- Email: [email protected]
PERFORMANCE OF URINE TEST IN PATIENTS MONITORED FOR RECURRENCE OF BLADDER CANCER: A MULTICENTER STUDY IN THE UNITED STATES
0022-5347/05/1744-1238/0 THE JOURNAL OF UROLOGY® Copyright © 2005 by AMERICAN UROLOGICAL ASSOCIATION
Vol. 174, 1238 –1241, October 2005 Printed in U.S.A.
DOI: 10.1097/01.ju.0000173918.84006.4d
PERFORMANCE OF URINE TEST IN PATIENTS MONITORED FOR RECURRENCE OF BLADDER CANCER: A MULTICENTER STUDY IN THE UNITED STATES EDWARD M. MESSING,*, † LISA TEOT, HOWARD KORMAN,‡ JEANNE UNDERHILL, EDWARD BARKER, BRIAN STORK, JUNQI QIAN AND DAVID G. BOSTWICK From the Departments of Urology (EMM) and Pathology (LT), University of Rochester Medical Center, Rochester, New York, Departments of Urology (HK) and Pathology (JE), William Beaumont Hospital, Detroit, Michigan, Medical Lab Associates (EB, BS), Seattle, Washington, and Bostwick Laboratories, (JQ, DGB), Glen Allen, Virginia
ABSTRACT
Purpose: We assessed the performance of the ImmunoCyt® immunocytochemical test for detecting bladder cancer recurrence in patients with prior superficial bladder cancers compared with cystoscopic and histological findings. Materials and Methods: A total of 341 patients with a history of bladder cancer undergoing monitoring were evaluated at 4 sites. The results of cytology and/or ImmunoCyt® were analyzed for sensitivity and specificity compared with biopsy confirmed cancer. Results: The overall sensitivity of cytology alone, ImmunoCyt® alone and the 2 methods combined was 23%, 81% and 81%, respectively. The specificity of cytology alone, ImmunoCyt® alone and of the 2 methods combined was 93%, 75% and 73%, respectively. The immunocytochemical test was more sensitive than cytology for detecting grades 1 and 2, and stages Ta, T1, and T2 urothelial carcinoma, and it was equally sensitive for detecting grade 3 cancers and carcinoma in situ (CIS). The sensitivity of the combined tests for grades 1 to 3/CIS was 79%, 90% and 82%, while for stages Ta, T1, T2⫹ and CIS it was 83%, 75%, 100% and 100%, respectively. The overall positive and negative predictive values of the combined tests were 37% and 95%, respectively. Importantly the immunocytochemical test could detect 71% of small (less than 1 cm) tumors. Conclusions: ImmunoCyt® is a sensitive test for detecting bladder cancer. Because of its high sensitivity for detecting small tumors, even those of low histological grade, and its high negative predictive value, this test may have a role in decreasing the frequency of cystoscopic examinations for monitoring patients with low risk bladder cancer. KEY WORDS: bladder, bladder neoplasms, neoplasm recurrence, cytology
More than 55,000 cases of bladder cancer are diagnosed yearly in the United States.1 Within 2 years of endoscopic resection of papillary superficial tumors approximately 50% of patients experience recurrence and after 5 years the risk of muscle invasive cancer is 5% to 25%.2 Most urologists consider cystoscopy to be the best method for the diagnosis of bladder cancer and for followup after surgery and treatment. Many urologists recommend followup cystoscopy for noninvasive tumors every 3 months for 1 or 2 years and less frequently thereafter. In patients with high grade tumors who are at 73% risk for recurrence during year 1 after initial treatment and at about 25% risk for subsequent progression to muscle invasion3 frequent followup by cystoscopy may be advisable. However, in patients with lower grade tumors who are at lower risk for recurrence during year 1 after initial treatment3 it may be possible to use alternative, noninvasive monitoring methods to space out cystoscopic examinations.
The most common noninvasive method for monitoring bladder cancer is the detection of exfoliated tumor cells by urinary cytology. Cytology has high sensitivity for detecting high grade tumors but poor sensitivity for detecting low grade tumors, which are the most common type of bladder cancer.4 However, the limitations of cytology, especially for detecting recurrent low grade tumors, and the invasiveness of cystoscopy have increased interest in tests using urinary diagnostic markers. Examples are tests based on human complement related factor, nuclear mitotic apparatus protein and DNA loss of heterozygosity or chromosomal gain in tumor cells exfoliated in urine, as detected by fluorescence in situ hybridization.5 Tests using these markers are more sensitive than cytology but sensitivity is no greater than 60% for low grade tumors and specificity tends to be lower than that of cytology. The ImmunoCyt® test uses a cocktail of 3 monoclonal antibodies directed against urothelial carcinoma antigens in exfoliated cells, including 2 cytoplasmic mucin related antigens and high molecular weight carcinoembryonic antigen, to identify tumor cells shed in urine. The antibodies are labeled with fluorescent markers, which allow detection by fluorescence microscopy. In 5,370 cases evaluated in multiple studies done in Canada and Europe cumulative sensitivity was 85% and specificity was 71% (table 3).2, 6 –17 In addition, ImmunoCyt® also improved sensitivity for detecting low grade, low stage tumors from 50% to 90%.3 However, interval cystoscopy can detect small tumors, for which many marker
Submitted for publication January 25, 2005. Ethics approval for the study was obtained from individual institutions. Supported by DiagnoCure, Quebec, Canada. * Correspondence: Department of Urology, University of Rochester, 601 Elmwood Ave., Box 655, Rochester, New York 14642 (telephone: 585-275-3345; FAX: 585-442-8350; e-mail: edward_messing@ urmc.rochester.edu). † Financial interest and/or other relationship with AstraZeneca, Matritech, DynaCure and Fujirebio. ‡ Financial interest and/or other relationship with DiagnoCure and AstraZeneca. 1238
1239
URINE TEST AND RECURRENT BLADDER CANCER
tests have had difficulties.18 We assessed the performance of ImmunoCyt® and cytology to detect recurrent cancers in a population of patients with prior superficial bladder tumors. MATERIALS AND METHODS
