in the prostate gland during the procedure. All cores were mapped for location and submitted separately in specific single-core specimen containers filled with 10% buffered formalin for pathological evaluation. All specimens were processed and evaluated by our dedicated senior uro-pathologist who was blinded to the results from the urine and serum biomarkers. PCa was identified and graded according to definitions of the 2005 consensus conference of the International Society of Urological Pathology [17]. Significant PCa was defined using pathological updated Epstein criteria as follows: a Gleason score 7, more than three positive cores, or >50% cancer involvement in any core [18]. High-grade prostate intra-epithelial neoplasia or atypical small acinar proliferations were not considered as PCa. Study Endpoints The primary endpoint of the study was to directly compare the accuracy of PHI and PCA-3 scores to determine the most accurate biomarker to predict overall and significant PCa at an initial biopsy. We also compared the accuracy of these biomarkers with the PSAD to confirm that PHI and PCA-3 outperformed classic PSA testing when screening for overall and significant PCa. Statistical Analyses All statistical analyses were performed using R software (version 2.12.1; R Foundation for Statistical Computing, Vienna, Austria). Results were expressed as medians for continuous variables, and categorical variables were reported as percentages. The Shapiro– Wilk test assessed the normal distribution of variables. Student’s t-test and the Mann–Whitney U-test were used to compare normally and non-normally distribut-
105
ed continuous variables, respectively. The Chi-squared test was used to compare non-normally distributed categorical variables. All tests were two-sided with a P-value of 0.15, PHI >40, PCA-3 >35). Multivariate logistic regression analyses were performed to assess whether PCA-3 >35 and PHI >40 were independent predictors for the presence of overall and significant PCa at an initial biopsy.
RESULTS Participants The STARD flowchart is shown in Figure 1. The demographic characteristics of the study population are listed in Table I. Overall, PCa was diagnosed in 62 of the 138 men (44.9%) undergoing an initial biopsy, with a median PSAD of 0.17 (range: 0.02–0.55), median PHI of 42.4 (range: 5.9–171.7) and median PCA-3 of 49.2 (range: 1.0–304.0). When comparing patients with and without PCa, no significant difference was found in terms of clinical and biochemical characteristics except for PSAD, PHI, and PCA-3. Indeed, the proportions of patients presenting with PSAD >0.15 (67.7% vs. 57.9%; P ¼ 0.02), PHI >40 (43.5% vs. 32.9%; P ¼ 0.01), PCA-3 >35 (66.1% vs. 31.6%; P ¼ 0.01) and PHI > 40 combined with PCA-3 > 35 (27.4% vs. 14.5%; P ¼ 0.04) were significantly higher in the PCa group. Similarly, median PSAD (0.19 vs. 0.08; P ¼ 0.04), PHI (50.6 vs. 31.1; P ¼ 0.02), and PCA-3 (67.3 vs. 30.8; P < 0.01) significantly differed between men with and without PCa. Significant PCa was diagnosed in 39 of the 138 men (28.3%) undergoing an initial biopsy. When comparing patients with and without significant PCa, no significant difference was found in terms of clinical and biochemical characteristics except for PSAD, PHI, and PCA-3. Indeed, greater proportions of patients presented with PSAD >0.15 (79.5% vs. 55.6%; P < 0.01), PHI >40 (66.7% vs. 26.3%; P ¼ 0.01), PCA-3 The Prostate
106
Seisen et al.
Fig. 1. Flowchart showing patients inclusion process according to the Standards for Reporting of Diagnostic Accuracy (STARD). BPH, Benign prostatic hyperplasia; PCa, prostate cancer; S-PCa, significant PCa.
