- Email: [email protected]
CAN URINE BOUND DIAGNOSTIC TESTS REPLACE CYSTOSCOPY IN THE MANAGEMENT OF BLADDER CANCER?
Vol. 169,1876-1880, June 1998 Printed in US&
CAN URINE BOUND DIAGNOSTIC TESTS REPLACE CYSTOSCOPY IN
THE MANAGEMENT OF BLADDER CANCER? H.G. WIENER,* CH. MIAN, A. HAITEL, A. PYCHA, G. SCHATZL AND M. MARBERGER From the Deparments of Urology and Pathology, University of Vienna, Vienna, Austria
ABSTRACT
Purpose: We compare the diagnostic value of NMP22t and BTA stat$ testing, and QUANTICYTg computer assisted dual parameter image analysis to cytology and cystoscopy in patients who had symptoms suggestive of transitional cell cancer or were being followed after treatment for that disease. Materials and Methods: We prospectively evaluated voided urine and/or barbotage specimens from 291 patients a mean of 65.2 years old. All voiqed urine samples were evaluated by quick staining and standard cytology, the BTA stat 1-step qualitative assay (which detects a bladder tumor associated antigen) and the NMP22 test (which detects a nuclear mitotic apparatus protein). In addition, barbotage specimens were evaluated by QUANTICYT computer assisted dual parameter image analysis. All patients underwent subsequent cystoscopy and biopsy evaluation of any suspicious lesion. Sensitivity, specificity, and the predictive value of positive and negative results were determined in correlation with endoscopic and histological findings. Results: In 91 patients with histologically proved transitional cell carcinoma overall sensitivity was 48,57, 58, 59 and 59% for the NMP22 test, the BTA stat test, rapid staining cytology of barbotage samples, rapid staining cytology of voided urine specimens and image analysis, respectively. For histological grades 1 to 3 underlying transitional cell carcinoma sensitivity was 17,61and 90% for urinary cytology, 48,58 and 63% for the BTA stat test, and 52,45 and 50% for the NMP22 test, respectively. Specificity was 100%for cytology, 93% for image analysis, 70% for the NMP22 test and 68% for the BTA stat test. Conclusions: Immunological markers are superior to cytological evaluation and image analysis for detecting low grade transitional cell carcinoma but they have low specificity and sensitivity in grade 3 transitional cell carcinoma. Urine bound diagnostic tools cannot replace cystoscopy. KEYWORDS: carcinoma, transitional cell; bladder; cytology;cystoscopy
Cystoscopy has the pivotal role for detecting primary and recurrent transitional cell carcinoma of the bladder but even with flexible instruments it remains invasive and bothersome to the patient. Exophytic tumors are reliably diagnosed but flat transitional cell carcinoma, particularly carcinoma in situ, remains an endoscopic dilemma. Therefore, exfoliative urinary cytology of voided urine and barbotage specimens is routinely performed to complement ~ystoscopy.~ Its sensitivity is poor in low grade transitional cell carcinoma, which accounts for 60 to 70% of all primary bladder tumors: but it is unsurpassed in high grade transitional cell carcinoma, particularly carcinoma in situ.' Unfortunately the results of urinary cytology depend on the training and expertise of the cytopathologist and, as in all morphological evaluations, they have high intra-observer and interobserver ~ a r i a t i o n . ~ Recently simple methods for detecting tumor antigens in the urine have become commercially a ~ a i l a b l eRapid .~ staining techniques and quantitative computer assisted image analysis systems' have been developed in parallel in an attempt to decrease processing time and eliminate investigator dependent cytological errors. We prospectively assessed the diagnostic value of these new approaches, specifically to determine whether they eliminate the need for cystocopy.
MATERIALS AND METHODS
From January to October 1996 barbotage and/or voided urine specimens were prospectively obtained from 199 men and 92 women 17 to 90 years old (mean age 65.2) who had symptoms suggestive of bladder tumors or were being followed after treatment of transitional cell carinoma of the bladder (101 patients). To exclude from study patients with tumors persisting after transurethral resection only those proved to be tumor-free by at least 1 negative cystoscopy 3 months after transurethral resection were enrolled. In all cases voided urine samples were available and rigid or flexible cystoscopy was performed. Barbotage specimens were obtained from 200 patients by rinsing the bladder with 100 to 150 ml. saline, usually through the cystoscope. Any suspicious lesion detected was biopsied and/or removed by transurethral resection. Histopathological grading and staging were done according to the World Health Organization' and TNM7 classifications, respectively. Voided urine samples of 50 to 100 ml. were divided into 3 aliquots, coded, and evaluated by cytology with rapid and standard staining as well as by the BTA stat and NMP22 tests. In accordance with method specific protocols specimens were used fresh or were immediately frozen and stored at -2oc. Accepted for publication December 26, 1997. BTA stat test. In principle the BTA stat test is a 1-step *FLequesta for re rints: Department of Urology, University of qualitative assay that recognizes a tumor antigen reported to Vienna, Wiihringer 8iirtel 18-20, A-1090 Vienna, Austria. be a member of the complement factor H family that is tMatritech, Inc., Newton, Massachusetts. associated with transitional cell carcinoma.' Five drops of 1 SBard Dia ostics, Redmond, Washington. $Gentian gentific Software, Niawier, The Netherlands. aliquot of fresh urine or urine previously frozen at - 20C are 1876
1877
URINE BOUND DIAGNOSTIC TESTS AND CYSTOSCOPY IN BLADDER CANCER
placed on the entrance pool of a small immunochromatoTABLE1. Results of all methods tested based on voided urine graphic environment. The urine reacts with a colloidal gold samples of 291 patients, including 200 free of transitional cell carcinoma conjugated antibladder tumor associated antigen antibody. Antigen conjugate complexes are captured by another anti96 Urinarv Cvtolow 9b NMP22 96 BTA Stat bladder tumor associated antigen antibody, forming a visible Sensitivity 59 48 57 line. A positive result, as indicated by a red line, is shown Specificity 100 69 68 after 5 minutes. The plate includes an internal control zone Predictive value: POS.result 100 41 45 wntaining an immobilized reagent that captures the conjuNeg. result 84 74 78 gate independently of the presence or absence of the antigen and, thus, serves as the test quality control. NMPZZ test. Another aliquot of all voided urine was added TABLE2. Results of dl methods based on voided urine and to a stabilizing solution containing protein stabilizers, protease inhibitors and buffers, and frozen at -2OC. The nuclear hiarbotage specimens of 208 patients, including 11 7 free of tmnsitwna1 cell carcinoma mitotic apparatus protein known to correlate with transitional cell carcinomag is detected using an NMP22 test kit. It represents an enzyme-linked immunoassay in a microplate strip-well format using the monoclonal antibodies mAb 59 58 59 48 57 302-18 and mAb 302-22. According to Miyanaga et al the Sensitivity speciiicity 100 100 93 70 68 cutoff point for bladder cancer positivity is 10 units per ml.' Predictive value: QUANTZCYT analysis. Dual parameter morphometry of 100 100 76 54 56 POS. result Nen. result 78 77 76 66 70 the Feulgen stained barbotage specimens was done using the QUANTICYT computer assisted static image analyzing system to determine the individual risk of bladder cancer.' The estimation of a low, intermediate or high risk of bladder cancer is based on the combination of the deoxyribonucleic logical and morphometric evaluation, since the cellular conacid (DNA) parameter 2cDI and a morphometric shape tent varied significantly. Of all cytospin preparations of parameter (PASS). The DNA parameters 