Original Paper
Urologia
Received: June 15, 2015 Accepted after revision: August 31, 2015 Published online: November 11, 2015
Urol Int 2016;96:194–201 DOI: 10.1159/000440814
Internationalis
Sextant-Specific Analysis of Detection and Tumor Volume by HistoScanningTM Daniel Porres a Timur Hasan Kuru a Robin Epplen a Andreas Eck e Vahudin Zugor e Lieven Niels Kennes c Samir Afram d Till Braunschweig b Ruth Knüchel-Clarke b David Pfister a Axel Heidenreich a
Departments of a Urology, b Pathology, and c Medical Statistics, RWTH University, Aachen, Departments of d Pathology and e Urology, St. Antonius-Hospital, Gronau, Germany
Key Words Prostate cancer · Diagnostic · Ultrasound · HistoScanning
ing to pre-operational PSA values did not improve performance characteristics of HS. Conclusions: Our results do not support a significant contribution of HS to PCa diagnostics. © 2015 S. Karger AG, Basel
© 2015 S. Karger AG, Basel 0042–1138/15/0962–0194$39.50/0 E-Mail [email protected] www.karger.com/uin
Introduction
In the search of optimized diagnostics for prostate cancer (PCa), HistoScanning (HS) has been introduced to support the detection and assessment of cancer foci. In 2008, Braeckman et al. [1] reported their first promising results with this technique. However, later study by Schiffmann et al. [2] was not able to reproduce these observations. He was not able to show a significant correlation between tumor volume by HS and radical prostatectomy (RP) specimen in a cohort of 148 patients. In the mean time, other diagnostic tools like elastography, multiparametric magnetic-resonance imaging (mp-MRI) and MRI/transrectal ultrasound (TRUS) fusion biopsies showed promising results in PCa detection [3–8]. This study aims at validating the previous results of HS on the gold standard of RP specimen and put them into the context of other available diagnostic modalities for PCa detection. Daniel Porres, MD Department of Urology, RWTH University Pauwelsstr. 30, 52074 Aachen (Germany) E-Mail daniel.porres @ uk-koeln.de
Downloaded by: Univ.of Adelaide 129.127.145.240 - 8/24/2017 4:28:18 PM
Abstract Introduction: Published results of HistoScanningTM (HS) for prostate cancer (PCa) diagnostics are inconsistent and their value remains unclear. We prospectively analyzed the detection rate and tumor volume concordance in PCa patients. Material and Methods: Two hundred and eighty-two patients with biopsy-proven PCa scheduled for radical prostatectomy (RP) were included. All patients underwent ultrasonographical examination by HS prior to surgery. HS was evaluated compared to RP specimen as to (1) the prediction of overall tumor volume and (2) accuracy of HS in detection of PCa lesions larger than 0.2/0.5 ml, separated for each sextant. For each sextant, receiver operating characteristic (ROC)-analysis and area under the curve were determined. Sensitivity and specificity were calculated and visualized in ROC-curves. Results: HS tends to underestimate volume of cancerous lesions, particularly larger lesions >8 ml. Using a 0.2 ml detection threshold, specificity and sensitivity of HS were between 29–68% and 46–78%. For a 0.5 ml detection threshold, sextant-specific specificity increased to 59–92% and sensitivity decreased to 16–54%. Stratification accord-
The study was conducted in accordance with the Declaration of Helsinki. Institutional review board approval was obtained and all subjects provided written informed consent. Two hundred and eighty-two patients with histologically confirmed, organ-confined PCa, scheduled for RP were included. Initial histology was received by a 12-core TRUS-guided biopsy without information of a prebiopsy MRI. Therefore, no targeted biopsies were performed. All patients underwent TRUS including HS at the time of admission prior to surgery in a lateral decubitus position. No special rectal preparation was administered. In case of stool disturbing TRUS, digital removal was performed. The time interval between admission and surgery was 24 h in general. Patients who previously had a transurethral resection of the prostate and/or had a clinical tumor stage T3b or higher were not eligible. In case of hyperechogenic structures with dorsal echo extinction (i.e. prostatolithiasis) larger than 0.5 ml, patients were excluded from this study due to potentially missed ventral tumor foci. Further exclusion criteria were prostate volume >60 ml, pre-treatment of PCa and previous rectal surgery. The Reference Test – RP Histology Prostatectomy specimens were processed according to the Stanford protocol [9]. Briefly, after fixation and weighing, surgical margins were marked. Three differently colored inks were used for the ventral, dorsal and base area. First, seminal vesicles and vas deferens were cut off, cross-sectioned and embedded, separated by the side. Next, base margin – at the proximal ending of the urethra – and apex margin – at the distal ending of the urethra – were first