>35 (61.5% vs. 41.4%; P ¼ 0.03) and PHI > 40 combined with PCA-3 > 35 (28.2% vs. 17.2%; P ¼ 0.02) in the significant PCa group. Furthermore, patients with significant PCa had significantly higher median PSAD (0.26 vs. 0.13; P ¼ 0.02), PHI (65.7 vs. 36.5; P < 0.01), and PCA-3 (53.1 vs. 45.5; P ¼ 0.01). ROC-Curves Analysis In univariate accuracy analysis, PCA-3 (AUC ¼ 0.71) was the most accurate predictor of PCa overall at the initial biopsy compared to PHI (AUC ¼ 0.65; P ¼ 0.03), age (AUC ¼ 0.59; P ¼ 0.04) and PSAD (AUC ¼ 0.57; P ¼ 0.04). In contrast, PHI (AUC ¼ 0.80) remained the most accurate predictor when screening exclusively for significant PCa, and significantly outperformed PSAD (AUC ¼ 0.68; P ¼ 0.02), age (AUC ¼ 0.60; P ¼ 0.01) and PCA-3 (AUC ¼ 0.55; P ¼ 0.03). ROC curves depicting individual predictors for overall and significant PCa are shown in Figure 2. The Prostate
Among the PHI components, [2]proPSA was the most accurate at detecting PCa overall (AUC ¼ 0.64) and significant PCa (AUC ¼ 0.77) at an initial biopsy. Indeed, this PSA isoform significantly outperformed tPSA (AUC ¼ 0.53; P ¼ 0.003 and AUC ¼ 0.67; P ¼ 0.01, respectively) and fPSA (AUC ¼ 0.53; P ¼ 0.001 and AUC ¼ 0.61; P ¼ 0.02, respectively) in both cases. ROC curves depicting the ability of PHI to predict overall and significant PCa are shown in Figure 3. Diagnostic Performances of PSAD >0.15, PHI > 40, PCA-3 >35, and PHI > 40 Combined With PCA-3 >35 A detailed analysis of diagnostic performances of PSAD >0.15, PHI > 40, PCA-3 >35 and PHI > 40 combined with PCA-3 >35 for screening for overall and significant PCa is given in Table II. Sensitivity of PSAD >0.15 and specificity of PHI > 40 combined
76 (55.1) 62.5 [42.0–81.0] 13 (17.1) 47.2 [19.0–107.0] 3.9 [0.6–15.9] 0.08 [0.03–0.53] 32 (42.1) 44 (57.9) 1.6 [0.04–11.6] 20.8 [5.6–78.9] 17.6 [1.1–84.1] 31.1 [4.2–93.2] 51 (67.1) 25 (32.9) 30.8 [0.6–155.4] 52 (68.4) 24 (31.6) 65 (85.5) 11 (14.5)
110 (79.7) 28 (20.3)
Absence of PCa
138 (100) 63.4 [44.0–83.0] 26 (18.8) 46.4 [18.0–107.0] 7.8 [0.7–19.8] 0.17 [0.04–0.55] 52 (37.7) 86 (62.3) 1.7 [0.1–12.9] 22.2 [5.8–81.5] 21.9 [1.2–86.0] 42.4 [5.9–171.7] 86 (62.3) 52 (37.7) 49.2 [1.0–304.0] 73 (52.9) 65 (47.1)
Overall population
45 (72.6) 17 (27.4)
62 (44.9) 65.6 [45.0–83.0] 13 (20.9) 45.4 [17.0–105.0] 8.7 [0.7–19.8] 0.19 [0.03–0.56] 20 (32.3) 42 (67.7) 1.8 [0.3–13.1] 24.7 [6.4–83.6] 25.9 [1.6–89.4] 50.6 [8.7–187.2] 35 (56.5) 27 (43.5) 67.3 [18.1–311.2] 21 (33.9) 41 (66.1)
Presence of PCa
Screening for PCa overall
0.04
0.15 Median [fPSA], ng/ml [IQR] Median %[fPSA] [IQR] Median [2]proPSA, pg/ml [IQR] Median PHI [IQR] No PHI 40 (%) No PHI >40 (%) Median PCA-3 [IQR] No PCA-3 35(%) No PCA-3 >35(%) No PHI >40 þ PCA-3 >35(%) Absent Present
Characteristics
TABLE I. Descriptive Characteristics of the Study Population
0.02
0.01 0.03
0.18 0.07 0.15 35 had the best accuracy and positive or negative predictive values when screening for PCa overall whereas these diagnostic performances were greater for PHI >40 when exclusively screening for significant PCa. Predictors for PCa Overall and for Clinically Signif|cant PCa The biomarkers (i.e., PSAD >0.15, PHI >40, PCA-3 >35, PHI > 40 combined with PCA-3 >35) were tested in multivariate analyses for their effectiveness in overall and significant PCa screening. Only PCA-3 >35 (OR ¼ 5.68; 95%CI ¼ [2.21–14.59]; P < 0.001) The Prostate
Fig. 3. ROC curves depicting the ability of the Prostate Health Index (PHI) components to predict the presence of overall (A) and significant (B) prostate cancer (PCa) at aninitialbiopsy.
