Original Paper
Urologia
Published online: June 3, 2015
Urol Int DOI: 10.1159/000431233
Internationalis
Prediction of Significant Prostate Cancer at Prostate Biopsy and Per Core Detection Rate of Targeted and Systematic Biopsies Using Real-Time Shear Wave Elastography Katharina Boehm a, d Lars Budäus a Pierre Tennstedt a Burkhard Beyer a Jonas Schiffmann a Alessandro Larcher d, e Kathrin Simonis a, b Markus Graefen a Dirk Beyersdorff a, c Georg Salomon a
a
Martini-Clinic, Prostate Cancer Centre, b Department of Urology and c Diagnostic and Interventional Radiology Department and Clinic, University Medical Center Hamburg-Eppendorf, Hamburg, Germany; d Cancer Prognostics and Health Outcomes Unit, University of Montreal Health Center, Montreal, Que., Canada; e Division of Oncology, Unit of Urology, URI, IRCCS Ospedale San Raffaele, Milan, Italy
Abstract Introduction: Prostate cancer (PCa) detection is accompanied by overdiagnosis and mischaracterization of PCa. Therefore, new imaging modalities like shear wave elastography (SWE) are required. Aim: The aim of this study was to evaluate per-core detection rates (DRs) of targeted biopsies and systematic biopsies and to test if SWE findings can predict presence of clinically significant PCa (csPCa) at biopsy. Patients and Methods: Overall, 95 patients scheduled for prostate biopsy in our center underwent SWE. SWE findings were classified into suspicious or normal. Targeted biopsies were taken in up to 3 SWE-suspicious areas. csPCa was defined as the presence of Gleason pattern ≥4, level of prostate-specific antigen ≥10 ng/ml or >2 positive cores. Results: Overall DR for csPCa in our study cohort was 40%. Per-core DR for exclusively SWE-targeted cores versus systematic samples cores was 10.5 vs. 8.6% (p = 0.3). In the logistic regression models,
© 2015 S. Karger AG, Basel 0042–1138/15/0000–0000$39.50/0 E-Mail [email protected] www.karger.com/uin
individuals with suspicious SWE findings are at 6.4-fold higher risk of harboring csPCa (p = 0.03). Gain in predictive accuracy was 2.3% (0.82 vs. 0.84, p = 0.01). Conclusions: Presence of suspicious SWE findings is an independent predictor of csPCa. Therefore, SWE may be helpful in selecting patients for biopsy. Nonetheless, per-core DR for SWE-targeted cores was not statistically significant higher than DR of systematic sampled cores. Therefore, additional systematic biopsy is © 2015 S. Karger AG, Basel mandatory.
Introduction
Diagnosis of prostate cancer (PCa) relies on the level of prostate-specific antigen (PSA) and the digital rectal examination (DRE). Additionally, PCa diagnosis is the only solid cancer diagnosis, based on non-targeted biopsy schemes [1]. This might be an explanation for 2 major is-
K.B. and L.B. contributed equally to this work.
Katharina Boehm, MD Cancer Prognostics and Health Outcomes Unit 264 Blvd. Rene-Levesque E. Room 228 Montreal, QC H2X 1P1 (Canada) E-Mail boehm @ martini-klinik.de
Downloaded by: Univ. of California San Diego 132.239.1.230 - 6/11/2015 3:06:43 AM
Key Words Biopsy · Diagnostic · Elastography · Imaging · Prostate cancer
Color version available online
Transducer
Acoustic push pulses Shear waves
Region of interest
a
b
Fig. 1. a Principle of SWE: the transducer sends acoustic push pulses in the tissue. Compared to strain elastography, manually induced pressure by the examiner is not necessary. b Reproducible suspicious lesion in SWE imaging (red).
2
Urol Int DOI: 10.1159/000431233
were compared in SWE-targeted biopsies and systematic sampled biopsies, to test the hypothesis that SWE-targeted biopsies can increase the overall DR of csPCa.