2c-deviation bladder washings 19.5% contained less than 50 transitional index (2cDI) and 5c-exceeding rate are used, as defined epithelial cells, as necessary for valid cytological examinaby B o ~ k i n g .Briefly, ~ cytospin preparations of ethanol- tion, and were discarded. Of all barbotage specimens 30% did polyethylenglycol fixed barbotage specimens are stained ac- not contain the minimum of 100 nuclei of transitional epicording to the Feulgen SchifF method using cold hydrolyza- thelial cells needed by definition for computer assisted image tion.' Nuclei are imported to the system via a microscope, analysis. Based on negative and positive cytological classifications analyzed by software, evaluated visually again and then finally calculated using lymphocytes as an internal reference. there was 100% agreement for the 2 cytological staining For statistical evaluation only specimens at high risk for methods. In addition, the results of the primary microscopic bladder cancer are defined as positive. Those in which less evaluation were in accordance with those of the independent than 100 epithelial nuclei are counted are considered insuf- review. Ten specimens each of the BTA stat and NMP22 tests ficient. were evaluated twice and results were equivalent each time Cytology. An aliquot of voided urine and barbotage speci- when NMP22 test results were reduced to a qualitative statemens was centrifuged and 2 or 3 cytospin slides were pre- ment of values greater than or less than 10 units per ml. The pared from the sediment. One cytospin preparation per spec- reproducibility of the QUANTICYT study was evaluated in h e n was air dried by gently warming the glass slide. Air 20 bladder washings. Freshly prepared cytospin specimens dried specimens were immediately stained within 40 seconds were tested at our institution and another laboratory. Docuby the rapid Hemacolor staining kit* technique, which re- mented risk factors were concordant in all cases. Table 3 shows the sensitivity of all methods tested corresults in cellular features equivalent to those of MayGriinwald Giemsa staining. Therefore, detailed cytolo@cal lated with underlying transitional cell carcinoma. We noted descriptions were available 20 minutes after the start of the major differences. Sensitivity of the cytological evaluation of Preparation and before cystoscopy. Diagnostic results were voided urine and barbotage specimens increased as transicategorized as negative, including diagnoses negative for tional cell carcinoma grade and stage increased. We observed and atypia regardlei%of atypk grade, a similar increase in sensitivity for computer assisted dual a b i a andor -cy and positive, inchding all specimens described as suspicious parameter image analysis. In contrast, for the immunological and/or positive for malignancy. The remaining specimens were " 2 2 and BTA stat tests sensitivity was 48 and 52,58 and wet fixed with a polyethylenglyoolethanol mixture. Wet 6xed Specimens underwent the Papanicohu staining protocol. logical evaluation of the PapanicouIaou stained specimens was TABLE3. Sensitivity of all tested methods s t m t i h d according to done independently on the subsequent day following the same classification scheme. According to routine quality Control histological gmde and stage of underlying transitional cell cancer in 91 patients methods 10% of the cytological specimens were reviewed inde% Quick Staining pendently by another cytologist. Specificity, sensitiGty and p s w % 96 Itwe predictive value were calculated for all tests ev+uated Urinary Barbotage Q U A " " 2 2 BTAStat wing the results of cystoscopic evaluation and histologd exCytology Cytology amination of the resected lesions as the gold standard. RESULTS
Tables 1 and 2 list overdl sensitivity, specificity, and the Predictive values of positive and negative results for all l l based on the examination of methods studied.~ techniques voided urine p d u c e d representative results. However, barhtage samples proved to be a less reliable source for cyto-
* Merck, Darmstadt, Germany.
Grade (No. pts.): 1(23*) 2 (38*) 3 (30*)
17 61 90
22(38)
55(65) W(96)
.
26(46) 55(65)
90(96)
52 45 50
48
58 63
(19')
*Values in parenthesea represent valid specimens only and reflect true results.
1878
URINE BOUND DIAGNOSTIC TESTS AND CYSTOSCOPY IN BLADDER CANCER
45,and 50 and 63% for grades 1 to 3 transitional cell carcinoma, respectively, showing that these testa depended less on resected tumor grade. Cytological evaluation had the highest specificity and predictive value of a positive result when calculated for the study population overall. Table 4 lists the clinical diagnoses in patients free of bladder malignancy in whom the immunological NMP22 and BTA stat tests indicated false-positive results. Tables 5 and 6 show sensitivity, specificity,and predictive value of a positive and negative result using all tested methods in patients followed after treatment of transitional cell carcinoma. DISCUSSION
Various predominantly monoclonal antibodies directed against tumor associated antigens are receiving increasing attention for the primary diagnosis and followup of bladder ~ a n c e r Methods .~ using immunocytochemical tests or the immunological detection of soluble antibodies in voided urine specimens are of particular interest because they are based on specimens obtained noninvasively. For example, Banks et a1 reported that a soluble version of E-cadherin, previously used as an immunohistochemical marker only, also has diagnostic potential in bladder cancer." Klein et al found that CK-20 in the voided urine of patients with bladder cancer has 82.8% sensitivity and 100.0% specificity, and it a pears to be a most interesting future diagnostic parameter.J'Kavaler et al reported that the presence of telomerase activity in the voided urine of patients with bladder cancer has 91% sensitivity.12 Unfortunately these assays are technically complicated, and require highly sophisticated laboratory expertise and equipment. They are not currently available for routine urological testing. The BTA test to detect basement membrane complexes was the fist technically simple, urine bound test for bladder cancer.13 However, sensitivity and specificity proved to be unsatisfactory in clinical use.14 The BTA stat test that we used identifies a distinctly different tumor antigen of the complement factor H family? It is designed as a rapid and simple o0ice test that provides qualitative results within minute^,'^ and it is superior to conventional cytological evaluation for detecting grade 1 transitional cell carcinoma. Unfortunately its low 68% specificity in our study limits its clinical impact. Our study population included patients with urolithiasis and chronic cystitis, as expected in clinical practice when evaluating for potential transitional cell carcinoma. False-positive results were documented in 27 and 59% of these patients, respectively. Specificity in our study is comparable to that in the double-blind multicenter study of Leyh et al, who noted 64% specificity for the BTA stat test.15
TABLE5. Sensitivity, specifiity, and predictive value of positive and negative results for all methods tested based on voided urine samples for 101 patients followed a@r tmnsurethml resection for tmnsitional cell carcinoma (87 free of recurrence) Sensitivity speeiscity M i d i v e value: POS. result Neg. result
96 Urinary Cytology 64 100
%" 2 2
100 93
% BTA Stat
14 12
29 72
8
14 86
84
TABLE6. Specificity, and predictive value of positive and negative results of all methods based on voided urine and barbotage specimens of 79 patients followed after transurethral resection of tmnsitional cell carcinoma (69 free of recurrence) %
%
%
%
%
Cytology urinaryBarbotage Cytology QUANTICYT NMp22 BTAStat Sensitivity Specificity
80 100
80 100
P;edictivi value: Pos. result 100 100 Neg. result 97 97 * Valid specimens only and true results.