was significantly correlated with the presence of PCa overall at an initial biopsy, but PHI >40 (OR ¼ 9.60; 95%CI ¼ [1.72–91.32]; P ¼ 0.001) was the only independent predictor for detecting significant PCa (Table III). DISCUSSION The low accuracy of opportunistic PSA-based screening has led to the development of more specific plasma- and urine-based biomarkers, including PHI and PCA-3 scores, which have been widely described to improve detection of PCa compared to classic PSA testing [20–22]. To our knowledge, no significant
PHI vs. PCA-3 to Detect Overall and Significant Prostate Cancer
109
TABLE II. Sensitivity, Specif|city, and PredictiveValues of PSAD >0.15, PHI >40, and PCA-3 >35 for Screening for Overall and Signif|cant Prostate Cancer PCa overall Diagnostic performances Accuracy Sensitivity Specificity PPV NPV
Significant PCa
PSAD >0.15
PHI >40
PCA-3 >35
PHI >40 þ PCA-3 >35
PSAD >0.15
PHI >40
PCA-3 >35
PHI >40 þ PCA-3 >35
53.6 67.7 42.1 48.8 61.5
56.5 43.5 67.1 51.9 59.3
67.4 66.1 68.4 63.1 71.2
59.4 27.4 85.5 60.7 59.1
54.3 79.5 44.4 36.0 84.6
71.7 66.7 73.7 50.0 84.9
59.4 61.5 58.6 36.9 79.5
67.4 28.2 82.8 39.3 74.5
PCa, prostate cancer; PSAD, PSA density; PHI, Prostate Health Index; PCA-3, Prostate-Cancer Antigen-3 score; PPV, positivity predictive value; NPV, negative predictive value. Bold values are the greatest ones for each diagnostic performance when screening for PCa overall and significant PCa.
TABLE III. Multivariate Analyses of the Biomarkers to Predict Overall and Signif|cant Prostate Cancer at an Initial Biopsy Presence of PCa overall Biomarkers PSAD >0.15 PHI >40 PCA-3 >35 PHI >40 þ PCA-3 >35
Presence of significant PCa
OR [95% CI]
P
OR [95% CI]
P
1.35 [0.54–2.57] 2.07 [0.71–6.01] 5.68 [2.21–14.59] 0.62 [0.14–2.74]
0.53 0.21 35 remained the only independent predictor after adjusting for PSAD >0.15 and PHI >40 in multivariate analysis. However, we believe that the increased propensity of PCA-3 to detect PCa overall in our study was largely because of its potential to predict insignificant tumors [23]. This agrees with the recent concerns that have been raised regarding the interest of the PCA-3 score for a populationbased screening considering the actual trend to avoid over-diagnosis of low stage and grade PCa. The quest to find relevant biomarkers in PCa screening no longer consists of conducting a simple blood or urine test leading to a positive biopsy core. Indeed, there is now growing interest in screening only for significant
PCa. Interestingly, several authors have reported that PHI can be highly predictive at detecting the more aggressive PCa phenotype, and particularly high-grade disease [6,9,24]. Accordingly, we have demonstrated in our study that PHI was the most accurate predictor for significant PCa at an initial biopsy, and significantly outperformed the PCA-3 score. This finding was consistent in multivariate analysis, which found that a PHI >40 remained the only independent predictor after adjusting for PSAD >0.15 and PCA-3 >35. Consequently, these results suggested that instead of searching for completely different biomarkers, we should simply refine the way we use PSA testing by combining different accurate PSA-based measurements. Indeed, each individual component of PHI makes sense. In particular, [2]proPSA has been reported to be the best marker in terms of detecting significant PCa while also maintaining relevant sensitivity in population-based screening [25]. Similarly, we found that [2]proPSA was the most accurate biomarker amongst the PHI components when screening for PCa overall and for significant PCa but, logically, the combination of PHI components performed better than [2]proPSA, fPSA, or tPSA individually. The Prostate
110
Seisen et al.