Methods Participants Between January 2013 and November 2013, 109 consecutive patients were included in this study after obtaining informed consent. The study has been performed in accordance with the ethical standards laid down by the 1964 Declaration of Helsinki and its later amendments. A total of 14 patients were excluded. Specifically, 3 patients with no information about SWE finding stratification, 1 patient with a history of transurethral resection of the prostate and 10 patients with no information about the number of prior biopsies undergone were excluded (fig. 2). The remaining 95 patients were scheduled for 12-core prostate biopsy in a single tertiary center. In concordance to the European guidelines for detection of PCa, indication for biopsy was elevated PSA levels (≥4 ng/ml) or suspicious DRE [1]. Patients with proven PCa in history, for example, patients under active surveillance or those who had undergone primary radiation therapy, were not included. Subsequently, patients were stratified according to suspicious and not suspicious SWE findings. Definition of csPCa was the presence of Gleason pattern 4, >2 biopsy cores affected by any PCa or PSA levels ≥10 ng/ml. Data were prospectively collected and stored in an institutional database (FileMaker pro 10; FileMaker, Inc.). SWE Principles An experienced single observer (>50 SWE procedures, >500 transrectal biopsies) performed all SWE examinations, using the TRUS Aixplorer System (Supersonic Imagine, Aix-en-Provence,
Boehm et al.
Downloaded by: Univ. of California San Diego 132.239.1.230 - 6/11/2015 3:06:43 AM
sues in PCa diagnosis, namely overdiagnosis and mischaracterization of the tumor [2–4]. Therefore, new imaging modalities are required. However, the present role of ultrasound-based targeting procedures still remains unclear [5]. To confront the issue of overdiagnosis, some authors report the detection rates (DRs) of clinically significant PCa (csPCa) [6–8]. Definitions of csPCa are inconsistent, but most study groups define Gleason 3 + 3 disease, a PSA level 2 positive cores at biopsy.
cs prostate cancer n = 36
France). The SWE principle is based on measuring waves resulting from the reaction of prostate tissue on acoustic push pulses through the ultrasound transducer [14, 15]. Thereby, stiffness of the tissue can be calculated, and stiffer areas are visualized (fig. 1a). Specifically, stiffness of the tissue is visualized in a color scale from blue (soft tissue, 0 kPa) to red (hard tissue, >90 kPa). Prior to biopsy, the observer classified all prostate glands into suspicious or not suspicious for harboring PCa (fig. 1b). To determine if stiffness values were elevated and therefore suspicious, the color scale was used. Previous reported cutoffs of 50 kPa for overall stiffness or a density ratio between surrounding tissue and suspicious area of ≥2.2 were used to define suspicious areas [11]. Suspicious areas had to be reproducible on the screen (fig. 1b). The SWE examination focused on the peripheral zone of the prostate as most PCa presents in this area, and in comparison, stiffness values in the central zone are generally higher [11].
Normal tissue n = 38
cs prostate cancer n=2
Normal tissue n = 19
Pathological Principles Genitourinary pathologists analyzed all biopsy cores separately and recorded origin of the core, length, percentage of tumor, Gleason grade and sum of every single core [17].
Biopsy Principles All ultrasound procedures were performed in the left lateral decubital position in a standardized fashion. SWE was performed prior taking the targeted biopsies. After performance of SWE, targeted biopsies were taken in up to 3 SWE-suspicious areas. Another 3 cores were sampled from hypoechogenic areas in gray scale ultrasound, followed by 6 to 18 systematic biopsies, depending on the number of previous negative biopsies. Generally, patients with no previous biopsy underwent the 12-core biopsy (6 systematic, 3 SWE-targeted and 3 gray-scale-targeted cores). Conversely, in patients with previous biopsies, the number of systematic sampled cores was escalated to 12 systematic sampled cores. As the peripheral zone of the prostate is at higher risk of harboring PCa than other areas, systematic biopsies were sampled from the apex, midgland and base of the prostate [1, 16].