60 (67)* 87 (93):
68
40 68
40 94
8 85
15 89
20
Sarosdy et al reported 72% specificity in the voided urine of patients with benign genitourinary abnormalities.16 The same antigen may also be analyzed quantitatively using the BTA-TRAK* test, a modified enzyme immunoassay. Ishak et al reported 74% specificity with this method, which was slightly higher than with the BTA stat test,17 as described. They also reported 85% specificity for the original bladder tumor associated antigen test, which has not been reproduced by others. Sarosdy et al reported greater than 90% specificity for the BTA stat test in healthy volunteers only.16 However, as stated by Layfield, specificity based on healthy volunteers does not reflect the regular clinical situation." Our calculation of specificity was based on concurrent cystoscopic findings. Theoretically tumor antigens may appear in the urine before the tumor becomes visible on endoscopy, similar to when urinary cytology is positive months before cystoscopy is positive.'. Therefore, the significance of a positive BTA stat test with otherwise negative findings must still be considered unknown. Leyh16 and Sarosdy" et a1 reported 83 to 95%sensitivity of the BTA stat test for grade 3 transitional cell carcinoma. We did not reproduce these data in our study. This discrepancy may result from the low numbers of investigated patients per study, since both included less than 50 with grade 3 transitional cell carcinoma. Calculation of the mean of documented values results in 80% sensitivity of the BTA stat test for grade 3 transitional cell carcinoma. Compared with voided urine cytology the BTA stat test appears to be less sensitive TABLE4. Clinical features of patients free of malignant bladder for this aggressive tumor. disease and false-positive results of immunological markers In our study specificity and sensitivity for detecting the No. False-pos. nuclear matrix protein nuclear mitotic apparatus protein Results (%) WMl"MP2 test) in voided urine were equivalent or similar to No. F'ta. those of the BTA stat test. The data compare favorably with BTAStat "22 the findings of Rodriguez-Villanueva et al, who reported Followup aRer tranurethral reseetion, 87 24 (38) 24 (38) overall 53.8% specificity for the NMP22 test.'' In contrast, neg. C Y ~ ~ P Y Cystitis 22 13 (20) 9 (14) Miyanaga et al noted 80.9% sensitivity and 64.3% specificiUpper tract urnlithiasis 15 4 (6) 4 (6) ty.' The discrepancy may again result from the number of Benign lesions of lower urinary tract 25 12 (19) 10 (16) voided urine specimens in the studies. Our series included and followup (benign prostatic hypernearly twice the specimens evaluated by Miyanaga et al. plasia, hemangioma, nephrogenic adFurthermore, differences were evident for the reported perenoma and so forth) 11 Malignancies of nonurological sites and centage of false-positive test results, especially in cases of followup cystitis. Miyanaga et a1 observed 19 of 19 false-positive reMicrohematuria, no evidence of transi40 sults (100%) in cystitis, whereas we noted only 41%. In both tional cell Ca - Totals
200
64 (100)
63 (100)
* Bard Diagnostics, Redmond, Washington.
URINE BOUND DIAGNOSTIC TESTS AND CYSTOSCOPY IN BLADDER CANCER
I
1879
CONCLUSIONS studies technical errors may be excluded because standards and controls are included in the test kits. Urinary erythroThe BTA stat and " 2 2 tests, and QUANnCYT comeytes have only marginal impact on the results? and so puter assisted dual parameter image analysis improve the microhematuria does not explain the differences. Variation detection of transitional cell carcinoma, particularly grade 1 in the storage time of the voided urine samples may be a tumors, compared to cytological evaluation. The value of the possible explanation. In our study the storage time was lim- immunological markers is limited by low sensitivity in grade ited to a maximum of 3 weeks. At least for the NMP22 test 3 transitional cell carcinoma as well as low specificity. None storage conditions appear to be of high importance. Voided of the urine bound diagnostic tools that we investigated reurine specimens must be stored in special vials containing a places cystoscopy. stabilizer, which complicates clinical management. However, The BTA stat and some NMP22 test kits were provided by the test is designed as an enzyme-linked immunoassay that allows quantitative evaluation of nuclear matrix protein 22 Bard Diagnostics, Redmond, Washingtan, and Fa. Bender, levels. Although not the aim of the present study, the quan- Vienna, Austria. The QUANTICYT system was provided titative NMP22" and BTA-TRAK17 tests may help to eval- by Gentian Scientific Software, Niawier, The Netherlands. uate the individual risk of tumor recurrence. Soloway et al Peter van Stratum and Pim Peelen, Urological Development previously proved the usefulness of the NMP22 test for de- and Diagnostics Laboratory, Wijchen, The Netherlands, performed double analyses of bladder washings, and 0. Haitel tecting persistent or rapidly recurring bladder cancer?' performed the statistical analysis. Whereas the value of computerized image analysis of voided urine specimens remains a matter of controversy: the determination of nuclear total DNA content in barbotage REFERENCES samples has been shown to improve diagnostic power, par1. Wiener, H.G., Vooijs, G. P. and van't Hof-Gmotenboer, B.: Acticularly in recurrent superficial bladder cancer. Q U A " curacy of urinary cytology in the diagnosis of primary and software combines 2 morphometric parameters, a total DNA recurrent bladder cancer. Acta Cytol., 37: 163,1993. and nuclear shape descriptor! Specificity and grade 3 trans2. Raghavan, D.,Shipley, W. U., Garnick, M. B., Russell, P. J. and itional cell carcinoma related sensitivity of dual parameter Richie, J. P.: Biology and management of bladder cancer. New image analysis surpassed those of the immunological markEngl. J. Med., 322 1129,1990. 