- Email: [email protected]
OPTICAL BIOPSY OF HUMAN BLADDER NEOPLASIA WITH IN VIVO CONFOCAL LASER ENDOMICROSCOPY
414
THE JOURNAL OF UROLOGY®
Vol. 181, No. 4, Supplement, Monday, April 27, 2009
Bladder Cancer: Detection and Screening Moderated Poster 38 Monday, April 27, 2009
3:30 pm - 5:30 pm
1160 DETECTION OF BLADDER CANCER WITH NARROW-BAND IMAGING SYSTEM Katsunori Tatsugami*, Fukuoka, Japan; Toshiyuki Kamoto, Hiroyuki Nishiyama, Atsushi Nishiyama, Kyoto, Japan; Satoru Ishikawa, Hitachi, Japan; Nobuo Shinohara, Ataru Sazawa, Hokkaido, Japan; Shoji Fukushima, Iwaki, Japan; Seiji Naito, Fukuoka, Japan INTRODUCTION AND OBJECTIVES: Recent reports have shown the advantage of fluorescence endoscopy using some dyes such as 5-aminolevulinic acid (ALA) or hypericin in the detection of bladder tumors. The fluorescence endoscope demonstrated the sensitivity above 90% and specificity ranging from 70% to 90%. The narrow-band imaging (NBI) technique, in which modified optical filters were used in the light source of a video endoscope system, narrows the bandwidth of spectral transmittance and enhances the mucosal surface contrast without the use of dyes. The clinical efficacy of the NBI system was evaluated for detecting imperceptible and/or flat bladder carcinomas, which are difficult to detect with the standard endoscope. METHODS: From July 2007 to September 2008, a prospective controlled study of NBI endoscope in comparison to a standard endoscope was conducted on the patients suspected of having bladder cancer at four hospitals. Patients were enrolled after giving their informed consent, and this study was approved by the institutional review board of each hospital. Transurethral targeted biopsies were performed after the examinations with both endoscopes for detecting the bladder tumor and compared the histological outcome. RESULTS: The biopsies were taken from 147 places, including 67 sites which were detected with both endoscopes. Fifty-eight (71%) of 82 patients were diagnosed to have bladder tumors by histological examinations and 47 (83%) of 58 patients had abnormal lesions which were not identified using the standard endoscope but were observed with the NBI endoscope. The percentage of patients who had malignant lesions detected only with the NBI endoscope was 34.1% (28/82 patients). The percentage of malignancy in the lesions detected only with NBI endoscope was 56.9% (33/58 places). The sensitivity, specificity, positive predictive value and likelihood ratio of negative test with the NBI endoscope for detecting bladder tumors were 94.3%, 60.2%, 65.6% and <0.1, respectively. CONCLUSIONS: Although the specificity of NBI endoscope is slightly lower than that of fluorescence endoscope, extremely subtle lesions can be identified with the NBI endoscope. Since the sensitivity is more than 90% and the likelihood ratio of negative test is less than 0.1, it is effective for identifying imperceptible and/or flat abnormal lesions without omission and is useful for exclusion diagnosis of bladder tumor. In particular, the NBI system may be effective for making a primary diagnosis in outpatients, because it is not necessary to use dyes such as ALA or hypericin for accurate detection. Source of Funding: None
1161 OPTICAL COHERENCE TOMOGRAPHY (OCT) - A MINIMALLY INVASIVE TECHNIQUE FOR THE DETECTION OF LESIONS WITHIN THE BLADDER Alexander Karl*, Herbert Stepp, Eva Willmann, Stefan Tritschler, Derya Tilki, Michael D Staehler, Christian G Stief, Munich, Germany INTRODUCTION AND OBJECTIVES: The feasibility of Optical Coherence Tomography (OCT) as a minimally-invasive technique was shown in different fields. Few recent studies also used OCT to detect and characterize lesions within the urinary bladder in vivo. Monocenter