Statistical Methods Baseline characteristics of the patients with normal and suspicious findings in the SWE examination were compared, using the Chi-square likelihood test for categorical variables and the Wilcoxon test for continuous variables. First, we compared the DRs per biopsy core stratified by the targeting method (SWE targeted, hypoechogenic targeted or systematically targeted). Second, we computed the DR per patient. Per-core DRs were compared using the Chi-square likelihood test. To compare the SWE DR to the systematic biopsy DR by patient we used the McNemar Chi-square test. A risk factor analysis for detection of csPCa was performed. Subsequently, 2 multivariable logistic regression models were fitted for the prediction of csPCa diagnosis at biopsy. The base model included PSA level, age at biopsy, prostate volume, findings of the DRE and number of prior biopsies. The second model additionally included the findings of the SWE examination. Predictive accuracy for both models was quantified as the area under the receiver operating characteristic curve (AUC). To evaluate whether including SWE findings in the predictive model improved the accuracy of the prognostic model and if this imaging procedure might be helpful for the clinical management of patients, we used decision curve analysis (DCA), as previously described [18]. Decision curves depict the net benefit against the threshold probability. As an example for different threshold probabilities in our study, some patients might opt for biopsy if they had a 10% risk of cancer, others only if the risk was 20% or higher. In DCA, the net benefit is calculated across a range of probabilities. Additionally, we con-
Shear Wave Elastography for Prostate Biopsy
Urol Int DOI: 10.1159/000431233
3
Downloaded by: Univ. of California San Diego 132.239.1.230 - 6/11/2015 3:06:43 AM
Fig. 2. Workflow of 109 individuals sched-
Normal SWE n = 21
Table 1. Characteristics of 95 patients scheduled for prostate biopsy between January 2013 and November 2013,
stratified according to suspicious or normal findings in the SWE examination
Patients, n Age, years, median (IQR) Prostate volume, median (IQR) PSA level, ng/ml, median (IQR) Biopsy result Positive Negative csPCa Yes No Cores, n, median (range) DRE Negative Positive Previous biopsies First biopsy Repeated biopsy
Overall
Patients with suspicious SWE
Patients without suspicious SWE
p value
95 67 (61–72) 50 (36–69) 6.7 (4.9–9.3)
74 (77.9) 67 (60–72) 50 (36–68) 7.1 (5.1–9.7)
21 (22.1) 65 (62–68) 56 (38–92) 6.0 (4.3–9.4)
47 (49) 48 (51)
42 (57) 32 (43)
5 (24) 16 (76)
38 (40) 57 (60) 12 (9–14)
36 (49) 38 (51) 12 (10–14)
2 (9.5) 19 (90.5) 12 (9–14)
80 (84) 15 (16)
59 (80) 15 (20)
21 (100) 0 (0)
Urologia
Published online: June 3, 2015
Urol Int DOI: 10.1159/000431233
Internationalis
Prediction of Significant Prostate Cancer at Prostate Biopsy and Per Core Detection Rate of Targeted and Systematic Biopsies Using Real-Time Shear Wave Elastography Katharina Boehm a, d Lars Budäus a Pierre Tennstedt a Burkhard Beyer a Jonas Schiffmann a Alessandro Larcher d, e Kathrin Simonis a, b Markus Graefen a Dirk Beyersdorff a, c Georg Salomon a
a
Martini-Clinic, Prostate Cancer Centre, b Department of Urology and c Diagnostic and Interventional Radiology Department and Clinic, University Medical Center Hamburg-Eppendorf, Hamburg, Germany; d Cancer Prognostics and Health Outcomes Unit, University of Montreal Health Center, Montreal, Que., Canada; e Division of Oncology, Unit of Urology, URI, IRCCS Ospedale San Raffaele, Milan, Italy
Abstract Introduction: Prostate cancer (PCa) detection is accompanied by overdiagnosis and mischaracterization of PCa. Therefore, new imaging modalities like shear wave elastography (SWE) are required. Aim: The aim of this study was to evaluate per-core detection rates (DRs) of targeted biopsies and systematic biopsies and to test if SWE findings can predict presence of clinically significant PCa (csPCa) at biopsy. Patients and Methods: Overall, 95 patients scheduled for prostate biopsy in our center underwent SWE. SWE findings were classified into suspicious or normal. Targeted biopsies were taken in up to 3 SWE-suspicious areas. csPCa was defined as the presence of Gleason pattern ≥4, level of prostate-specific antigen ≥10 ng/ml or >2 positive cores. Results: Overall DR for csPCa in our study cohort was 40%. Per-core DR for exclusively SWE-targeted cores versus systematic samples cores was 10.5 vs. 8.6% (p = 0.3). In the logistic regression models,
© 2015 S. Karger AG, Basel 0042–1138/15/0000–0000$39.50/0 E-Mail [email protected] www.karger.com/uin
individuals with suspicious SWE findings are at 6.4-fold higher risk of harboring csPCa (p = 0.03). Gain in predictive accuracy was 2.3% (0.82 vs. 0.84, p = 0.01). Conclusions: Presence of suspicious SWE findings is an independent predictor of csPCa. Therefore, SWE may be helpful in selecting patients for biopsy. Nonetheless, per-core DR for SWE-targeted cores was not statistically significant higher than DR of systematic sampled cores. Therefore, additional systematic biopsy is © 2015 S. Karger AG, Basel mandatory.