3. Ooms, E. C. M., Kurver, P. H. J., Veldhiuzen, R. W., Alons, C. L. ers. Depending on whether insufficient barbotage specimens and Boon, M. E.: Morphometric grading of bladder tumors in were excluded we noted sensitivity similar to or higher than comparison with histologic gradmg by pathologists. Hum. that of conventional cytologicalevaluation but 7% lower specPath., 144: 140, 1983. ificity. The test decreases investigator dependent variation 4. Fradet, Y. and Cordon-Cardo, C.: Critical appraisal of tumor and, thus, complements clinically important rapid cytological markers in bladder cancer. Sem. Urol., 11: 145,1993. evaluation at the cost of slightly lower specificity. However, 5. van der Poel, H. G., Witjes, J. A., van Stratum, P., Boon, M. E., this test is more time-consuming and barbotage samples are Debruyne, F. M. and Schalken, J. A.: Quanticyt: karyometric more invasive than voided urine samples. The value of all analysis ofbladder washings for patients with superficial bladder cancer. Urology, 48.357, 1996. bladder washing, cell bound diagnostic tools depends on the 6. Mostofi, F.K,Solbin, L. H. and Torloni, H.: Histological Typing quality and amount of cells in the specimen. Of the speciof Urinary Bladder Tumors.Geneva: World Health Organizamens included in our study 30% proved insufficient for tion, 1973. QUANTICYT analysis, while van der Poel et al reported 8% 7. TMN Classification of Malignant Tumors, 4th ed. Union Interinsufficient specimens after obtaining only 25 ml. of bladder nationale Contre le Cancer. Geneva: International Union washings through a catheter or cystoscope.' Most of our Against Cancer, 1987. barbotage specimens were obtained by irrigation through the 8. Kinders. R., Jones, T.. Root, R., Murchison. H., Bruce. C.. port of flexible cystoscopes, which is extremely thin. NeverWilliams, L. and Hass; G. M.: H&an bladder.tumor antigen $ a member of the RCA (regulators of complement activation) theless, flexible cystocopy is the standard technique, at least gene family. J. Urol., part 2,157: 284 abstract 110, 1997. in men. Laboratory bound procedures followed the same pro9. Miyanaga, N., h a , H., Ishikawa, S., Ohtani, M., Noguchi, R., tocol. Kawai, K, Koiso, K, Kobayashi, M., Koyama, A. and The clinical usefulness of a diagnostic test also depends on Takahashi, T.: Clinical evaluation of nuclear matrix protein procedure time, personnel and material needed (table 7). ( " 2 2 ) in urine as a novel marker for urothelial cancer. Eur. With overall procedure times of 5 and 25 minutes, respecUrol., 31: 163, 1997. tively, the BTA stat test and rapid staining cytology are the 10. Banks,R. E., Porter, W. H.,Whelan, P., Smith, P. H. and Selby, most rapid. Calculating expenses for staff and materials for P. J.: Soluble forms of the adhesion E-cadherin molecule in 1,000 specimens per year and amortization times of 3 and 5 urine. J. Clin. Path., 48: 179,1995. Years, respectively, the BTA stat test and rapid staining 11. Klein, A., Zemer, R., Buchumensky, V., Klaper, R., Nissenkom, I. and Saba, K: Detection of bladder carcinoma: a urine test, urinary cytology are definitively the most cost-effectivemethbased on cytokeratin expression. J. Urol., part 2, 157: 3394 ods. Cytological evaluation adds information on all corpuscuabstract 1326,1997. 1ar elements found within the specimens but it requires a 12. Kavaler, E., Shu, W.-P., Chang, Y., Droller, M. J. and Liu, highly trained and qualified staff. In contrast, the qualitative B. C.-S.: Detection of human bladder cancer cells in voided immunochromatographic BTA stat assay is a simple office urine samples by assaying for the presence of telomerase acprocedure that may be performed by the urologist. tivity. J. Urol., part 2, 157: 338A,abstract 1321,1997. 13. Sarosdy, M.F.,deVere White, R. W., Soloway, M. S., Sheinfeld, J., Hudson, M. A, Schellhammer, P. F., Jarowenko, M. V., TABLE 7. Amortized costs in United States dollars and evalwtwn
time for all methods tested BTA
5
s i n sJspeeimen
Quick Staining
Stat
N m 2
15 5
180
54
QUmW
41 210
Urinary cytology
42 25
Adams, G. and Blumenstein, B. A,: Results of a multicenter trial using the BTA test to monitor for and diagnose recurrent bladder cancer. J. Urol., 161:379,1995. 14. D'Hallewin, M. A. and Baert, L.: Initial evaluation of the bladder tumor antigen test in superficial bladder cancer. J. Urol., 165: 475,1996. 15. Leyh, H., Marberger, M., Pagano, F., Bassi, P., Sternberg, C. N., Pansadom, V., Conort, P., Bmn-Gibod, L. and Thoeke, K. R: Results of a European multicenter trial comparing the BTAstat test to urine cytology in patients suspected of having bladder cancer. J. Urol., part 2, 151: 3374 abstract 1316, 1997.
1880
URINE BOUND DIAGNOSTIC TESTS AND CYSTOSCOW IN BLADDER CANCER
16. samedy. M. F., Hudson, M. A, Ellis, W.J., Soloway, M. S., 18. Layfield, L.:Editorial comment. J. Urol., 15k 383, 1995. deVere White, R. W.,Sheinfield. J., Jarowenk, M. V., 19. Rodriguez-Villanueva,J., Dinney, C. P. N., Grossman, H. B. and Fritsche, H. A.: Evaluation of the NMP22 immunoassay in the Schellhammer, P. F., Messing, E. M., Schervish, E. W.,Patel, detedion of transitional cell carcinoma (TCC) of the urinary J. V., Chodak, G. W.,Lamm, D. L.,Johnson, R. D., Henderson, M., Adams, G., Blumenstein, B. A, Thoelke, K. R., Brunelle, tract. J. Urol., part 2,151: 3364 abstract 1314,1997. S. L.,Pfalzgraf, R. D. and Murchimn, H. A.: Detection of 20. Soloway, M. S., Briggman, J. V., Carpinito, G. A., Chodak, G. W., Church, P. A.,Lamm,D.L.,Lange, P., Messing, E., Pasciak, recurrent bladder cancer using a new one-step test for bladder tumor antigen. J. Urol., part 2,161: 3374 abstract 1318,1997. R. M., Reservitz, G. B., Rukstalis, D. B., Sarosdy, M. F., 17. Ishak, L.M.. Enfield, D. L.,Sarosdy, M. F. and the Multicenter Stadler, W.M., Thiel, R. P. and Hayden, C. L.: Use of a new Group: Detection of recurrent bladder cancer using a new tumor marker, urinary NMP22, in the detection of occult or quantitative assay for bladder tumor. J. Urol., part 2, 161: rapidly recurring transitional cell carcinoma of the urinary 337A,abstract 1317,1997. tract following surgical treatment. J. Urol., 156: 363,1996.