results and experiences using this new technique are presented. METHODS: In 52 patients 166 lesions were evaluated using OCT as additional diagnostic method during TUR-BT/bladder biopsy. Involved in this study were thirty-two patients with a positive history for urothelial cancer and twenty patients with suspicion for malignancy (positive cytology, gross hematuria, suspicious cystoscopical finding etc.). In our study we used an OCT-device (Niris®, Imalux®, Cleveland, US), that utilizes infrared light guided through a flexible fibre-based applicator. OCT -scanning was performed using a 1300-nm 10-mW superluminescent diode with an applicator of 2.7mm outer diameter positioned through a regular cystoscope sheath (Storz, Tuttlingen, Germany, 25 Chr). OCT scans of about 1.5 seconds duration each generated 200 x 200-pixel images. OCT was used additionally to white light cystoscopy in the interpretation of lesions detected within the bladder. Suspicious lesions were scanned using OCT and either biopsy or TUR-BT was performed. All procedures were videotaped for subsequent validation of correct matching between biopsy and OCT site. All OCT images were evaluated by the surgeon during the procedure. RESULTS: Of 166 recorded OCT images, 102 lesions (61.4%) could be verified by subsequent videoanalysis to be matching exactly with the site of biopsy taken. Only these verified lesions were used for further analysis. Of these lesions 88 were benign (inflammation, edema, hyperplasia etc.) and 14 showed malignancy (CIS, Ta, T1, T2) in histopathology. All malignant lesions were detected correctly using this approach. OCT showed a sensitivity of 100% and a specificity of 65%. CONCLUSIONS: OCT as a minimally invasive detection method for bladder cancer showed a high sensitivity for malignant lesions in this dataset. As the used OCT probe has only a small diameter and no further marking tool was used, only 61% of taken biopsies could be matched with the OCT site as verified by videoanalysis. Therefore a special marking tool or the development of an adapted sheath that allows switching easily between OCT and biopsy procedure would be supportive. Source of Funding: None
1162 OPTICAL BIOPSY OF HUMAN BLADDER NEOPLASIA WITH IN VIVO CONFOCAL LASER ENDOMICROSCOPY Geoffrey A Sonn*, Sha-Nita Jones, Kathleen E Mach, Christine B Du, Kristin Jensen, Joseph C Liao, Stanford, CA INTRODUCTION AND OBJECTIVES: Great interest exists in enhancing the diagnostic accuracy of white light cystoscopy. Confocal laser endomicroscopy is a new tool that provides real-time imaging of live tissues with cellular detail. We describe its first in vivo application in the urinary tract to evaluate normal and neoplastic urothelium. METHODS:After IRB approval, 18 consecutive patients scheduled for TURBT were recruited. Following diagnostic cystoscopy, fluorescein sodium was administered as a contrast dye either intravenously (n=6), intravesically (n=3), or both (n=9). A 2.6 mm confocal probe (Mauna Kea Technologies, Paris, France) was passed through the working channel of a 26 Fr resectoscope. Images were acquired with 488-nm excitation at 12 frames/s. Normal and abnormal appearing areas were imaged through direct probe contact. The confocal images were visualized in real-time along with cystoscopic images and later further analyzed. After confocal imaging, the TURBT was completed. Confocal findings were compared with standard H&E analysis. RESULTS: With either intravenous or intravesical fluorescein, confocal laser endomicroscopy of normal appearing mucosa enabled realtime differentiation of the urothelial layers (umbrella cells, intermediate cells, and lamina propria) by varying probe contact pressure. On final pathology, 4 patients had benign lesions, 7 had low-grade, and 7 had high-grade urothelial carcinoma. Confocal endomicroscopy revealed morphologic differences between normal, low-grade, and high-grade tumors. Low-grade papillary tumors demonstrated densely arranged but uniformly shaped small cells surrounding fibrovascular cores, while high-grade tumors showed markedly irregular architecture and pleomorphic cells. With a 60μm penetration depth, muscularis propria was not visualized with the current probe. CONCLUSIONS: We report the first in vivo application of confocal laser endomicroscopy in the urinary tract. Real-time cellular
THE JOURNAL OF UROLOGY®
Vol. 181, No. 4, Supplement, Monday, April 27, 2009
imaging revealed histologic distinctions between normal and neoplastic urothelium comparable to standard H&E analysis. Pending further technologic development and clinical validation, confocal endomicroscopy may emerge as a valuable adjunct to cystoscopy for minimally invasive diagnosis of bladder neoplasms.