Introduction
Diagnosis of prostate cancer (PCa) relies on the level of prostate-specific antigen (PSA) and the digital rectal examination (DRE). Additionally, PCa diagnosis is the only solid cancer diagnosis, based on non-targeted biopsy schemes [1]. This might be an explanation for 2 major is-
K.B. and L.B. contributed equally to this work.
Katharina Boehm, MD Cancer Prognostics and Health Outcomes Unit 264 Blvd. Rene-Levesque E. Room 228 Montreal, QC H2X 1P1 (Canada) E-Mail boehm @ martini-klinik.de
Downloaded by: Univ. of California San Diego 132.239.1.230 - 6/11/2015 3:06:43 AM
Key Words Biopsy · Diagnostic · Elastography · Imaging · Prostate cancer
Color version available online
Transducer
Acoustic push pulses Shear waves
Region of interest
a
b
Fig. 1. a Principle of SWE: the transducer sends acoustic push pulses in the tissue. Compared to strain elastography, manually induced pressure by the examiner is not necessary. b Reproducible suspicious lesion in SWE imaging (red).
2
Urol Int DOI: 10.1159/000431233
were compared in SWE-targeted biopsies and systematic sampled biopsies, to test the hypothesis that SWE-targeted biopsies can increase the overall DR of csPCa.
Methods Participants Between January 2013 and November 2013, 109 consecutive patients were included in this study after obtaining informed consent. The study has been performed in accordance with the ethical standards laid down by the 1964 Declaration of Helsinki and its later amendments. A total of 14 patients were excluded. Specifically, 3 patients with no information about SWE finding stratification, 1 patient with a history of transurethral resection of the prostate and 10 patients with no information about the number of prior biopsies undergone were excluded (fig. 2). The remaining 95 patients were scheduled for 12-core prostate biopsy in a single tertiary center. In concordance to the European guidelines for detection of PCa, indication for biopsy was elevated PSA levels (≥4 ng/ml) or suspicious DRE [1]. Patients with proven PCa in history, for example, patients under active surveillance or those who had undergone primary radiation therapy, were not included. Subsequently, patients were stratified according to suspicious and not suspicious SWE findings. Definition of csPCa was the presence of Gleason pattern 4, >2 biopsy cores affected by any PCa or PSA levels ≥10 ng/ml. Data were prospectively collected and stored in an institutional database (FileMaker pro 10; FileMaker, Inc.). SWE Principles An experienced single observer (>50 SWE procedures, >500 transrectal biopsies) performed all SWE examinations, using the TRUS Aixplorer System (Supersonic Imagine, Aix-en-Provence,
Boehm et al.
Downloaded by: Univ. of California San Diego 132.239.1.230 - 6/11/2015 3:06:43 AM
sues in PCa diagnosis, namely overdiagnosis and mischaracterization of the tumor [2–4]. Therefore, new imaging modalities are required. However, the present role of ultrasound-based targeting procedures still remains unclear [5]. To confront the issue of overdiagnosis, some authors report the detection rates (DRs) of clinically significant PCa (csPCa) [6–8]. Definitions of csPCa are inconsistent, but most study groups define Gleason 3 + 3 disease, a PSA level 2 positive cores at biopsy.
cs prostate cancer n = 36
France). The SWE principle is based on measuring waves resulting from the reaction of prostate tissue on acoustic push pulses through the ultrasound transducer [14, 15]. Thereby, stiffness of the tissue can be calculated, and stiffer areas are visualized (fig. 1a). Specifically, stiffness of the tissue is visualized in a color scale from blue (soft tissue, 0 kPa) to red (hard tissue, >90 kPa). Prior to biopsy, the observer classified all prostate glands into suspicious or not suspicious for harboring PCa (fig. 1b). To determine if stiffness values were elevated and therefore suspicious, the color scale was used. Previous reported cutoffs of 50 kPa for overall stiffness or a density ratio between surrounding tissue and suspicious area of ≥2.2 were used to define suspicious areas [11]. Suspicious areas had to be reproducible on the screen (fig. 1b). The SWE examination focused on the peripheral zone of the prostate as most PCa presents in this area, and in comparison, stiffness values in the central zone are generally higher [11].