CAN URINE BOUND DIAGNOSTIC TESTS REPLACE CYSTOSCOPY IN
THE MANAGEMENT OF BLADDER CANCER? H.G. WIENER,* CH. MIAN, A. HAITEL, A. PYCHA, G. SCHATZL AND M. MARBERGER From the Deparments of Urology and Pathology, University of Vienna, Vienna, Austria
ABSTRACT
Purpose: We compare the diagnostic value of NMP22t and BTA stat$ testing, and QUANTICYTg computer assisted dual parameter image analysis to cytology and cystoscopy in patients who had symptoms suggestive of transitional cell cancer or were being followed after treatment for that disease. Materials and Methods: We prospectively evaluated voided urine and/or barbotage specimens from 291 patients a mean of 65.2 years old. All voiqed urine samples were evaluated by quick staining and standard cytology, the BTA stat 1-step qualitative assay (which detects a bladder tumor associated antigen) and the NMP22 test (which detects a nuclear mitotic apparatus protein). In addition, barbotage specimens were evaluated by QUANTICYT computer assisted dual parameter image analysis. All patients underwent subsequent cystoscopy and biopsy evaluation of any suspicious lesion. Sensitivity, specificity, and the predictive value of positive and negative results were determined in correlation with endoscopic and histological findings. Results: In 91 patients with histologically proved transitional cell carcinoma overall sensitivity was 48,57, 58, 59 and 59% for the NMP22 test, the BTA stat test, rapid staining cytology of barbotage samples, rapid staining cytology of voided urine specimens and image analysis, respectively. For histological grades 1 to 3 underlying transitional cell carcinoma sensitivity was 17,61and 90% for urinary cytology, 48,58 and 63% for the BTA stat test, and 52,45 and 50% for the NMP22 test, respectively. Specificity was 100%for cytology, 93% for image analysis, 70% for the NMP22 test and 68% for the BTA stat test. Conclusions: Immunological markers are superior to cytological evaluation and image analysis for detecting low grade transitional cell carcinoma but they have low specificity and sensitivity in grade 3 transitional cell carcinoma. Urine bound diagnostic tools cannot replace cystoscopy. KEYWORDS: carcinoma, transitional cell; bladder; cytology;cystoscopy
Cystoscopy has the pivotal role for detecting primary and recurrent transitional cell carcinoma of the bladder but even with flexible instruments it remains invasive and bothersome to the patient. Exophytic tumors are reliably diagnosed but flat transitional cell carcinoma, particularly carcinoma in situ, remains an endoscopic dilemma. Therefore, exfoliative urinary cytology of voided urine and barbotage specimens is routinely performed to complement ~ystoscopy.~ Its sensitivity is poor in low grade transitional cell carcinoma, which accounts for 60 to 70% of all primary bladder tumors: but it is unsurpassed in high grade transitional cell carcinoma, particularly carcinoma in situ.' Unfortunately the results of urinary cytology depend on the training and expertise of the cytopathologist and, as in all morphological evaluations, they have high intra-observer and interobserver ~ a r i a t i o n . ~ Recently simple methods for detecting tumor antigens in the urine have become commercially a ~ a i l a b l eRapid .~ staining techniques and quantitative computer assisted image analysis systems' have been developed in parallel in an attempt to decrease processing time and eliminate investigator dependent cytological errors. We prospectively assessed the diagnostic value of these new approaches, specifically to determine whether they eliminate the need for cystocopy.
MATERIALS AND METHODS
From January to October 1996 barbotage and/or voided urine specimens were prospectively obtained from 199 men and 92 women 17 to 90 years old (mean age 65.2) who had symptoms suggestive of bladder tumors or were being followed after treatment of transitional cell carinoma of the bladder (101 patients). To exclude from study patients with tumors persisting after transurethral resection only those proved to be tumor-free by at least 1 negative cystoscopy 3 months after transurethral resection were enrolled. In all cases voided urine samples were available and rigid or flexible cystoscopy was performed. Barbotage specimens were obtained from 200 patients by rinsing the bladder with 100 to 150 ml. saline, usually through the cystoscope. Any suspicious lesion detected was biopsied and/or removed by transurethral resection. Histopathological grading and staging were done according to the World Health Organization' and TNM7 classifications, respectively. Voided urine samples of 50 to 100 ml. were divided into 3 aliquots, coded, and evaluated by cytology with rapid and standard staining as well as by the BTA stat and NMP22 tests. In accordance with method specific protocols specimens were used fresh or were immediately frozen and stored at -2oc. Accepted for publication December 26, 1997. BTA stat test. In principle the BTA stat test is a 1-step *FLequesta for re rints: Department of Urology, University of qualitative assay that recognizes a tumor antigen reported to Vienna, Wiihringer 8iirtel 18-20, A-1090 Vienna, Austria. be a member of the complement factor H family that is tMatritech, Inc., Newton, Massachusetts. associated with transitional cell carcinoma.' Five drops of 1 SBard Dia ostics, Redmond, Washington. $Gentian gentific Software, Niawier, The Netherlands. aliquot of fresh urine or urine previously frozen at - 20C are 1876
1877
URINE BOUND DIAGNOSTIC TESTS AND CYSTOSCOPY IN BLADDER CANCER
placed on the entrance pool of a small immunochromatoTABLE1. Results of all methods tested based on voided urine graphic environment. The urine reacts with a colloidal gold samples of 291 patients, including 200 free of transitional cell carcinoma conjugated antibladder tumor associated antigen antibody. Antigen conjugate complexes are captured by another anti96 Urinarv Cvtolow 9b NMP22 96 BTA Stat bladder tumor associated antigen antibody, forming a visible Sensitivity 59 48 57 line. A positive result, as indicated by a red line, is shown Specificity 100 69 68 after 5 minutes. The plate includes an internal control zone Predictive value: POS.result 100 41 45 wntaining an immobilized reagent that captures the conjuNeg. result 84 74 78 gate independently of the presence or absence of the antigen and, thus, serves as the test quality control. NMPZZ test. Another aliquot of all voided urine was added TABLE2. Results of dl methods based on voided urine and to a stabilizing solution containing protein stabilizers, protease inhibitors and buffers, and frozen at -2OC. The nuclear hiarbotage specimens of 208 patients, including 11 7 free of tmnsitwna1 cell carcinoma mitotic apparatus protein known to correlate with transitional cell carcinomag is detected using an NMP22 test kit. It represents an enzyme-linked immunoassay in a microplate strip-well format using the monoclonal antibodies mAb 59 58 59 48 57 302-18 and mAb 302-22. According to Miyanaga et al the Sensitivity speciiicity 100 100 93 70 68 cutoff point for bladder cancer positivity is 10 units per ml.' Predictive value: QUANTZCYT analysis. Dual parameter morphometry of 100 100 76 54 56 POS. result Nen. result 78 77 76 66 70 the Feulgen stained barbotage specimens was done using the QUANTICYT computer assisted static image analyzing system to determine the individual risk of bladder cancer.' The estimation of a low, intermediate or high risk of bladder cancer is based on the combination of the deoxyribonucleic logical and morphometric evaluation, since the cellular conacid (DNA) parameter 2cDI and a morphometric shape tent varied significantly. Of all cytospin preparations of parameter (PASS). The DNA parameters 2c-deviation bladder washings 19.5% contained less than 50 transitional index (2cDI) and 