cut off tangentially and then cross-sectioned and embedded, also separated by the side. Finally, whole prostate specimens were serially sectioned in a perpendicular plane to the rectal surface in 3 mm thick slices, beginning at the apex. Slices were laid out on the cutting table and embedded after separation of left and right side. After paraffination, sectioning and staining, histological examination was carried out. The Index Test Ultrasound data were acquired using the Pro Focus UltraviewTM device from B&K Medical and the Transducer 8818 (Triplane/140°) combined with a Magnetic Wheel Mover (UA 0513). HS was performed in 3 steps. First, a motorized TRUS recorded the complete prostate. Second, the examiner defined the region of interest within the HS system software. In a non-real-time fashion, the computerized HS analyses provided color-coded areas suspicious for PCa, and calculated the corresponding tumor volume. For prostate volumes up to 50 ml, 2 scans were performed; for volumes larger than 50 ml 3 scans were performed. All scans were analyzed to detect the presence of cancerous lesions. For volume measurements, the qualitatively best scan was chosen for each, base, middle and apex. The distance between prostate and probe was 4 mm. HS analysis occurred offline with the 2.2 software version after acquisition of the ultrasound data and without prior knowledge of the biopsy results. During analysis, the prostate gland was outlined and the 3-dimensional latitudes were measured. Foci were identified per sextant (base right and left, mid right and left, apex right and left) and total volumes of foci per sextant were quantified using the incorporated software program.
Sextant-Specific Analysis by HSTM
Statistical Analysis Statistical analyses were performed using SPSS Statistics® version 20 (IBM, Armonk, USA). A p value 20 ng/ml (n = 15), AUCs show greater variation between 0.39 and 0.88 (data not shown). Table 2 summarizes the specificity and sensitivity of HS for patients grouped by pre-operational PSA values, using either 0.2 or 0.5 ml as the detection limit. For base and middle sextants, sensitivity tends to be highest in patients with PSA values up to 5 ng/ml compared to those with higher PSA values. Specificity of HS in the apex is higher than in the base or the middle sextants, for both left and right and for all three ranges of PSA values. Specificity in the left-middle sextant is particularly low being just over 20% for PSA values
Urologia
Received: June 15, 2015 Accepted after revision: August 31, 2015 Published online: November 11, 2015
Urol Int 2016;96:194–201 DOI: 10.1159/000440814
Internationalis
Sextant-Specific Analysis of Detection and Tumor Volume by HistoScanningTM Daniel Porres a Timur Hasan Kuru a Robin Epplen a Andreas Eck e Vahudin Zugor e Lieven Niels Kennes c Samir Afram d Till Braunschweig b Ruth Knüchel-Clarke b David Pfister a Axel Heidenreich a
Departments of a Urology, b Pathology, and c Medical Statistics, RWTH University, Aachen, Departments of d Pathology and e Urology, St. Antonius-Hospital, Gronau, Germany
Key Words Prostate cancer · Diagnostic · Ultrasound · HistoScanning
ing to pre-operational PSA values did not improve performance characteristics of HS. Conclusions: Our results do not support a significant contribution of HS to PCa diagnostics. © 2015 S. Karger AG, Basel
© 2015 S. Karger AG, Basel 0042–1138/15/0962–0194$39.50/0 E-Mail [email protected] www.karger.com/uin
Introduction
In the search of optimized diagnostics for prostate cancer (PCa), HistoScanning (HS) has been introduced to support the detection and assessment of cancer foci. In 2008, Braeckman et al. [1] reported their first promising results with this technique. However, later study by Schiffmann et al. [2] was not able to reproduce these observations. He was not able to show a significant correlation between tumor volume by HS and radical prostatectomy (RP) specimen in a cohort of 148 patients. In the mean time, other diagnostic tools like elastography, multiparametric magnetic-resonance imaging (mp-MRI) and MRI/transrectal ultrasound (TRUS) fusion biopsies showed promising results in PCa detection [3–8]. This study aims at validating the previous results of HS on the gold standard of RP specimen and put them into the context of other available diagnostic modalities for PCa detection. Daniel Porres, MD Department of Urology, RWTH University Pauwelsstr. 30, 52074 Aachen (Germany) E-Mail daniel.porres @ uk-koeln.de
Downloaded by: Univ.of Adelaide 129.127.145.240 - 8/24/2017 4:28:18 PM