Source of Funding: Stanford Cancer Center Developmental Cancer Research Award
1163 NARROW BAND IMAGING FLEXIBLE CYSTOSCOPY: AN UPDATE AND A NEW USER’S EXPERIENCE Richard T Bryan*, Birmingham, United Kingdom; Z H Shah, Reading, United Kingdom; Stuart I Collins, D. Michael A Wallace, Birmingham, United Kingdom INTRODUCTION AND OBJECTIVES: We have previously demonstrated a significantly improved detection rate for bladder cancer recurrences with narrow band imaging (NBI) flexible cystoscopy when compared with conventional white-light imgaing (WLI) flexible cystoscopy. We investigated whether a “new user” of NBI flexible cystoscopy, previously unfamiliar with this technique, could reproduce our results. METHODS: The same protocol from our initial study was continued into this second study at The Queen Elizabeth Hospital, Birmingham, UK, but with a “new user” (ZHS): between September 2007 and May 2008 NBI flexible cystoscopy was performed on 23 patients with known recurrences of urothelial cancer (UC) of the bladder following initial conventional WLI flexible cystoscopy with the same switchable Olympus Lucera sequential RGB instrument. RESULTS: In our first series (published), NBI detected a total of 15 additional UCs in 12 of 29 patients (41%), as compared with WLI: a mean difference of 0.52 UCs per patient (standard deviation 0.74; Wilcoxon paired signed-rank test p-value 0.0005). In this second series, NBI detected 15 additional UCs in 8 of the 23 patients (35%): six of these patients had one additional UC, one had four additional UCs, and one had five additional UCs when compared with WLI, with a mean of 0.65 additional UCs per patient (SD1.30; Wilcoxon p=0.01). When the second series with the “new user” is compared with the first series, there is no statistical evidence that the excess number of UCs detected by NBI is different (Wilcoxon (unpaired) signed-rank test p=0.74), suggesting that there is no difference between a new user and an experienced user in the application of NBI. When the two series are combined, NBI detected 30 additional UCs in 20 of 52 patients (38%): 16 patients had one additional UC, and one each had two, three, four and five additional UCs, with a mean of 0.58 additional UCs per patient overall (SD 1.02; Wilcoxon paired signed-rank test p<0.001). CONCLUSIONS: A new user can obtain the same significant benefits with NBI as an experienced user in improving the detection rate of bladder UC recurrences. We continue to demonstrate a significant benefit of NBI flexible cystoscopy in the follow-up of patients with bladder cancer. Source of Funding: KeyMed Olympus (UK) and Olympus (Japan) provided the authors with the flexible cystoscopy equipment required to carryout this study.
415
1164 MAIN PREDICTORS OF FALSE POSITIVES IN PHOTODYNAMIC DIAGNOSIS OF TRANSITIONAL CELL CARCINOMA OF THE BLADDER Ronald O.P. Draga*, Matthijs C.M. Grimbergen, Esther T. Kok, Trudy Jonges, Ruud Bosch, Utrecht, Netherlands INTRODUCTION AND OBJECTIVES: Photodynamic diagnosis (PDD) is a technique that enhances the detection of tumors during cystoscopy using a photosensitizer which accumulates primarily in cancerous cells and will fluoresce when illuminated by violet-blue light. The specificity is approximately 50%. The aim of the study is to identify the main predictors of false positives in multivariate analysis. The second objective is to determine the optimal waiting period after transurethral resection of bladder tumors (TURBT) and intravesical therapy (IVT) to minimize the number of false positives in PDD. METHODS: Data of 366 procedures and 200 patients were collected. Patients were instilled with 5-aminolevulinic acid (5-ALA) intravesically and 1253 biopsies were taken from tumors and suspicious lesions. Age, gender, recent TURBT, previous IVT, urinary tract infections (UTIs) and tumor classification were examined for association with the false-positive rates in fluorescence cystoscopy. RESULTS: The sensitivity and specificity of white light endoscopy (WLE) and PDD are 71% and 76%, 96% and 43%, respectively. Significant univariate associations are found between false positives and female gender (p=0.007, OR=2.01), IVT instillations (p=0.02, OR=1.85), Bacille Calmette-Guérin instillations (BCG; p=0.03, OR=2.16) and TURBT in the past 90 days (p=0.01, OR=2.23). In multivariate analysis female gender (p=0.001, OR=2.68) and TURBT within 90 days before PDD (p=0.01, OR=2.27) are the only significant independent predictors of false positives. In a second multivariate model the dichotomous variable ‘>1 BCG instillation in the past 90 days’ is the only predictor for falsepositive findings in PDD (p=0.002, OR=4.67). Tangential illumination of the bladder wall does not seem to result in additional false positives. The false-positive rate decreases during the first 12 weeks after the latest TURBT and the latest BCG instillation. CONCLUSIONS: Female gender, previous TURBT and recent BCG instillations are important predictors of false positives in PDD. The false-positive rate decreases during the first 12 weeks after the latest TURBT and the latest BCG instillation. We recommend to perform a fluorescence guided TURBT 9-12 weeks after an incomplete first resection of low or intermediate risk non-muscle invasive tumors and 9-12 weeks after the latest BCG instillation. Source of Funding: The Dutch Cancer Society, grant-number UU 2007-3922
1165 URINARY - BASED TUMOUR MARKERS EVALUATED EITHER BY WHITE LIGHT OR PHOTODYNAMIC TRANSURETHRAL RESECTION IN BLADDER CANCER. IS THERE ANY DIFFERENCE? Marcus Horstmann*, Tuebingen, Germany; Oliver Patschan, Malmoe, Sweden; Joerg Hennenlotter, Daniela Colleselli, Gerhard Feil, Arnulf Stenzl, Tuebingen, Germany INTRODUCTION AND OBJECTIVES: Photodynamic diagnostic (PDD) improves the detection of transitional cell carcinoma of the bladder (TCC). Still sensitivity and specificity of urinary-based tumour markers and cytology have been validated only by white light cystoscopy or resection (WL). Aim of this study was to evaluate how the sensitivity and specificity of urinary-based tumor markers changes if either WL TUR or photodynamic TUR were used. METHODS: Cytology, FISH, U-Cyt+, and NMP22 were evaluated prior to transurethral resection in 335 serial patients with the suspicion of primary TCC (n=163) or TCC recurrence (n=172) of the bladder. In each case histopathological results served as a control. PDD was used in 112 patients during resection. BCa was found in 128 (57%) out of 223 (100%) patients in WL and in 70 (62%) out of 112 (100) patients in PDD.
THE JOURNAL OF UROLOGY®
Vol. 181, No. 4, Supplement, Monday, April 27, 2009
Bladder Cancer: Detection and Screening Moderated Poster 38 Monday, April 27, 2009
3:30 pm - 5:30 pm
1160 DETECTION OF BLADDER CANCER WITH NARROW-BAND IMAGING SYSTEM Katsunori Tatsugami*, Fukuoka, Japan; Toshiyuki Kamoto, Hiroyuki Nishiyama, Atsushi Nishiyama, Kyoto, Japan; Satoru Ishikawa, Hitachi, Japan; Nobuo Shinohara, Ataru Sazawa, Hokkaido, Japan; Shoji Fukushima, Iwaki, Japan; Seiji Naito, Fukuoka, Japan INTRODUCTION AND OBJECTIVES: Recent reports have shown the advantage of fluorescence endoscopy using some dyes such as 5-aminolevulinic acid (ALA) or hypericin in the detection of bladder tumors. The fluorescence endoscope demonstrated the sensitivity above 90% and specificity ranging from 70% to 90%. The narrow-band imaging (NBI) technique, in which modified optical filters were used in the light source of a video endoscope system, narrows the bandwidth of spectral transmittance and enhances the mucosal surface contrast without the use of dyes. The clinical efficacy of the NBI system was evaluated for detecting imperceptible and/or flat bladder carcinomas, which are difficult to detect with the standard endoscope. METHODS: From July 2007 to September 2008, a prospective controlled study of NBI endoscope in comparison to a standard endoscope was conducted on the patients suspected of having bladder cancer at four hospitals. Patients were enrolled after giving their informed consent, and this study was approved by the institutional review board of each hospital. Transurethral targeted biopsies were performed after the examinations with both endoscopes for detecting the bladder tumor and compared the histological outcome. RESULTS: The biopsies were taken from 147 places, including 67 sites which were detected with both endoscopes. Fifty-eight (71%) of 82 patients were diagnosed to have bladder tumors by histological examinations and 47 (83%) of 58 patients had abnormal lesions which were not identified using the standard endoscope but were observed with the NBI endoscope. The percentage of patients who had malignant lesions detected only with the NBI endoscope was 34.1% (28/82 patients). The percentage of malignancy in the lesions detected only with NBI endoscope