Furthermore, the recent development of a PHI based-nomogram, which includes age, digital rectal examination, repeat-biopsy status, prostate volume, and PHI value, fits with the general idea that we should combine all relevant predictors of PCa during the biopsy process [26]. A rational approach to avoid over-diagnosis of indolent disease could be to only perform biopsies when there is high predicted probability of significant PCa based on the patient’s clinical, biochemical, but also morphological characteristics. Indeed, multiparametric magnetic-resonance imaging (mp-MRI) has previously demonstrated a highly negative predictive value for ruling out significant disease, especially Gleason patterns 4 or 5 [27]. However, we believe that candidates for mp-MRI should be initially selected using more adequate biomarkers than PSA testing, and our results suggest that it should be PHI. Two strengths of this study include its prospective design and the use of STARD guidelines to improve the accuracy and generalizability of the results. We limited selection bias by screening all prospective consecutive patients undergoing an initial prostate biopsy at our institution over a specific time period; also, the diagnostic procedures in this study adhered rigorously to welldefined protocols that included double-blind assessment of biomarkers and prostate-biopsy cores by a single referent physician, which avoided observer bias. However, we acknowledge that there are several limitations, such as its monocentric design in which only a small sample size is represented. Nevertheless, we chose to use standardized and widespread methods to perform biomarker assessment and prostate biopsies to guarantee reproducibility of the results.
CONCLUSION Although PCA-3 was the best predictor for PCa overall at initial biopsy, our findings strongly indicate that PHI should be used instead in population-based screening to avoid over-diagnosis of indolent tumors that are unlikely to cause death. We also consider that only patients with highly suspected significant PCa defined according PHI measurement, should be screened for a prostatic target using mp-MRI before undergoing any initial biopsies. Further prospective studies with a larger sample size are needed to confirm these results and to determine the optimal diagnostic sequence.
ACKNOWLEDGMENT The authors acknowledge Cecile Gaffory and Valerie Ondet for excellent technical assistance.
The Prostate
REFERENCES 1. Bill-Axelson A, Holmberg L, Ruutu M, Garmo H, Stark JR, Busch C, Nordling S, Haggman M, Andersson SO, Bratell S, Spangberg A, Palmgren J, Steineck G, Adami HO, Johansson JE, SPCG-4 Investigators. Radical prostatectomy versus watchful waiting in early prostate cancer. N Engl J Med 2011;364:1708–1717. 2. Schr€ oder FH, Hugosson J, Roobol MJ, Tammela TLJ, Ciatto S, Nelen V, Kwiatkowski M, Lujan M, Lilja H, Zappa M, Denis LJ, Recker F, Paez A, Maattanen L, Bangma CH, Aus G, Carlsson S, Villers A, Rebillard X, van der Kwast T, Kujala PM, Blijenberg BG, Stenman UH, Huber A, Taari K, Hakama M, Moss SM, de Koning HJ, Auvinen A, ERSPC Investigators. Prostate-cancer mortality at 11 years of follow-up. N Engl J Med 2012;366:981–990. 3. Andriole GL, Crawford ED, Grubb RL III, Buys SS, Chia D, Church TR, Fouad MN, Isaacs C, Kvale PA, Reding DJ, Weissfeld JL, Yokochi LA, O’Brien B, Ragard LR, CLapp JD, Rathmell JM, Riley TL, Hsing AW, Izmirlian G, Pinsky PF, Kramer BS, Miller AB, Gohagan JK, Prorock PC, PLCO Project Team. Prostate cancer screening in the randomized Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial: Mortality results after 13 years of follow-up. J Natl Cancer Inst 2012;104:125–132. €tt€ 4. Schr€ oder FH, Hugosson J, Carlsson S, Tammela T, M€ aa anen L, Auvinen A, Kwiatkowski M, Recker F, Roobol MJ. Screening for prostate cancer decreases the risk of developing metastatic disease: Findings from the European Randomized Study of Screening for Prostate Cancer (ERSPC). Eur Urol 2012;62:745–752. 