5c-exceeding rate are used, as defined epithelial cells, as necessary for valid cytological examinaby B o ~ k i n g .Briefly, ~ cytospin preparations of ethanol- tion, and were discarded. Of all barbotage specimens 30% did polyethylenglycol fixed barbotage specimens are stained ac- not contain the minimum of 100 nuclei of transitional epicording to the Feulgen SchifF method using cold hydrolyza- thelial cells needed by definition for computer assisted image tion.' Nuclei are imported to the system via a microscope, analysis. Based on negative and positive cytological classifications analyzed by software, evaluated visually again and then finally calculated using lymphocytes as an internal reference. there was 100% agreement for the 2 cytological staining For statistical evaluation only specimens at high risk for methods. In addition, the results of the primary microscopic bladder cancer are defined as positive. Those in which less evaluation were in accordance with those of the independent than 100 epithelial nuclei are counted are considered insuf- review. Ten specimens each of the BTA stat and NMP22 tests ficient. were evaluated twice and results were equivalent each time Cytology. An aliquot of voided urine and barbotage speci- when NMP22 test results were reduced to a qualitative statemens was centrifuged and 2 or 3 cytospin slides were pre- ment of values greater than or less than 10 units per ml. The pared from the sediment. One cytospin preparation per spec- reproducibility of the QUANTICYT study was evaluated in h e n was air dried by gently warming the glass slide. Air 20 bladder washings. Freshly prepared cytospin specimens dried specimens were immediately stained within 40 seconds were tested at our institution and another laboratory. Docuby the rapid Hemacolor staining kit* technique, which re- mented risk factors were concordant in all cases. Table 3 shows the sensitivity of all methods tested corresults in cellular features equivalent to those of MayGriinwald Giemsa staining. Therefore, detailed cytolo@cal lated with underlying transitional cell carcinoma. We noted descriptions were available 20 minutes after the start of the major differences. Sensitivity of the cytological evaluation of Preparation and before cystoscopy. Diagnostic results were voided urine and barbotage specimens increased as transicategorized as negative, including diagnoses negative for tional cell carcinoma grade and stage increased. We observed and atypia regardlei%of atypk grade, a similar increase in sensitivity for computer assisted dual a b i a andor -cy and positive, inchding all specimens described as suspicious parameter image analysis. In contrast, for the immunological and/or positive for malignancy. The remaining specimens were " 2 2 and BTA stat tests sensitivity was 48 and 52,58 and wet fixed with a polyethylenglyoolethanol mixture. Wet 6xed Specimens underwent the Papanicohu staining protocol. logical evaluation of the PapanicouIaou stained specimens was TABLE3. Sensitivity of all tested methods s t m t i h d according to done independently on the subsequent day following the same classification scheme. According to routine quality Control histological gmde and stage of underlying transitional cell cancer in 91 patients methods 10% of the cytological specimens were reviewed inde% Quick Staining pendently by another cytologist. Specificity, sensitiGty and p s w % 96 Itwe predictive value were calculated for all tests ev+uated Urinary Barbotage Q U A " " 2 2 BTAStat wing the results of cystoscopic evaluation and histologd exCytology Cytology amination of the resected lesions as the gold standard. RESULTS
Tables 1 and 2 list overdl sensitivity, specificity, and the Predictive values of positive and negative results for all l l based on the examination of methods studied.~ techniques voided urine p d u c e d representative results. However, barhtage samples proved to be a less reliable source for cyto-
* Merck, Darmstadt, Germany.
Grade (No. pts.): 1(23*) 2 (38*) 3 (30*)
17 61 90
22(38)
55(65) W(96)
.
26(46) 55(65)
90(96)
52 45 50
48
58 63
(19')
*Values in parenthesea represent valid specimens only and reflect true results.
1878
URINE BOUND DIAGNOSTIC TESTS AND CYSTOSCOPY IN BLADDER CANCER
45,and 50 and 63% for grades 1 to 3 transitional cell carcinoma, respectively, showing that these testa depended less on resected tumor grade. Cytological evaluation had the highest specificity and predictive value of a positive result when calculated for the study population overall. Table 4 lists the clinical diagnoses in patients free of bladder malignancy in whom the immunological NMP22 and BTA stat tests indicated false-positive results. Tables 5 and 6 show sensitivity, specificity,and predictive value of a positive and negative result using all tested methods in patients followed after treatment of transitional cell carcinoma. DISCUSSION
Various predominantly monoclonal antibodies directed against tumor associated antigens are receiving increasing attention for the primary diagnosis and followup of bladder ~ a n c e r Methods .~ using immunocytochemical tests or the immunological detection of soluble antibodies in voided urine specimens are of particular interest because they are based on specimens obtained noninvasively. For example, Banks et a1 reported that a soluble version of E-cadherin, previously used as an immunohistochemical marker only, also has diagnostic potential in bladder cancer." Klein et al found that CK-20 in the voided urine of patients with bladder cancer has 82.8% sensitivity and 100.0% specificity, and it a pears to be a most interesting future diagnostic parameter.J'Kavaler et al reported that the presence of telomerase activity in the voided urine of patients with bladder cancer has 91% sensitivity.12 Unfortunately these assays are technically complicated, and require highly sophisticated laboratory expertise and equipment. They are not currently available for routine urological testing. The BTA test to detect basement membrane complexes was the fist technically simple, urine bound test for bladder cancer.13 However, sensitivity and specificity proved to be unsatisfactory in clinical use.14 The BTA stat test that we used identifies a distinctly different tumor antigen of the complement factor H family? It is designed as a rapid and simple o0ice test that provides qualitative results within minute^,'^ and it is superior to conventional cytological evaluation for detecting grade 1 transitional cell carcinoma. Unfortunately its low 68% specificity in our study limits its clinical impact. Our study population included patients with urolithiasis and chronic cystitis, as expected in clinical practice when evaluating for potential transitional cell carcinoma. False-positive results were documented in 27 and 59% of these patients, respectively. Specificity in our study is comparable to that in the double-blind multicenter study of Leyh et al, who noted 64% specificity for the BTA stat test.15
TABLE5. Sensitivity, specifiity, and predictive value of positive and negative results for all methods tested based on voided urine samples for 101 patients followed a@r tmnsurethml resection for tmnsitional cell carcinoma (87 free of recurrence) Sensitivity speeiscity M i d i v e value: POS. result Neg. result
96 Urinary Cytology 64 100
%" 2 2
100 93
% BTA Stat
14 12
29 72
8
14 86
84
TABLE6. Specificity, and predictive value of positive and negative results of all methods based on voided urine and barbotage specimens of 79 patients followed after transurethral resection of tmnsitional cell carcinoma (69 free of recurrence) %
%
%
%
%
Cytology urinaryBarbotage Cytology QUANTICYT NMp22 BTAStat Sensitivity Specificity
80 100
80 100
P;edictivi value: Pos. result 100 100 Neg. result 97 97 * Valid specimens only and true results.