Normal tissue n = 38
cs prostate cancer n=2
Normal tissue n = 19
Pathological Principles Genitourinary pathologists analyzed all biopsy cores separately and recorded origin of the core, length, percentage of tumor, Gleason grade and sum of every single core [17].
Biopsy Principles All ultrasound procedures were performed in the left lateral decubital position in a standardized fashion. SWE was performed prior taking the targeted biopsies. After performance of SWE, targeted biopsies were taken in up to 3 SWE-suspicious areas. Another 3 cores were sampled from hypoechogenic areas in gray scale ultrasound, followed by 6 to 18 systematic biopsies, depending on the number of previous negative biopsies. Generally, patients with no previous biopsy underwent the 12-core biopsy (6 systematic, 3 SWE-targeted and 3 gray-scale-targeted cores). Conversely, in patients with previous biopsies, the number of systematic sampled cores was escalated to 12 systematic sampled cores. As the peripheral zone of the prostate is at higher risk of harboring PCa than other areas, systematic biopsies were sampled from the apex, midgland and base of the prostate [1, 16].
Statistical Methods Baseline characteristics of the patients with normal and suspicious findings in the SWE examination were compared, using the Chi-square likelihood test for categorical variables and the Wilcoxon test for continuous variables. First, we compared the DRs per biopsy core stratified by the targeting method (SWE targeted, hypoechogenic targeted or systematically targeted). Second, we computed the DR per patient. Per-core DRs were compared using the Chi-square likelihood test. To compare the SWE DR to the systematic biopsy DR by patient we used the McNemar Chi-square test. A risk factor analysis for detection of csPCa was performed. Subsequently, 2 multivariable logistic regression models were fitted for the prediction of csPCa diagnosis at biopsy. The base model included PSA level, age at biopsy, prostate volume, findings of the DRE and number of prior biopsies. The second model additionally included the findings of the SWE examination. Predictive accuracy for both models was quantified as the area under the receiver operating characteristic curve (AUC). To evaluate whether including SWE findings in the predictive model improved the accuracy of the prognostic model and if this imaging procedure might be helpful for the clinical management of patients, we used decision curve analysis (DCA), as previously described [18]. Decision curves depict the net benefit against the threshold probability. As an example for different threshold probabilities in our study, some patients might opt for biopsy if they had a 10% risk of cancer, others only if the risk was 20% or higher. In DCA, the net benefit is calculated across a range of probabilities. Additionally, we con-
Shear Wave Elastography for Prostate Biopsy
Urol Int DOI: 10.1159/000431233
3
Downloaded by: Univ. of California San Diego 132.239.1.230 - 6/11/2015 3:06:43 AM
Fig. 2. Workflow of 109 individuals sched-
Normal SWE n = 21
Table 1. Characteristics of 95 patients scheduled for prostate biopsy between January 2013 and November 2013,
stratified according to suspicious or normal findings in the SWE examination
Patients, n Age, years, median (IQR) Prostate volume, median (IQR) PSA level, ng/ml, median (IQR) Biopsy result Positive Negative csPCa Yes No Cores, n, median (range) DRE Negative Positive Previous biopsies First biopsy Repeated biopsy
Overall
Patients with suspicious SWE
Patients without suspicious SWE
p value
95 67 (61–72) 50 (36–69) 6.7 (4.9–9.3)
74 (77.9) 67 (60–72) 50 (36–68) 7.1 (5.1–9.7)
21 (22.1) 65 (62–68) 56 (38–92) 6.0 (4.3–9.4)
47 (49) 48 (51)
42 (57) 32 (43)
5 (24) 16 (76)
38 (40) 57 (60) 12 (9–14)
36 (49) 38 (51) 12 (10–14)
2 (9.5) 19 (90.5) 12 (9–14)
80 (84) 15 (16)
59 (80) 15 (20)
21 (100) 0 (0)