Patients. This study was performed at 4 sites in the United States. From November 2000 to November 2003 voided urine specimens were obtained from 341 patients who were being monitored for bladder cancer recurrence. Except for a history of completely resected stage T1 or less urothelial cancer no specific tumor characteristics were prescribed. An attempt was made to enlist subjects with a history of grades 1 and 2, rather than high grade cancer. Cytological examination of urine samples was performed as well as evaluation using an ImmunoCyt® immunocytochemical test kit at each site. Ethics approval for the study was obtained from individual institutions. Informed consent, obtained from all patients, is on file at the sites. Bladder cancer was diagnosed based on histological confirmation from cystoscopic biopsy. Biopsy samples were prepared for histological staining and grading at each study site. Urine sample preparation. Each 20 to 40 ml voided urine specimen was obtained prior to cystoscopic examination and at least 4 weeks after any urethral instrumentation. Specimens were fixed with an equal volume of 50% ethanol or 50% isopropyl alcohol and then stored at 2C to 8C until testing. ImmunoCyt® fixative buffered solution (1 ml) was mixed with urine and the sample was incubated for 15 minutes. at room temperature (18C to 30C). The sample was then filtered through a polycarbonate membrane with a porosity of 8 m in a device connected to a 50 ml syringe. Cells captured on the filter were blotted onto a silane treated slide and spray fixed with 50% isopropanol before immunocytochemical assay was performed. Control slides were also prepared using positive and negative control cells provided with the kit. ImmunoCyt® assay. Slides were prepared for immunofluorescence staining, as described in the package insert. They were pretreated with 80%, 70% and 50% ethanol, distilled water and Harris hematoxylin with 4% acetic acid. Cells were then incubated with 4 drops of blocking solution for 15 minutes at room temperature in a closed humid chamber and with the ImmunoCyt® antibody cocktail for 1 hour. Slides were then rinsed in phosphate buffered saline containing 0.5% Tween 20 and in pure phosphate buffered saline. Mounting medium (1 drop) was added on the slide before the cover slip. Slides were stored at 4C until evaluation using an epifluorescence microscope fitted with a double filter to detect fluorescein and Texas red emissions. Red fluorescent cells were positive for a glycosylated form of carcinoembryonic antigen and green fluorescent cells were positive for bladder cancer mucins. The test was scored positive as soon as 1 confirmed green or red fluorescent cell was detected. According to manufacturer instructions slides with fewer than 500 cells were considered inadequate for to validate a negative result with ImmunoCyt®. More than 500 microscopic fields (20⫻ objective) per slide were screened. Techni-
cians and cytopathologists were blinded to the results of cystoscopy, biopsy and cytology. Cytology. Slides for conventional urinary cytology were stained according to the standard Papanicolaou procedure. Three cytology categories were defined as negative, suspicious and positive for malignancy, respectively. In this study cases with atypia were included in the negative category and only the presence of malignant cells was considered positive. Sensitivity and specificity. Sensitivity was calculated as the number of samples testing positive by immunocytochemical and/or cytology divided by the number of samples that were positive by biopsy. Specificity was calculated as the number of negative tests divided by the number of samples that were negative by cystoscopy or biopsy, if the patient had suspicious cystoscopy but negative biopsy. For grade, stage and tumor size sensitivity was calculated as the number of samples positive by each testing method divided by the number of samples positive for cancer for which a grade, stage or size could be accurately assessed. Positive and negative predictive values were calculated in the entire cohort tested for which a satisfactory immunocytochemical or cytology specimen was obtained. RESULTS
Of 341 patients who were evaluated by cystoscopy for bladder cancer recurrence 14 (4%) had urine specimens that were excluded because of inadequate sampling, defined as less than 500 cells per slide. Of the remaining 327 patients 61 (19%) had cystoscopically evident tumors, of which 9 were fulgurated, providing no tissue for histology. Of the 52 histologically confirmed cancers 50 were evaluable for tumor stage, 49 were evaluable for tumor grade and 41 were evaluable for tumor size determinations. Tables 1 to 4 show the performance of ImmunoCyt® and cytology overall and for various grades, stages and sizes. In the 42 cases of histologically confirmed bladder cancer in which ImmunoCyt® was positive 5 or fewer cells were positive in slides in 16 cases, 6 to 20 were positive in 13 and greater than 20 were positive in 13. In no instance with a specimen evaluable by immunocytochemistry was cytology positive when immunocytochemistry was not. Importantly the sensitivity for stage T2 cancers and carcinoma in situ (CIS) (7 samples) was 100% for ImmunoCyt® and 71% (5 of 7) for cytology. The sensitivity of ImmunoCyt® and cytology was each the same for grade 3 cancers (82% or 9 of 11). Of tumors smaller than 1 cm 71% were positive by ImmunoCyt®. Sensitivity for grade 1 and stage Ta tumors was 79% and 83%, respectively. DISCUSSION
Cystoscopy is the most specific and sensitive method for detecting new and recurrent bladder cancers. However, it is invasive, resulting in patient discomfort and the risk of infection. In addition, it is difficult to detect flat tumors, such as (CsIS).19 There is growing evidence that fluorescent dye
TABLE 1. ImmunoCyt® and cytology performance % (No./total No.) ImmunoCyt®
Cytology*
ImmunoCyt® ⫹ Cytology
Pos predictive value True pos/true pos ⫹ false-pos 38 (42/110) 37 (42/115) Neg predictive value True neg/true neg ⫹ false-neg 95 (206/216) 95 (206/216) Overall sensitivity True pos/true pos ⫹ false-neg 81 (42/52) 23 (12/52) 81 (42/52) Specificity True neg/true neg ⫹ false-pos 75 (206/274) 93 (255/274) 73 (201/327) Only histologically confirmed cancer for which a test was positive was considered positive and if additional cases for which fulguration was performed without biopsy were each considered to show cancer, positive predictive value would have changed to 51% and 48% for ImmunoCyt® and for Immunocyt® and cytology, respectively, and sensitivity would have changed to 51 of 61 cases (84%) for Immunocyt®, and for Immunocyt® and cytology. * Atypia and dysplasia were not considered positive.