60 (67)* 87 (93):
68
40 68
40 94
8 85
15 89
20
Sarosdy et al reported 72% specificity in the voided urine of patients with benign genitourinary abnormalities.16 The same antigen may also be analyzed quantitatively using the BTA-TRAK* test, a modified enzyme immunoassay. Ishak et al reported 74% specificity with this method, which was slightly higher than with the BTA stat test,17 as described. They also reported 85% specificity for the original bladder tumor associated antigen test, which has not been reproduced by others. Sarosdy et al reported greater than 90% specificity for the BTA stat test in healthy volunteers only.16 However, as stated by Layfield, specificity based on healthy volunteers does not reflect the regular clinical situation." Our calculation of specificity was based on concurrent cystoscopic findings. Theoretically tumor antigens may appear in the urine before the tumor becomes visible on endoscopy, similar to when urinary cytology is positive months before cystoscopy is positive.'. Therefore, the significance of a positive BTA stat test with otherwise negative findings must still be considered unknown. Leyh16 and Sarosdy" et a1 reported 83 to 95%sensitivity of the BTA stat test for grade 3 transitional cell carcinoma. We did not reproduce these data in our study. This discrepancy may result from the low numbers of investigated patients per study, since both included less than 50 with grade 3 transitional cell carcinoma. Calculation of the mean of documented values results in 80% sensitivity of the BTA stat test for grade 3 transitional cell carcinoma. Compared with voided urine cytology the BTA stat test appears to be less sensitive TABLE4. Clinical features of patients free of malignant bladder for this aggressive tumor. disease and false-positive results of immunological markers In our study specificity and sensitivity for detecting the No. False-pos. nuclear matrix protein nuclear mitotic apparatus protein Results (%) WMl"MP2 test) in voided urine were equivalent or similar to No. F'ta. those of the BTA stat test. The data compare favorably with BTAStat "22 the findings of Rodriguez-Villanueva et al, who reported Followup aRer tranurethral reseetion, 87 24 (38) 24 (38) overall 53.8% specificity for the NMP22 test.'' In contrast, neg. C Y ~ ~ P Y Cystitis 22 13 (20) 9 (14) Miyanaga et al noted 80.9% sensitivity and 64.3% specificiUpper tract urnlithiasis 15 4 (6) 4 (6) ty.' The discrepancy may again result from the number of Benign lesions of lower urinary tract 25 12 (19) 10 (16) voided urine specimens in the studies. Our series included and followup (benign prostatic hypernearly twice the specimens evaluated by Miyanaga et al. plasia, hemangioma, nephrogenic adFurthermore, differences were evident for the reported perenoma and so forth) 11 Malignancies of nonurological sites and centage of false-positive test results, especially in cases of followup cystitis. Miyanaga et a1 observed 19 of 19 false-positive reMicrohematuria, no evidence of transi40 sults (100%) in cystitis, whereas we noted only 41%. In both tional cell Ca - Totals
200
64 (100)
63 (100)
* Bard Diagnostics, Redmond, Washington.
URINE BOUND DIAGNOSTIC TESTS AND CYSTOSCOPY IN BLADDER CANCER
I
1879
CONCLUSIONS studies technical errors may be excluded because standards and controls are included in the test kits. Urinary erythroThe BTA stat and " 2 2 tests, and QUANnCYT comeytes have only marginal impact on the results? and so puter assisted dual parameter image analysis improve the microhematuria does not explain the differences. Variation detection of transitional cell carcinoma, particularly grade 1 in the storage time of the voided urine samples may be a tumors, compared to cytological evaluation. The value of the possible explanation. In our study the storage time was lim- immunological markers is limited by low sensitivity in grade ited to a maximum of 3 weeks. At least for the NMP22 test 3 transitional cell carcinoma as well as low specificity. None storage conditions appear to be of high importance. Voided of the urine bound diagnostic tools that we investigated reurine specimens must be stored in special vials containing a places cystoscopy. stabilizer, which complicates clinical management. However, The BTA stat and some NMP22 test kits were provided by the test is designed as an enzyme-linked immunoassay that allows quantitative evaluation of nuclear matrix protein 22 Bard Diagnostics, Redmond, Washingtan, and Fa. Bender, levels. Although not the aim of the present study, the quan- Vienna, Austria. The QUANTICYT system was provided titative NMP22" and BTA-TRAK17 tests may help to eval- by Gentian Scientific Software, Niawier, The Netherlands. uate the individual risk of tumor recurrence. Soloway et al Peter van Stratum and Pim Peelen, Urological Development previously proved the usefulness of the NMP22 test for de- and Diagnostics Laboratory, Wijchen, The Netherlands, performed double analyses of bladder washings, and 0. Haitel tecting persistent or rapidly recurring bladder cancer?' performed the statistical analysis. Whereas the value of computerized image analysis of voided urine specimens remains a matter of controversy: the determination of nuclear total DNA content in barbotage REFERENCES samples has been shown to improve diagnostic power, par1. Wiener, H.G., Vooijs, G. P. and van't Hof-Gmotenboer, B.: Acticularly in recurrent superficial bladder cancer. Q U A " curacy of urinary cytology in the diagnosis of primary and software combines 2 morphometric parameters, a total DNA recurrent bladder cancer. Acta Cytol., 37: 163,1993. and nuclear shape descriptor! Specificity and grade 3 trans2. Raghavan, D.,Shipley, W. U., Garnick, M. B., Russell, P. J. and itional cell carcinoma related sensitivity of dual parameter Richie, J. P.: Biology and management of bladder cancer. New image analysis surpassed those of the immunological markEngl. J. Med., 322 1129,1990. 