was 56.9% (33/58 places). The sensitivity, specificity, positive predictive value and likelihood ratio of negative test with the NBI endoscope for detecting bladder tumors were 94.3%, 60.2%, 65.6% and <0.1, respectively. CONCLUSIONS: Although the specificity of NBI endoscope is slightly lower than that of fluorescence endoscope, extremely subtle lesions can be identified with the NBI endoscope. Since the sensitivity is more than 90% and the likelihood ratio of negative test is less than 0.1, it is effective for identifying imperceptible and/or flat abnormal lesions without omission and is useful for exclusion diagnosis of bladder tumor. In particular, the NBI system may be effective for making a primary diagnosis in outpatients, because it is not necessary to use dyes such as ALA or hypericin for accurate detection. Source of Funding: None
1161 OPTICAL COHERENCE TOMOGRAPHY (OCT) - A MINIMALLY INVASIVE TECHNIQUE FOR THE DETECTION OF LESIONS WITHIN THE BLADDER Alexander Karl*, Herbert Stepp, Eva Willmann, Stefan Tritschler, Derya Tilki, Michael D Staehler, Christian G Stief, Munich, Germany INTRODUCTION AND OBJECTIVES: The feasibility of Optical Coherence Tomography (OCT) as a minimally-invasive technique was shown in different fields. Few recent studies also used OCT to detect and characterize lesions within the urinary bladder in vivo. Monocenter
results and experiences using this new technique are presented. METHODS: In 52 patients 166 lesions were evaluated using OCT as additional diagnostic method during TUR-BT/bladder biopsy. Involved in this study were thirty-two patients with a positive history for urothelial cancer and twenty patients with suspicion for malignancy (positive cytology, gross hematuria, suspicious cystoscopical finding etc.). In our study we used an OCT-device (Niris®, Imalux®, Cleveland, US), that utilizes infrared light guided through a flexible fibre-based applicator. OCT -scanning was performed using a 1300-nm 10-mW superluminescent diode with an applicator of 2.7mm outer diameter positioned through a regular cystoscope sheath (Storz, Tuttlingen, Germany, 25 Chr). OCT scans of about 1.5 seconds duration each generated 200 x 200-pixel images. OCT was used additionally to white light cystoscopy in the interpretation of lesions detected within the bladder. Suspicious lesions were scanned using OCT and either biopsy or TUR-BT was performed. All procedures were videotaped for subsequent validation of correct matching between biopsy and OCT site. All OCT images were evaluated by the surgeon during the procedure. RESULTS: Of 166 recorded OCT images, 102 lesions (61.4%) could be verified by subsequent videoanalysis to be matching exactly with the site of biopsy taken. Only these verified lesions were used for further analysis. Of these lesions 88 were benign (inflammation, edema, hyperplasia etc.) and 14 showed malignancy (CIS, Ta, T1, T2) in histopathology. All malignant lesions were detected correctly using this approach. OCT showed a sensitivity of 100% and a specificity of 65%. CONCLUSIONS: OCT as a minimally invasive detection method for bladder cancer showed a high sensitivity for malignant lesions in this dataset. As the used OCT probe has only a small diameter and no further marking tool was used, only 61% of taken biopsies could be matched with the OCT site as verified by videoanalysis. Therefore a special marking tool or the development of an adapted sheath that allows switching easily between OCT and biopsy procedure would be supportive. Source of Funding: None
1162 OPTICAL BIOPSY OF HUMAN BLADDER NEOPLASIA WITH IN VIVO CONFOCAL LASER ENDOMICROSCOPY Geoffrey A Sonn*, Sha-Nita Jones, Kathleen E Mach, Christine B Du, Kristin Jensen, Joseph C Liao, Stanford, CA INTRODUCTION AND OBJECTIVES: Great interest exists in enhancing the diagnostic accuracy of white light cystoscopy. Confocal laser endomicroscopy is a new tool that provides real-time imaging of live tissues with cellular detail. We describe its first in vivo application in the urinary tract to evaluate normal and neoplastic urothelium. METHODS:After IRB approval, 18 consecutive patients scheduled for TURBT were recruited. Following diagnostic cystoscopy, fluorescein sodium was administered as a contrast dye either intravenously (n=6), intravesically (n=3), or both (n=9). A 2.6 mm confocal probe (Mauna Kea Technologies, Paris, France) was passed through the working