1240
URINE TEST AND RECURRENT BLADDER CANCER TABLE 2. ImmunoCyt® and cytology sensitivity by available tumor grade or stage
Grade or Stage
No. Pts
G1 G2 G3* Ta T1 T2 CIS * G3 and CIS sensitivity 9 of 11
% ImmunoCyt® (No. pts)
Cytology (No. pts)
ImmunoCyt® ⫹ Cytology
79 (22) 90 (9) 67 (4) 83 (29) 75 (6) 100 (2) 100 (5)
7 (2) 10 (1) 67 (4) 6 (2) 50 (4) 0 100 (5)
79 (22) 90 (9) 67 (4) 83 (29) 75 (6) 100 (2) 100 (5)
28 10 6 35 8 2 5 patients (82%).
TABLE 3. Performance of combined ImmunoCyt® and cytology tests for bladder cancer % G1–G2 Sensitivity (No. pts)
References
89–97 (70) Fradet and Lockhard6 87–90 (98) Lodde et al14 Bunting et al15 100 (22) Olsson and Zackrisson7 Vriesema et al13 59–89 (93) Lopez et al11 81–96 (71) Lodde et al8 67–78 (71) Pfister et al2 84–92 (55) Mian et al9 14–64 (21) Ceil et al16 71–79 (64) Tetu et al10 86–91 (30) Toma et al12 79–91 (209) Mian et al17 Present series 79–90 (38) * Low sensitivity due to improper technical performance.
Less than 1 1–3 Greater than 3
% Specificity (No. pts)
100 (25) 95 (45)
76 (102) 74 (286) 52 (68) 69 (83) 64 (64) 76 (287) 68 (123) 81 (548) 66 (92) 82 (87) 61 (734) 72 (82) 73 (1,585) 73 (274)
45 (23)* 100 (8) 52 (22)* 93 (54) 100 (31) 87 (68) 96 (25) 80 (5) 93 (40) 83 (12) 99 (89) 82 (11)
TABLE 4. ImmunoCyt® and cytology sensitivity on available tumor sizes Size (cm)
% G3/CIS Sensitivity (No. pts)
% (No./total No.) ImmunoCyt®
Cytology
71 (12/17) 84 (16/19) 60 (3/5)
18 (3/17) 26 (5/19) 20 (1/5)
increases the sensitivity of cystoscopy. In 1 study standard cystoscopy did not detect a substantial number of bladder tumors that were visible using fluorescence endoscopy.3, 20 In contrast, cytological examination of exfoliated cells shed in urine is a specific but not a sensitive method. New methods for detecting bladder cancer that are based on urinary markers have become available in the last 10 years. Diagnostic tests for recurrent bladder cancer that are available in the United States, Canada and Europe have a sensitivity range of 46% to 100% and a specificity range of 60% to 100%.16 Four tests are available in the United States. The BTA Stat® and BTA Trak® tests detect human complement factor H related protein in voided urine.21 Sensitivity is 57% to 83% depending on grade and stage, and specificity is 46% to 73%. The nuclear mitotic apparatus protein assay, which detects a urinary nuclear mitotic apparatus protein, is reported to have 47% to 100% sensitivity and 60% to 70% specificity. The UroVysion™ fluorescence in situ hybridization assay, which uses probes to the centromeres of chromosomes 3, 7 and 17, and to the 9p21 region, combines fluorescence in situ hybridization with cytology. It has been reported to have 36% to 95% sensitivity, and 89% to 96% specificity.5, 22 For some of these assays the lowest sensitivity is for detecting small tumors18 and for all it is for detecting low grade tumors.13, 18, 19 These features would be of concern if these tests were used to replace some cystoscopic examinations for monitoring patients with a history of low risk bladder cancer. The results of several other studies done in Canada and
Europe show the performance of the combined ImmunoCyt® (or UCyt™ in Europe) and cytology tests for detecting bladder cancer (table 3).2, 6 –17 Sensitivity in the current study was 81%. Specificity in studies done outside of the United States was 52% to 82%. In studies with poorer sensitivity and specificity technical errors in reporting the assays were identified which, when corrected, improved results (table 3).2, 6 –17 As shown in this study, the ImmunoCyt® test has 81% sensitivity and 73% specificity. It is able to detect recurrent bladder cancer of all grades and stages. Reported variation in the sensitivity of this test depends in part on the degree of training and experience of the technologist performing the assay.15 However, in this study after 1 day of training pathologists were able to pass an interobserver training test, achieving 100% concordance on 5 slides. Additionally, at 1 participating laboratory (Bostwick Laboratories) 40% of cases were reviewed by 2 observers independently. There was 90% agreement between observers with the final diagnosis of disputed cases agreed on by the 2 pathologists who reviewed these cases together. Thus, with training the interpretations of slides were quite reproducible. In addition to the high sensitivity for low stage, low grade tumors observed in this study, the ImmunoCyt® assay combined with cytology detected 100% of carcinomas in situ. Similar results to those in the current study were reported in populations including patients with newly diagnosed cancers and nonbladder cancer controls as well as those being monitored for bladder cancer recurrence.2, 3, 6 – 8 Also, intravesical treatments, such as epirubicin and bacillus Calmette-Guerin, do not appear to significantly affect the performance of the test.23 However, to our knowledge this is the first study to assess the performance of this test exclusively in 1 major area for which a noninvasive marker test would appear to have great usefulness and immediate applicability, namely the surveillance of patients with low and intermediate risk superficial bladder cancer. Highly pertinent to this potential use is the critical new finding that even low risk tumors smaller than 1 cm can be detected with good reliability by ImmunoCyt®. Because the strategy behind cystoscopic surveillance is to permit the detection and treatment of cancers