5. Gittelman MC, Hertzman B, Bailen J, Williams T, Koziol I, Henderson RJ, Efros M, Bidair M, Ward JF. PCA3 molecular urine test as a predictor of repeat prostate biopsy outcome in men with previous negative biopsies: A prospective multicenter clinical study. J Urol 2013;190:64–69. 6. Lazzeri M, Haese A, de la Taille A, Palou Redorta J, McNicholas T, Lughezzani G, Scattoni V, Bini V, Freschi M, Sussman A, Ghaleh B, Le Corvoisier P, Alberola Bou J, Esquena Fernandez S, Graefen M, Guazzoni G. Serum isoform [2]proPSA derivatives significantly improve prediction of prostate cancer at initial biopsy in a total PSA range of 2–10 ng/ml: A multicentric European study. Eur Urol 2013;63:986–994. 7. Guazzoni G, Nava L, Lazzeri M, Scattoni V, Lughezzani G, Maccagnano C, Dorigatti F, Ceriotti F, Pontillo M, Bini V, Freschi M, Montorsi F, Rigatti P. Prostate-specific antigen (PSA) isoform p2PSA significantly improves the prediction of prostate cancer at initial extended prostate biopsies in patients with total PSA between 2.0 and 10 ng/ml: Results of a prospective study in a clinical setting. Eur Urol 2011;60:214–222. 8. Stephan C, Vincendeau S, Houlgatte A, Cammann H, Jung K, Semjonow A. Multicenter evaluation of [2]proprostate-specific antigen and the prostate health index for detecting prostate cancer. Clin Chem 2013;59:306–314. 9. Guazzoni G, Lazzeri M, Nava L, Lughezzani G, Larcher A, Scattoni V, Guadda GM, Bini V, Cestari A, Buffi NM, Freschi M, Rigatti P, Montorsi F. Preoperative prostate-specific antigen isoform p2PSA and its derivatives, %p2PSA and prostate health index, predict pathologic outcomes in patients undergoing radical prostatectomy for prostate cancer. Eur Urol 2012;61:455–466. 10. Ferro M, Bruzzese D, Perdonà S, Mazzarella C, Marino A, Sorrentino A, Di Carlo A, Autorino R, Di Lorenzo G, Buonerba C, Altieri V, Mariano A, Macchia V, Terracciano D. Predicting prostate biopsy outcome: Prostate health index (phi) and prostate cancer antigen 3 (PCA3) are useful biomarkers. Clin Chim Acta Int J Clin Chem 2012;413:1274–1278.
PHI vs. PCA-3 to Detect Overall and Significant Prostate Cancer 11. Stephan C, Jung K, Semjonow A, Schulze-Forster K, Cammann H, Hu X, Meyer HA, Bogemann M, Miller K, Friedersdorff F. Comparative assessment of urinary prostate cancer antigen 3 and TMPRSS2: ERG gene fusion with the serum [2]proprostatespecific antigen-based prostate health index for detection of prostate cancer. Clin Chem 2013;59:280–288. 12. Perdonà S, Bruzzese D, Ferro M, Autorino R, Marino A, Mazzarella C, Perruolo G, Longo M, Spinelli R, Di Lorenzo G, Oliva A, De Sio M, Damiano R, Altieri V, Terracciano D. Prostate health index (phi) and prostate cancer antigen 3 (PCA3) significantly improve diagnostic accuracy in patients undergoing prostate biopsy. Prostate 2013;73:227–235. 13. Scattoni V, Lazzeri M, Lughezzani G, De Luca S, Passera R, Bollito E, Randone D, Abdollah F, Capitanio U, Larcher A, Lista G, Gadda GM, Bini V, Montorsi F, Guazzoni G. Head-tohead comparison of prostate health index and urinary PCA3 for predicting cancer at initial or repeat biopsy. J Urol 2013;190: 496–501. 14. Ferro M, Bruzzese D, Perdonà S, Marino A, Mazzarella C, Perruolo G, D’Esposito V, Cosimato V, Buonerba C, Di Lorenzo G, Musi G, De Cobelli O, Chun FK, Terracciano D. Prostate Health Index (Phi) and Prostate Cancer Antigen 3 (PCA3) significantly improve prostate cancer detection at initial biopsy in a total PSA range of 2–10 ng/ml. PloS ONE 2013;8:e67687.