channel of a 26 Fr resectoscope. Images were acquired with 488-nm excitation at 12 frames/s. Normal and abnormal appearing areas were imaged through direct probe contact. The confocal images were visualized in real-time along with cystoscopic images and later further analyzed. After confocal imaging, the TURBT was completed. Confocal findings were compared with standard H&E analysis. RESULTS: With either intravenous or intravesical fluorescein, confocal laser endomicroscopy of normal appearing mucosa enabled realtime differentiation of the urothelial layers (umbrella cells, intermediate cells, and lamina propria) by varying probe contact pressure. On final pathology, 4 patients had benign lesions, 7 had low-grade, and 7 had high-grade urothelial carcinoma. Confocal endomicroscopy revealed morphologic differences between normal, low-grade, and high-grade tumors. Low-grade papillary tumors demonstrated densely arranged but uniformly shaped small cells surrounding fibrovascular cores, while high-grade tumors showed markedly irregular architecture and pleomorphic cells. With a 60μm penetration depth, muscularis propria was not visualized with the current probe. CONCLUSIONS: We report the first in vivo application of confocal laser endomicroscopy in the urinary tract. Real-time cellular
THE JOURNAL OF UROLOGY®
Vol. 181, No. 4, Supplement, Monday, April 27, 2009
imaging revealed histologic distinctions between normal and neoplastic urothelium comparable to standard H&E analysis. Pending further technologic development and clinical validation, confocal endomicroscopy may emerge as a valuable adjunct to cystoscopy for minimally invasive diagnosis of bladder neoplasms.
Source of Funding: Stanford Cancer Center Developmental Cancer Research Award
1163 NARROW BAND IMAGING FLEXIBLE CYSTOSCOPY: AN UPDATE AND A NEW USER’S EXPERIENCE Richard T Bryan*, Birmingham, United Kingdom; Z H Shah, Reading, United Kingdom; Stuart I Collins, D. Michael A Wallace, Birmingham, United Kingdom INTRODUCTION AND OBJECTIVES: We have previously demonstrated a significantly improved detection rate for bladder cancer recurrences with narrow band imaging (NBI) flexible cystoscopy when compared with conventional white-light imgaing (WLI) flexible cystoscopy. We investigated whether a “new user” of NBI flexible cystoscopy, previously unfamiliar with this technique, could reproduce our results. METHODS: The same protocol from our initial study was continued into this second study at The Queen Elizabeth Hospital, Birmingham, UK, but with a “new user” (ZHS): between September 2007 and May 2008 NBI flexible cystoscopy was performed on 23 patients with known recurrences of urothelial cancer (UC) of the bladder following initial conventional WLI flexible cystoscopy with the same switchable Olympus Lucera sequential RGB instrument. RESULTS: In our first series (published), NBI detected a total of 15 additional UCs in 12 of 29 patients (41%), as compared with WLI: a mean difference of 0.52 UCs per patient (standard deviation 0.74; Wilcoxon paired signed-rank test p-value 0.0005). In this second series, NBI detected 15 additional UCs in 8 of the 23 patients (35%): six of these patients had one additional UC, one had four additional UCs, and one had five additional UCs when compared with WLI, with a mean of 0.65 additional UCs per patient (SD1.30; Wilcoxon p=0.01). When the second series with the “new user” is compared with the first series, there is no statistical evidence that the excess number of UCs detected by NBI is different (Wilcoxon (unpaired) signed-rank test p=0.74), suggesting that there is no difference between a new user and an experienced user in the application of NBI. When the two series are combined, NBI detected 30 additional UCs in 20 of 52 patients (38%): 16 patients had one additional UC, and one each had two, three, four and five additional UCs, with a mean of 0.58 additional UCs per patient overall (SD 1.02; Wilcoxon paired signed-rank test p<0.001). CONCLUSIONS: A new user can obtain the same significant benefits with NBI as an experienced user in improving the detection rate of bladder UC recurrences. We continue to demonstrate a significant benefit of NBI flexible cystoscopy in the follow-up of patients with bladder cancer. Source of Funding: KeyMed Olympus (UK) and Olympus (Japan) provided the authors with the flexible cystoscopy equipment required to carryout this study.