URINE TEST AND RECURRENT BLADDER CANCER
before they are large enough to cause symptoms, this test seems well suited to be introduced into monitoring programs to decrease the frequency of followup cystoscopy, particularly in patients with low risk urothelial cancer. However, such a policy must be tested prospectively before it is adopted as an appropriate standard of care. CONCLUSIONS
ImmunoCyt® enhances the sensitivity of cytology, which is a specific but not a sensitive method for detecting bladder cancer. The ability of this immunocytochemical test to detect low grade, superficial, small tumors makes it the most suitable available marker to test for monitoring strategies in patients with low risk bladder cancer. REFERENCES
1. American Cancer Society. Cancer Facts and Figures, 2004. Statistics 2004 2. Pfister, C., Chautard, D., Devonec, M., Perrin, P., Chopin, D., Rischman, P. et al: Immunocyt test improves the diagnostic accuracy of urinary cytology: results of a French multicenter study. J Urol, 169: 921, 2003 3. Fradet, Y.: Recent advances in the management of superficial bladder tumors. Can J Urol, 9: 1544, 2002 4. Messing, E. M., Young, T. B., Hunt, V. B., Gilchrist, K. W., Newton, M. A., Bram, L. L. et al: Comparison of bladder cancer outcome in men undergoing hematuria home screening versus those with standard clinical presentations. Urology, 45: 387, 1995 5. Bubendorf, L., Grilli, B., Sauter, G., Mihatsch, M. J., Gasser, T. C. and Dalquen, P.: Multiprobe FISH for enhanced detection of bladder cancer in voided urine specimens and bladder washings. Am J Clin Pathol, 116: 79, 2001 6. Fradet, Y. and Lockhard, C.: Performance characteristics of a new monoclonal antibody test for bladder cancer: ImmunoCyt™. Can J Urol, 4: 400, 1997 7. Olsson, H. and Zackrisson, B.: ImmunoCyt a useful method in the follow-up protocol for patients with urinary bladder carcinoma. Scand J Urol Nephrol, 35: 280, 2001 8. Lodde, M., Mian, C., Negri, G., Berner, L., Maffei, N., Lusardi, L. et al: Role of uCyt⫹ in the detection and surveillance of urothelial carcinoma. Urology, 61: 243, 2003 9. Mian, C., Lodde, M., Comploj, E., Negri, G., Egarter-Vigl, E., Lusuardi, L. et al: Liquid-based cytology as a tool for the performance of uCyt⫹ and Urovysion® Multicolour-FISH in the detection of urothelial carcinoma. Cytopathology, 14: 338, 2003 10. Tetu, B., Tiguert, R., Harel, F. and Fradet, Y.: ImmunoCyt/ uCyt⫹ improves the sensitivity of urine cytology in patients
1241
followed for urothelial carcinoma. Mod Pathol, 18: 83, 2004 11. Lopez, B. M., Aparicio, T. F., Martinez-Valls, G., Villa-Plana, G. H., Garcia, J. L. and Benito, A. C.: Validez del ImmunoCyt en el diagnostico y seguimiento de las neoplasias vesicales uroteliales. Asoc Esp Urol, abstract C55, 2002 12. Toma, M. I., Friedrich, M. G., Hautmann, S. H., Jakel, K. T., Erbersdobler, A., Hellstern, A. et al: Comparison of the ImmunoCyt test and urinary cytology with other urine tests in the detection and surveillance of bladder cancer. World J Urol, 22: 145, 2004 13. Vriesema, J. L., Atsma, F., Kiemeney, L. A., Peelen, W. P., Witjes, J. A. and Schalken, J. A.: Diagnostic efficacy of the ImmunoCyt test to detect superficial bladder cancer recurrence. Urology, 58: 367, 2001 14. Lodde, M., Mian, C., Pycha, A., Wiener, H. and Marberger, M.: Immunocyt™ for the detection of transitional cell cancer of the urinary tract. J Urol, suppl., 161: 150, abstract 574, 1999 15. Bunting, P., Fleshner, N., Kapusta, L., Hersey, K. and Klotz, L.: Detection of bladder cancer via non-invasive methods: a direct comparison of the NMP-22 and immunocyte test. J Urol, suppl., 163: 134, abstract 592, 2000 16. Ceil, G., Zumbragel, A., Paulgen-Nelder, H. J., Hennenlotter, J., Maurer, S., Krauser, S. et al: Accuracy of the ImmunoCyt assay in the diagnosis of transitional cell carcinoma of the urinary bladder. Anticancer Res, 23: 963, 2003 17. Mian, C., Lodde, M., Comploj, E., Palmermo, S., Mian, M., Marziani, F. et al: The UCYT⫹ Test: an update. Eur Urol, 41: abstract 627, 2005 18. Boman, H., Hedelin, H. and Holma¨ng, S.: Four bladder tumor markers have a disappointing low sensitivity for small size and low grade recurrence. J Urol, 167: 80, 2002 19. Tinzl, M., Marberger, M., Horvath, S. and Chypre, C.: DD3PCA3 RNA analysis in urine—a new perspective for detecting prostate cancer. Eur Urol, 46: 182, 2004 20. Grossman, H. B., Gomella, L. G., Fradet, Y., Presti, J. C., Jr., Ritenour, C. W. M., Lerner, S. P. et al: The use of Hexvix and fluorescence cystoscopy as an adjuct in the diagnosis of stage TA/TI urothelial cancer in the urinary bladder. J Urol, suppl., 171: 69, abstract 263, 2004 21. Leyh, H., Marberger, M., Conort, P., Sternberg, C., Pasadoro, V., Pagano, F. et al: Comparison of the BTA stat test with voided urine cytology and bladder wash cytology in the diagnosis and monitoring of bladder cancer. Eur Urol, 35: 52, 1999 22. Sarosdy, M. F., Schellhammer, P., Bokinsky, G., Kahn, P., Chao, R., Yore, L. et al: Clinical evaluation of a multi-target fluorescent in situ hybridization assay for detection of bladder cancer. J Urol, 168: 1950, 2002 23. Lodde, M., Mian, C., Negri, G., Vittadello, F., Comploj, E., Palermo, S. et al: Effect of intravesical instillation on performance of uCYT⫹ test. Urology, 63: 878, 2004
Vol. 174, 1238 –1241, October 2005 Printed in U.S.A.