Abstract Introduction: Published results of HistoScanningTM (HS) for prostate cancer (PCa) diagnostics are inconsistent and their value remains unclear. We prospectively analyzed the detection rate and tumor volume concordance in PCa patients. Material and Methods: Two hundred and eighty-two patients with biopsy-proven PCa scheduled for radical prostatectomy (RP) were included. All patients underwent ultrasonographical examination by HS prior to surgery. HS was evaluated compared to RP specimen as to (1) the prediction of overall tumor volume and (2) accuracy of HS in detection of PCa lesions larger than 0.2/0.5 ml, separated for each sextant. For each sextant, receiver operating characteristic (ROC)-analysis and area under the curve were determined. Sensitivity and specificity were calculated and visualized in ROC-curves. Results: HS tends to underestimate volume of cancerous lesions, particularly larger lesions >8 ml. Using a 0.2 ml detection threshold, specificity and sensitivity of HS were between 29–68% and 46–78%. For a 0.5 ml detection threshold, sextant-specific specificity increased to 59–92% and sensitivity decreased to 16–54%. Stratification accord-
The study was conducted in accordance with the Declaration of Helsinki. Institutional review board approval was obtained and all subjects provided written informed consent. Two hundred and eighty-two patients with histologically confirmed, organ-confined PCa, scheduled for RP were included. Initial histology was received by a 12-core TRUS-guided biopsy without information of a prebiopsy MRI. Therefore, no targeted biopsies were performed. All patients underwent TRUS including HS at the time of admission prior to surgery in a lateral decubitus position. No special rectal preparation was administered. In case of stool disturbing TRUS, digital removal was performed. The time interval between admission and surgery was 24 h in general. Patients who previously had a transurethral resection of the prostate and/or had a clinical tumor stage T3b or higher were not eligible. In case of hyperechogenic structures with dorsal echo extinction (i.e. prostatolithiasis) larger than 0.5 ml, patients were excluded from this study due to potentially missed ventral tumor foci. Further exclusion criteria were prostate volume >60 ml, pre-treatment of PCa and previous rectal surgery. The Reference Test – RP Histology Prostatectomy specimens were processed according to the Stanford protocol [9]. Briefly, after fixation and weighing, surgical margins were marked. Three differently colored inks were used for the ventral, dorsal and base area. First, seminal vesicles and vas deferens were cut off, cross-sectioned and embedded, separated by the side. Next, base margin – at the proximal ending of the urethra – and apex margin – at the distal ending of the urethra – were first cut off tangentially and then cross-sectioned and embedded, also separated by the side. Finally, whole prostate specimens were serially sectioned in a perpendicular plane to the rectal surface in 3 mm thick slices, beginning at the apex. Slices were laid out on the cutting table and embedded after separation of left and right side. After paraffination, sectioning and staining, histological examination was carried out. The Index Test Ultrasound data were acquired using the Pro Focus UltraviewTM device from B&K Medical and the Transducer 8818 (Triplane/140°) combined with a Magnetic Wheel Mover (UA 0513). HS was performed in 3 steps. First, a motorized TRUS recorded the complete prostate. Second, the examiner defined the region of interest within the HS system software. In a non-real-time fashion, the computerized HS analyses provided color-coded areas suspicious for PCa, and calculated the corresponding tumor volume. For prostate volumes up to 50 ml, 2 scans were performed; for volumes larger than 50 ml 3 scans were performed. All scans were analyzed to detect the presence of cancerous lesions. For volume measurements, the qualitatively best scan was chosen for each, base, middle and apex. The distance between prostate and probe was 4 mm. HS analysis occurred offline with the 2.2 software version after acquisition of the ultrasound data and without prior knowledge of the biopsy results. During analysis, the prostate gland was outlined and the 3-dimensional latitudes were measured. Foci were identified per sextant (base right and left, mid right and left, apex right and left) and total volumes of foci per sextant were quantified using the incorporated software program.
Sextant-Specific Analysis by HSTM
Statistical Analysis Statistical analyses were performed using SPSS Statistics® version 20 (IBM, Armonk, USA). A p value 20 ng/ml (n = 15), AUCs show greater variation between 0.39 and 0.88 (data not shown). Table 2 summarizes the specificity and sensitivity of HS for patients grouped by pre-operational PSA values, using either 0.2 or 0.5 ml as the detection limit. For base and middle sextants, sensitivity tends to be highest in patients with PSA values up to 5 ng/ml compared to those with higher PSA values. Specificity of HS in the apex is higher than in the base or the middle sextants, for both left and right and for all three ranges of PSA values. Specificity in the left-middle sextant is particularly low being just over 20% for PSA values