3. Ooms, E. C. M., Kurver, P. H. J., Veldhiuzen, R. W., Alons, C. L. ers. Depending on whether insufficient barbotage specimens and Boon, M. E.: Morphometric grading of bladder tumors in were excluded we noted sensitivity similar to or higher than comparison with histologic gradmg by pathologists. Hum. that of conventional cytologicalevaluation but 7% lower specPath., 144: 140, 1983. ificity. The test decreases investigator dependent variation 4. Fradet, Y. and Cordon-Cardo, C.: Critical appraisal of tumor and, thus, complements clinically important rapid cytological markers in bladder cancer. Sem. Urol., 11: 145,1993. evaluation at the cost of slightly lower specificity. However, 5. van der Poel, H. G., Witjes, J. A., van Stratum, P., Boon, M. E., this test is more time-consuming and barbotage samples are Debruyne, F. M. and Schalken, J. A.: Quanticyt: karyometric more invasive than voided urine samples. The value of all analysis ofbladder washings for patients with superficial bladder cancer. Urology, 48.357, 1996. bladder washing, cell bound diagnostic tools depends on the 6. Mostofi, F.K,Solbin, L. H. and Torloni, H.: Histological Typing quality and amount of cells in the specimen. Of the speciof Urinary Bladder Tumors.Geneva: World Health Organizamens included in our study 30% proved insufficient for tion, 1973. QUANTICYT analysis, while van der Poel et al reported 8% 7. TMN Classification of Malignant Tumors, 4th ed. Union Interinsufficient specimens after obtaining only 25 ml. of bladder nationale Contre le Cancer. Geneva: International Union washings through a catheter or cystoscope.' Most of our Against Cancer, 1987. barbotage specimens were obtained by irrigation through the 8. Kinders. R., Jones, T.. Root, R., Murchison. H., Bruce. C.. port of flexible cystoscopes, which is extremely thin. NeverWilliams, L. and Hass; G. M.: H&an bladder.tumor antigen $ a member of the RCA (regulators of complement activation) theless, flexible cystocopy is the standard technique, at least gene family. J. Urol., part 2,157: 284 abstract 110, 1997. in men. Laboratory bound procedures followed the same pro9. Miyanaga, N., h a , H., Ishikawa, S., Ohtani, M., Noguchi, R., tocol. Kawai, K, Koiso, K, Kobayashi, M., Koyama, A. and The clinical usefulness of a diagnostic test also depends on Takahashi, T.: Clinical evaluation of nuclear matrix protein procedure time, personnel and material needed (table 7). ( " 2 2 ) in urine as a novel marker for urothelial cancer. Eur. With overall procedure times of 5 and 25 minutes, respecUrol., 31: 163, 1997. tively, the BTA stat test and rapid staining cytology are the 10. Banks,R. E., Porter, W. H.,Whelan, P., Smith, P. H. and Selby, most rapid. Calculating expenses for staff and materials for P. J.: Soluble forms of the adhesion E-cadherin molecule in 1,000 specimens per year and amortization times of 3 and 5 urine. J. Clin. Path., 48: 179,1995. Years, respectively, the BTA stat test and rapid staining 11. Klein, A., Zemer, R., Buchumensky, V., Klaper, R., Nissenkom, I. and Saba, K: Detection of bladder carcinoma: a urine test, urinary cytology are definitively the most cost-effectivemethbased on cytokeratin expression. J. Urol., part 2, 157: 3394 ods. Cytological evaluation adds information on all corpuscuabstract 1326,1997. 1ar elements found within the specimens but it requires a 12. Kavaler, E., Shu, W.-P., Chang, Y., Droller, M. J. and Liu, highly trained and qualified staff. In contrast, the qualitative B. C.-S.: Detection of human bladder cancer cells in voided immunochromatographic BTA stat assay is a simple office urine samples by assaying for the presence of telomerase acprocedure that may be performed by the urologist. tivity. J. Urol., part 2, 157: 338A,abstract 1321,1997. 13. Sarosdy, M.F.,deVere White, R. W., Soloway, M. S., Sheinfeld, J., Hudson, M. A, Schellhammer, P. F., Jarowenko, M. V., TABLE 7. Amortized costs in United States dollars and evalwtwn
time for all methods tested BTA
5
s i n sJspeeimen
Quick Staining
Stat
N m 2
15 5
180
54
QUmW
41 210
Urinary cytology
42 25
Adams, G. and Blumenstein, B. A,: Results of a multicenter trial using the BTA test to monitor for and diagnose recurrent bladder cancer. J. Urol., 161:379,1995. 14. D'Hallewin, M. A. and Baert, L.: Initial evaluation of the bladder tumor antigen test in superficial bladder cancer. J. Urol., 165: 475,1996. 15. Leyh, H., Marberger, M., Pagano, F., Bassi, P., Sternberg, C. N., Pansadom, V., Conort, P., Bmn-Gibod, L. and Thoeke, K. R: Results of a European multicenter trial comparing the BTAstat test to urine cytology in patients suspected of having bladder cancer. J. Urol., part 2, 151: 3374 abstract 1316, 1997.
1880
URINE BOUND DIAGNOSTIC TESTS AND CYSTOSCOW IN BLADDER CANCER
16. samedy. M. F., Hudson, M. A, Ellis, W.J., Soloway, M. S., 18. Layfield, L.:Editorial comment. J. Urol., 15k 383, 1995. deVere White, R. W.,Sheinfield. J., Jarowenk, M. V., 19. Rodriguez-Villanueva,J., Dinney, C. P. N., Grossman, H. B. and Fritsche, H. A.: Evaluation of the NMP22 immunoassay in the Schellhammer, P. F., Messing, E. M., Schervish, E. W.,Patel, detedion of transitional cell carcinoma (TCC) of the urinary J. V., Chodak, G. W.,Lamm, D. L.,Johnson, R. D., Henderson, M., Adams, G., Blumenstein, B. A, Thoelke, K. R., Brunelle, tract. J. Urol., part 2,151: 3364 abstract 1314,1997. S. L.,Pfalzgraf, R. D. and Murchimn, H. A.: Detection of 20. Soloway, M. S., Briggman, J. V., Carpinito, G. A., Chodak, G. W., Church, P. A.,Lamm,D.L.,Lange, P., Messing, E., Pasciak, recurrent bladder cancer using a new one-step test for bladder tumor antigen. J. Urol., part 2,161: 3374 abstract 1318,1997. R. M., Reservitz, G. B., Rukstalis, D. B., Sarosdy, M. F., 17. Ishak, L.M.. Enfield, D. L.,Sarosdy, M. F. and the Multicenter Stadler, W.M., Thiel, R. P. and Hayden, C. L.: Use of a new Group: Detection of recurrent bladder cancer using a new tumor marker, urinary NMP22, in the detection of occult or quantitative assay for bladder tumor. J. Urol., part 2, 161: rapidly recurring transitional cell carcinoma of the urinary 337A,abstract 1317,1997. tract following surgical treatment. J. Urol., 156: 363,1996.