DOI: 10.1097/01.ju.0000173918.84006.4d
PERFORMANCE OF URINE TEST IN PATIENTS MONITORED FOR RECURRENCE OF BLADDER CANCER: A MULTICENTER STUDY IN THE UNITED STATES EDWARD M. MESSING,*, † LISA TEOT, HOWARD KORMAN,‡ JEANNE UNDERHILL, EDWARD BARKER, BRIAN STORK, JUNQI QIAN AND DAVID G. BOSTWICK From the Departments of Urology (EMM) and Pathology (LT), University of Rochester Medical Center, Rochester, New York, Departments of Urology (HK) and Pathology (JE), William Beaumont Hospital, Detroit, Michigan, Medical Lab Associates (EB, BS), Seattle, Washington, and Bostwick Laboratories, (JQ, DGB), Glen Allen, Virginia
ABSTRACT
Purpose: We assessed the performance of the ImmunoCyt® immunocytochemical test for detecting bladder cancer recurrence in patients with prior superficial bladder cancers compared with cystoscopic and histological findings. Materials and Methods: A total of 341 patients with a history of bladder cancer undergoing monitoring were evaluated at 4 sites. The results of cytology and/or ImmunoCyt® were analyzed for sensitivity and specificity compared with biopsy confirmed cancer. Results: The overall sensitivity of cytology alone, ImmunoCyt® alone and the 2 methods combined was 23%, 81% and 81%, respectively. The specificity of cytology alone, ImmunoCyt® alone and of the 2 methods combined was 93%, 75% and 73%, respectively. The immunocytochemical test was more sensitive than cytology for detecting grades 1 and 2, and stages Ta, T1, and T2 urothelial carcinoma, and it was equally sensitive for detecting grade 3 cancers and carcinoma in situ (CIS). The sensitivity of the combined tests for grades 1 to 3/CIS was 79%, 90% and 82%, while for stages Ta, T1, T2⫹ and CIS it was 83%, 75%, 100% and 100%, respectively. The overall positive and negative predictive values of the combined tests were 37% and 95%, respectively. Importantly the immunocytochemical test could detect 71% of small (less than 1 cm) tumors. Conclusions: ImmunoCyt® is a sensitive test for detecting bladder cancer. Because of its high sensitivity for detecting small tumors, even those of low histological grade, and its high negative predictive value, this test may have a role in decreasing the frequency of cystoscopic examinations for monitoring patients with low risk bladder cancer. KEY WORDS: bladder, bladder neoplasms, neoplasm recurrence, cytology
More than 55,000 cases of bladder cancer are diagnosed yearly in the United States.1 Within 2 years of endoscopic resection of papillary superficial tumors approximately 50% of patients experience recurrence and after 5 years the risk of muscle invasive cancer is 5% to 25%.2 Most urologists consider cystoscopy to be the best method for the diagnosis of bladder cancer and for followup after surgery and treatment. Many urologists recommend followup cystoscopy for noninvasive tumors every 3 months for 1 or 2 years and less frequently thereafter. In patients with high grade tumors who are at 73% risk for recurrence during year 1 after initial treatment and at about 25% risk for subsequent progression to muscle invasion3 frequent followup by cystoscopy may be advisable. However, in patients with lower grade tumors who are at lower risk for recurrence during year 1 after initial treatment3 it may be possible to use alternative, noninvasive monitoring methods to space out cystoscopic examinations.
The most common noninvasive method for monitoring bladder cancer is the detection of exfoliated tumor cells by urinary cytology. Cytology has high sensitivity for detecting high grade tumors but poor sensitivity for detecting low grade tumors, which are the most common type of bladder cancer.4 However, the limitations of cytology, especially for detecting recurrent low grade tumors, and the invasiveness of cystoscopy have increased interest in tests using urinary diagnostic markers. Examples are tests based on human complement related factor, nuclear mitotic apparatus protein and DNA loss of heterozygosity or chromosomal gain in tumor cells exfoliated in urine, as detected by fluorescence in situ hybridization.5 Tests using these markers are more sensitive than cytology but sensitivity is no greater than 60% for low grade tumors and specificity tends to be lower than that of cytology. The ImmunoCyt® test uses a cocktail of 3 monoclonal antibodies directed against urothelial carcinoma antigens in exfoliated cells, including 2 cytoplasmic mucin related antigens and high molecular weight carcinoembryonic antigen, to identify tumor cells shed in urine. The antibodies are labeled with fluorescent markers, which allow detection by fluorescence microscopy. In 5,370 cases evaluated in multiple studies done in Canada and Europe cumulative sensitivity was 85% and specificity was 71% (table 3).2, 6 –17 In addition, ImmunoCyt® also improved sensitivity for detecting low grade, low stage tumors from 50% to 90%.3 However, interval cystoscopy can detect small tumors, for which many marker
Submitted for publication January 25, 2005. Ethics approval for the study was obtained from individual institutions. Supported by DiagnoCure, Quebec, Canada. * Correspondence: Department of Urology, University of Rochester, 601 Elmwood Ave., Box 655, Rochester, New York 14642 (telephone: 585-275-3345; FAX: 585-442-8350; e-mail: edward_messing@ urmc.rochester.edu). † Financial interest and/or other relationship with AstraZeneca, Matritech, DynaCure and Fujirebio. ‡ Financial interest and/or other relationship with DiagnoCure and AstraZeneca. 1238
1239
URINE TEST AND RECURRENT BLADDER CANCER
tests have had difficulties.18 We assessed the performance of ImmunoCyt® and cytology to detect recurrent cancers in a population of patients with prior superficial bladder tumors. MATERIALS AND METHODS