415
1164 MAIN PREDICTORS OF FALSE POSITIVES IN PHOTODYNAMIC DIAGNOSIS OF TRANSITIONAL CELL CARCINOMA OF THE BLADDER Ronald O.P. Draga*, Matthijs C.M. Grimbergen, Esther T. Kok, Trudy Jonges, Ruud Bosch, Utrecht, Netherlands INTRODUCTION AND OBJECTIVES: Photodynamic diagnosis (PDD) is a technique that enhances the detection of tumors during cystoscopy using a photosensitizer which accumulates primarily in cancerous cells and will fluoresce when illuminated by violet-blue light. The specificity is approximately 50%. The aim of the study is to identify the main predictors of false positives in multivariate analysis. The second objective is to determine the optimal waiting period after transurethral resection of bladder tumors (TURBT) and intravesical therapy (IVT) to minimize the number of false positives in PDD. METHODS: Data of 366 procedures and 200 patients were collected. Patients were instilled with 5-aminolevulinic acid (5-ALA) intravesically and 1253 biopsies were taken from tumors and suspicious lesions. Age, gender, recent TURBT, previous IVT, urinary tract infections (UTIs) and tumor classification were examined for association with the false-positive rates in fluorescence cystoscopy. RESULTS: The sensitivity and specificity of white light endoscopy (WLE) and PDD are 71% and 76%, 96% and 43%, respectively. Significant univariate associations are found between false positives and female gender (p=0.007, OR=2.01), IVT instillations (p=0.02, OR=1.85), Bacille Calmette-Guérin instillations (BCG; p=0.03, OR=2.16) and TURBT in the past 90 days (p=0.01, OR=2.23). In multivariate analysis female gender (p=0.001, OR=2.68) and TURBT within 90 days before PDD (p=0.01, OR=2.27) are the only significant independent predictors of false positives. In a second multivariate model the dichotomous variable ‘>1 BCG instillation in the past 90 days’ is the only predictor for falsepositive findings in PDD (p=0.002, OR=4.67). Tangential illumination of the bladder wall does not seem to result in additional false positives. The false-positive rate decreases during the first 12 weeks after the latest TURBT and the latest BCG instillation. CONCLUSIONS: Female gender, previous TURBT and recent BCG instillations are important predictors of false positives in PDD. The false-positive rate decreases during the first 12 weeks after the latest TURBT and the latest BCG instillation. We recommend to perform a fluorescence guided TURBT 9-12 weeks after an incomplete first resection of low or intermediate risk non-muscle invasive tumors and 9-12 weeks after the latest BCG instillation. Source of Funding: The Dutch Cancer Society, grant-number UU 2007-3922
1165 URINARY - BASED TUMOUR MARKERS EVALUATED EITHER BY WHITE LIGHT OR PHOTODYNAMIC TRANSURETHRAL RESECTION IN BLADDER CANCER. IS THERE ANY DIFFERENCE? Marcus Horstmann*, Tuebingen, Germany; Oliver Patschan, Malmoe, Sweden; Joerg Hennenlotter, Daniela Colleselli, Gerhard Feil, Arnulf Stenzl, Tuebingen, Germany INTRODUCTION AND OBJECTIVES: Photodynamic diagnostic (PDD) improves the detection of transitional cell carcinoma of the bladder (TCC). Still sensitivity and specificity of urinary-based tumour markers and cytology have been validated only by white light cystoscopy or resection (WL). Aim of this study was to evaluate how the sensitivity and specificity of urinary-based tumor markers changes if either WL TUR or photodynamic TUR were used. METHODS: Cytology, FISH, U-Cyt+, and NMP22 were evaluated prior to transurethral resection in 335 serial patients with the suspicion of primary TCC (n=163) or TCC recurrence (n=172) of the bladder. In each case histopathological results served as a control. PDD was used in 112 patients during resection. BCa was found in 128 (57%) out of 223 (100%) patients in WL and in 70 (62%) out of 112 (100) patients in PDD.