111
20. Jansen FH, van Schaik RHN, Kurstjens J, Horninger W, Klocker H, Bektic J, Wildhagen MF, Roobol MJ, Bangma CH, Bartsch G. Prostate-specific antigen (PSA) isoform p2PSA in combination with total PSA and free PSA improves diagnostic accuracy in prostate cancer detection. Eur Urol 2010;57:921–927. 21. Leyten GHJM, Hessels D, Jannink SA, Smit FP, de Jong H, Cornel EB, de Reijke TM, Vergunst H, Kil P, Knipscheer BC, van Oort IM, Mulders PF, Hulsbergen-van de Kaa CA, Schalken JA. Prospective multicentre evaluation of PCA3 and TMPRSS2-ERG gene fusions as diagnostic and prognostic urinary biomarkers for prostate cancer. Eur Urol 2014;65:534–542. 22. Chevli KK, Duff M, Walter P, Yu C, Capuder B, Elshafei A, Malczewski S, Kattan MW, Jones JS. Urinary PCA3 as a predictor of prostate cancer in a cohort of 3,073 men undergoing initial prostate biopsy. J Urol. In press. doi: 10.1016/j.juro.2013.12.005. 23. Auprich M, Chun FK-H, Ward JF, Pummer K, Babaian R, Augustin H, Luger F, Gutschi S, Budaus L, Fish M, Huland H, Graefen M, Haese A. Critical assessment of preoperative urinary prostate cancer antigen 3 on the accuracy of prostate cancer staging. Eur Urol 2011;59:96–105.
15. Semjonow A, K€ opke T, Eltze E, Pepping-Schefers B, Bürgel H, Darte C. Pre-analytical in-vitro stability of [2]proPSA in blood and serum. Clin Biochem 2010;43:926–928.
24. Catalona WJ, Partin AW, Sanda MG, Wei JT, Klee GG, Bangma CH, Slawin KM, Marks LS, Loeb S, Broyles DL, Shin SS, Cruz AB, Chan DW, Sokoll LJ, Roberts WL, va Schaik RH, Mizrahi IA. A multicenter study of [2]pro-prostate specific antigen combined with prostate specific antigen and free prostate specific antigen for prostate cancer detection in the 2.0 to 10.0 ng/ml prostate specific antigen range. J Urol 2011;185:1650–1655.
16. Groskopf J, Aubin SMJ, Deras IL, Blase A, Bodrug S, Clark C, Brentano S, Mathis J, Pham J, Meyer T, Cass M, Hodge P, Macairan ML, Marks LS, Rittenhouse H. APTIMA PCA3 molecular urine test: Development of a method to aid in the diagnosis of prostate cancer. Clin Chem 2006;52:1089–1095.
25. Catalona WJ, Bartsch G, Rittenhouse HG, Evans CL, Linton HJ, Horninger W, Klocker H, Mikolajczyk SD. Serum pro-prostate specific antigen preferentially detects aggressive prostate cancers in men with 2–4 ng/ml prostate specific antigen. J Urol 2004; 171:2239–2244.
17. Epstein JI, Allsbrook WC, Jr., Amin MB, Egevad LL, ISUP Grading Committee. The 2005 International Society of Urological Pathology (ISUP) Consensus Conference on Gleason Grading of Prostatic Carcinoma. Am J Surg Pathol 2005;29:1228–1242. 18. Ploussard G, Epstein JI, Montironi R, Carroll PR, Wirth M, Grimm M-O, Bjartell AS, Montorsi F, Freedland SJ, Erbersdobler A, van der Kwast TH. The contemporary concept of significant versus insignificant prostate cancer. Eur Urol 2011;60:291–303. 19. DeLong ER, DeLong DM, Clarke-Pearson DL. Comparing the areas under two or more correlated receiver operating characteristic curves: A nonparametric approach. Biometrics 1988;44: 837–845.
26. Lughezzani G, Lazzeri M, Larcher A, Lista G, Scattoni V, Cestari A, Buffi NM, Bini V, Guazzoni G. Development and internal validation of a Prostate Health Index based nomogram for predicting prostate cancer at extended biopsy. J Urol 2012;188: 1144–1150. 27. Hoeks CMA, Somford DM, van Oort IM, Vergunst H, Oddens JR, Smits GA, Roobol MJ, Bul M, Hambrock T, Witjes JA, Futterer JJ, Hulsbergen-van de Kaa CA, Barentsz JO. Value of 3-T multiparametric magnetic resonance imaging and magnetic resonance-guided biopsy for early risk restratification in active surveillance of low-risk prostate cancer: A prospective multicenter cohort study. Invest Radiol 2014;49:165–172.
The Prostate