Patients. This study was performed at 4 sites in the United States. From November 2000 to November 2003 voided urine specimens were obtained from 341 patients who were being monitored for bladder cancer recurrence. Except for a history of completely resected stage T1 or less urothelial cancer no specific tumor characteristics were prescribed. An attempt was made to enlist subjects with a history of grades 1 and 2, rather than high grade cancer. Cytological examination of urine samples was performed as well as evaluation using an ImmunoCyt® immunocytochemical test kit at each site. Ethics approval for the study was obtained from individual institutions. Informed consent, obtained from all patients, is on file at the sites. Bladder cancer was diagnosed based on histological confirmation from cystoscopic biopsy. Biopsy samples were prepared for histological staining and grading at each study site. Urine sample preparation. Each 20 to 40 ml voided urine specimen was obtained prior to cystoscopic examination and at least 4 weeks after any urethral instrumentation. Specimens were fixed with an equal volume of 50% ethanol or 50% isopropyl alcohol and then stored at 2C to 8C until testing. ImmunoCyt® fixative buffered solution (1 ml) was mixed with urine and the sample was incubated for 15 minutes. at room temperature (18C to 30C). The sample was then filtered through a polycarbonate membrane with a porosity of 8 m in a device connected to a 50 ml syringe. Cells captured on the filter were blotted onto a silane treated slide and spray fixed with 50% isopropanol before immunocytochemical assay was performed. Control slides were also prepared using positive and negative control cells provided with the kit. ImmunoCyt® assay. Slides were prepared for immunofluorescence staining, as described in the package insert. They were pretreated with 80%, 70% and 50% ethanol, distilled water and Harris hematoxylin with 4% acetic acid. Cells were then incubated with 4 drops of blocking solution for 15 minutes at room temperature in a closed humid chamber and with the ImmunoCyt® antibody cocktail for 1 hour. Slides were then rinsed in phosphate buffered saline containing 0.5% Tween 20 and in pure phosphate buffered saline. Mounting medium (1 drop) was added on the slide before the cover slip. Slides were stored at 4C until evaluation using an epifluorescence microscope fitted with a double filter to detect fluorescein and Texas red emissions. Red fluorescent cells were positive for a glycosylated form of carcinoembryonic antigen and green fluorescent cells were positive for bladder cancer mucins. The test was scored positive as soon as 1 confirmed green or red fluorescent cell was detected. According to manufacturer instructions slides with fewer than 500 cells were considered inadequate for to validate a negative result with ImmunoCyt®. More than 500 microscopic fields (20⫻ objective) per slide were screened. Techni-
cians and cytopathologists were blinded to the results of cystoscopy, biopsy and cytology. Cytology. Slides for conventional urinary cytology were stained according to the standard Papanicolaou procedure. Three cytology categories were defined as negative, suspicious and positive for malignancy, respectively. In this study cases with atypia were included in the negative category and only the presence of malignant cells was considered positive. Sensitivity and specificity. Sensitivity was calculated as the number of samples testing positive by immunocytochemical and/or cytology divided by the number of samples that were positive by biopsy. Specificity was calculated as the number of negative tests divided by the number of samples that were negative by cystoscopy or biopsy, if the patient had suspicious cystoscopy but negative biopsy. For grade, stage and tumor size sensitivity was calculated as the number of samples positive by each testing method divided by the number of samples positive for cancer for which a grade, stage or size could be accurately assessed. Positive and negative predictive values were calculated in the entire cohort tested for which a satisfactory immunocytochemical or cytology specimen was obtained. RESULTS
Of 341 patients who were evaluated by cystoscopy for bladder cancer recurrence 14 (4%) had urine specimens that were excluded because of inadequate sampling, defined as less than 500 cells per slide. Of the remaining 327 patients 61 (19%) had cystoscopically evident tumors, of which 9 were fulgurated, providing no tissue for histology. Of the 52 histologically confirmed cancers 50 were evaluable for tumor stage, 49 were evaluable for tumor grade and 41 were evaluable for tumor size determinations. Tables 1 to 4 show the performance of ImmunoCyt® and cytology overall and for various grades, stages and sizes. In the 42 cases of histologically confirmed bladder cancer in which ImmunoCyt® was positive 5 or fewer cells were positive in slides in 16 cases, 6 to 20 were positive in 13 and greater than 20 were positive in 13. In no instance with a specimen evaluable by immunocytochemistry was cytology positive when immunocytochemistry was not. Importantly the sensitivity for stage T2 cancers and carcinoma in situ (CIS) (7 samples) was 100% for ImmunoCyt® and 71% (5 of 7) for cytology. The sensitivity of ImmunoCyt® and cytology was each the same for grade 3 cancers (82% or 9 of 11). Of tumors smaller than 1 cm 71% were positive by ImmunoCyt®. Sensitivity for grade 1 and stage Ta tumors was 79% and 83%, respectively. DISCUSSION
Cystoscopy is the most specific and sensitive method for detecting new and recurrent bladder cancers. However, it is invasive, resulting in patient discomfort and the risk of infection. In addition, it is difficult to detect flat tumors, such as (CsIS).19 There is growing evidence that fluorescent dye
TABLE 1. ImmunoCyt® and cytology performance % (No./total No.) ImmunoCyt®
Cytology*
ImmunoCyt® ⫹ Cytology
Pos predictive value True pos/true pos ⫹ false-pos 38 (42/110) 37 (42/115) Neg predictive value True neg/true neg ⫹ false-neg 95 (206/216) 95 (206/216) Overall sensitivity True pos/true pos ⫹ false-neg 81 (42/52) 23 (12/52) 81 (42/52) Specificity True neg/true neg ⫹ false-pos 75 (206/274) 93 (255/274) 73 (201/327) Only histologically confirmed cancer for which a test was positive was considered positive and if additional cases for which fulguration was performed without biopsy were each considered to show cancer, positive predictive value would have changed to 51% and 48% for ImmunoCyt® and for Immunocyt® and cytology, respectively, and sensitivity would have changed to 51 of 61 cases (84%) for Immunocyt®, and for Immunocyt® and cytology. * Atypia and dysplasia were not considered positive.
1240
URINE TEST AND RECURRENT BLADDER CANCER TABLE 2. ImmunoCyt® and cytology sensitivity by available tumor grade or stage
Grade or Stage
No. Pts
G1 G2 G3* Ta T1 T2 CIS * G3 and CIS sensitivity 9 of 11
% ImmunoCyt® (No. pts)
Cytology (No. pts)
ImmunoCyt® ⫹ Cytology
79 (22) 90 (9) 67 (4) 83 (29) 75 (6) 100 (2) 100 (5)
7 (2) 10 (1) 67 (4) 6 (2) 50 (4) 0 100 (5)
79 (22) 90 (9) 67 (4) 83 (29) 75 (6) 100 (2) 100 (5)
28 10 6 35 8 2 5 patients (82%).
TABLE 3. Performance of combined ImmunoCyt® and cytology tests for bladder cancer % G1–G2 Sensitivity (No. pts)
References
89–97 (70) Fradet and Lockhard6 87–90 (98) Lodde et al14 Bunting et al15 100 (22) Olsson and Zackrisson7 Vriesema et al13 59–89 (93) Lopez et al11 81–96 (71) Lodde et al8 67–78 (71) Pfister et al2 84–92 (55) Mian et al9 14–64 (21) Ceil et al16 71–79 (64) Tetu et al10 86–91 (30) Toma et al12 79–91 (209) Mian et al17 Present series 79–90 (38) * Low sensitivity due to improper technical performance.
Less than 1 1–3 Greater than 3
% Specificity (No. pts)
100 (25) 95 (45)
76 (102) 74 (286) 52 (68) 69 (83) 64 (64) 76 (287) 68 (123) 81 (548) 66 (92) 82 (87) 61 (734) 72 (82) 73 (1,585) 73 (274)
45 (23)* 100 (8) 52 (22)* 93 (54) 100 (31) 87 (68) 96 (25) 80 (5) 93 (40) 83 (12) 99 (89) 82 (11)
TABLE 4. ImmunoCyt® and cytology sensitivity on available tumor sizes Size (cm)
% G3/CIS Sensitivity (No. pts)
% (No./total No.) ImmunoCyt®
Cytology
71 (12/17) 84 (16/19) 60 (3/5)
18 (3/17) 26 (5/19) 20 (1/5)
increases the sensitivity of cystoscopy. In 1 study standard cystoscopy did not detect a substantial number of bladder tumors that were visible using fluorescence endoscopy.3, 20 In contrast, cytological examination of exfoliated cells shed in urine is a specific but not a sensitive method. New methods for detecting bladder cancer that are based on urinary markers have become available in the last 10 years. Diagnostic tests for recurrent bladder cancer that are available in the United States, Canada and Europe have a sensitivity range of 46% to 100% and a specificity range of 60% to 100%.16 Four tests are available in the United States. The BTA Stat® and BTA Trak® tests detect human complement factor H related protein in voided urine.21 Sensitivity is 57% to 83% depending on grade and stage, and specificity is 46% to 73%. The nuclear mitotic apparatus protein assay, which detects a urinary nuclear mitotic apparatus protein, is reported to have 47% to 100% sensitivity and 60% to 70% specificity. The UroVysion™ fluorescence in situ hybridization assay, which uses probes to the centromeres of chromosomes 3, 7 and 17, and to the 9p21 region, combines fluorescence in situ hybridization with cytology. It has been reported to have 36% to 95% sensitivity, and 89% to 96% specificity.5, 22 For some of these assays the lowest sensitivity is for detecting small tumors18 and for all it is for detecting low grade tumors.13, 18, 19 These features would be of concern if these tests were used to replace some cystoscopic examinations for monitoring patients with a history of low risk bladder cancer. The results of several other studies done in Canada and
Europe show the performance of the combined ImmunoCyt® (or UCyt™ in Europe) and cytology tests for detecting bladder cancer (table 3).2, 6 –17 Sensitivity in the current study was 81%. Specificity in studies done outside of the United States was 52% to 82%. In studies with poorer sensitivity and specificity technical errors in reporting the assays were identified which, when corrected, improved results (table 3).2, 6 –17 As shown in this study, the ImmunoCyt® test has 81% sensitivity and 73% specificity. It is able to detect recurrent bladder cancer of all grades and stages. Reported variation in the sensitivity of this test depends in part on the degree of training and experience of the technologist performing the assay.15 However, in this study after 1 day of training pathologists were able to pass an interobserver training test, achieving 100% concordance on 5 slides. Additionally, at 1 participating laboratory (Bostwick Laboratories) 40% of cases were reviewed by 2 observers independently. There was 90% agreement between observers with the final diagnosis of disputed cases agreed on by the 2 pathologists who reviewed these cases together. Thus, with training the interpretations of slides were quite reproducible. In addition to the high sensitivity for low stage, low grade tumors observed in this study, the ImmunoCyt® assay combined with cytology detected 100% of carcinomas in situ. Similar results to those in the current study were reported in populations including patients with newly diagnosed cancers and nonbladder cancer controls as well as those being monitored for bladder cancer recurrence.2, 3, 6 – 8 Also, intravesical treatments, such as epirubicin and bacillus Calmette-Guerin, do not appear to significantly affect the performance of the test.23 However, to our knowledge this is the first study to assess the performance of this test exclusively in 1 major area for which a noninvasive marker test would appear to have great usefulness and immediate applicability, namely the surveillance of patients with low and intermediate risk superficial bladder cancer. Highly pertinent to this potential use is the critical new finding that even low risk tumors smaller than 1 cm can be detected with good reliability by ImmunoCyt®. Because the strategy behind cystoscopic surveillance is to permit the detection and treatment of cancers
URINE TEST AND RECURRENT BLADDER CANCER
before they are large enough to cause symptoms, this test seems well suited to be introduced into monitoring programs to decrease the frequency of followup cystoscopy, particularly in patients with low risk urothelial cancer. However, such a policy must be tested prospectively before it is adopted as an appropriate standard of care. CONCLUSIONS
ImmunoCyt® enhances the sensitivity of cytology, which is a specific but not a sensitive method for detecting bladder cancer. The ability of this immunocytochemical test to detect low grade, superficial, small tumors makes it the most suitable available marker to test for monitoring strategies in patients with low risk bladder cancer. REFERENCES
1. American Cancer Society. Cancer Facts and Figures, 2004. Statistics 2004 2. Pfister, C., Chautard, D., Devonec, M., Perrin, P., Chopin, D., Rischman, P. et al: Immunocyt test improves the diagnostic accuracy of urinary cytology: results of a French multicenter study. J Urol, 169: 921, 2003 3. Fradet, Y.: Recent advances in the management of superficial bladder tumors. Can J Urol, 9: 1544, 2002 4. Messing, E. M., Young, T. B., Hunt, V. B., Gilchrist, K. W., Newton, M. A., Bram, L. L. et al: Comparison of bladder cancer outcome in men undergoing hematuria home screening versus those with standard clinical presentations. Urology, 45: 387, 1995 5. Bubendorf, L., Grilli, B., Sauter, G., Mihatsch, M. J., Gasser, T. C. and Dalquen, P.: Multiprobe FISH for enhanced detection of bladder cancer in voided urine specimens and bladder washings. Am J Clin Pathol, 116: 79, 2001 6. Fradet, Y. and Lockhard, C.: Performance characteristics of a new monoclonal antibody test for bladder cancer: ImmunoCyt™. Can J Urol, 4: 400, 1997 7. Olsson, H. and Zackrisson, B.: ImmunoCyt a useful method in the follow-up protocol for patients with urinary bladder carcinoma. Scand J Urol Nephrol, 35: 280, 2001 8. Lodde, M., Mian, C., Negri, G., Berner, L., Maffei, N., Lusardi, L. et al: Role of uCyt⫹ in the detection and surveillance of urothelial carcinoma. Urology, 61: 243, 2003 9. Mian, C., Lodde, M., Comploj, E., Negri, G., Egarter-Vigl, E., Lusuardi, L. et al: Liquid-based cytology as a tool for the performance of uCyt⫹ and Urovysion® Multicolour-FISH in the detection of urothelial carcinoma. Cytopathology, 14: 338, 2003 10. Tetu, B., Tiguert, R., Harel, F. and Fradet, Y.: ImmunoCyt/ uCyt⫹ improves the sensitivity of urine cytology in patients
1241
followed for urothelial carcinoma. Mod Pathol, 18: 83, 2004 11. Lopez, B. M., Aparicio, T. F., Martinez-Valls, G., Villa-Plana, G. H., Garcia, J. L. and Benito, A. C.: Validez del ImmunoCyt en el diagnostico y seguimiento de las neoplasias vesicales uroteliales. Asoc Esp Urol, abstract C55, 2002 12. Toma, M. I., Friedrich, M. G., Hautmann, S. H., Jakel, K. T., Erbersdobler, A., Hellstern, A. et al: Comparison of the ImmunoCyt test and urinary cytology with other urine tests in the detection and surveillance of bladder cancer. World J Urol, 22: 145, 2004 13. Vriesema, J. L., Atsma, F., Kiemeney, L. A., Peelen, W. P., Witjes, J. A. and Schalken, J. A.: Diagnostic efficacy of the ImmunoCyt test to detect superficial bladder cancer recurrence. Urology, 58: 367, 2001 14. Lodde, M., Mian, C., Pycha, A., Wiener, H. and Marberger, M.: Immunocyt™ for the detection of transitional cell cancer of the urinary tract. J Urol, suppl., 161: 150, abstract 574, 1999 15. Bunting, P., Fleshner, N., Kapusta, L., Hersey, K. and Klotz, L.: Detection of bladder cancer via non-invasive methods: a direct comparison of the NMP-22 and immunocyte test. J Urol, suppl., 163: 134, abstract 592, 2000 16. Ceil, G., Zumbragel, A., Paulgen-Nelder, H. J., Hennenlotter, J., Maurer, S., Krauser, S. et al: Accuracy of the ImmunoCyt assay in the diagnosis of transitional cell carcinoma of the urinary bladder. Anticancer Res, 23: 963, 2003 17. Mian, C., Lodde, M., Comploj, E., Palmermo, S., Mian, M., Marziani, F. et al: The UCYT⫹ Test: an update. Eur Urol, 41: abstract 627, 2005 18. Boman, H., Hedelin, H. and Holma¨ng, S.: Four bladder tumor markers have a disappointing low sensitivity for small size and low grade recurrence. J Urol, 167: 80, 2002 19. Tinzl, M., Marberger, M., Horvath, S. and Chypre, C.: DD3PCA3 RNA analysis in urine—a new perspective for detecting prostate cancer. Eur Urol, 46: 182, 2004 20. Grossman, H. B., Gomella, L. G., Fradet, Y., Presti, J. C., Jr., Ritenour, C. W. M., Lerner, S. P. et al: The use of Hexvix and fluorescence cystoscopy as an adjuct in the diagnosis of stage TA/TI urothelial cancer in the urinary bladder. J Urol, suppl., 171: 69, abstract 263, 2004 21. Leyh, H., Marberger, M., Conort, P., Sternberg, C., Pasadoro, V., Pagano, F. et al: Comparison of the BTA stat test with voided urine cytology and bladder wash cytology in the diagnosis and monitoring of bladder cancer. Eur Urol, 35: 52, 1999 22. Sarosdy, M. F., Schellhammer, P., Bokinsky, G., Kahn, P., Chao, R., Yore, L. et al: Clinical evaluation of a multi-target fluorescent in situ hybridization assay for detection of bladder cancer. J Urol, 168: 1950, 2002 23. Lodde, M., Mian, C., Negri, G., Vittadello, F., Comploj, E., Palermo, S. et al: Effect of intravesical instillation on performance of uCYT⫹ test. Urology, 63: 878, 2004