- Email: [email protected]
Bladder tumor antigen assay as compared to voided urine cytology in the diagnosis of bladder cancer
Clinica Chimica Acta 305 (2001) 47–53 www.elsevier.com / locate / clinchim
Bladder tumor antigen assay as compared to voided urine cytology in the diagnosis of bladder cancer a, b a a Gabriella Priolo *, Paolo Gontero , Giovanna Martinasso , Giulio Mengozzi , Andrea Formiconi b , Gianni Pelucelli b , Andrea Zitella b , Giovanni Casetta b , Laura Viberti c , Giuseppe Aimo a , Alessandro Tizzani b a
Laboratory of Clinical Chemistry, San Giovanni Battista Hospital, Corso Bramante 88, 10126 Turin, Italy b Department of Urology, University of Turin, Turin, Italy c Institute of Pathology, University of Turin, Turin, Italy Received 31 May 2000; received in revised form 7 November 2000; accepted 20 November 2000
Abstract Background: The study was aimed at comparing the diagnostic accuracy of the quantitative bladder tumor antigen (BTA) TRAKE immunoassay with exfoliative urine cytology in the detection of primary and recurrent bladder cancer. Methods: The analysis was carried out on 194 high risk patients undergoing a diagnostic cystoscopy, 279 patients with previous history of transitional cell carcinoma awaiting a follow-up cystoscopy, and 45 healthy controls. Urine cytology was performed by a skilled cytopathologist on three consecutive samples. Results: BTA TRAK values resulted significantly higher in tumor positive cases than in absence of bladder tumor for both groups of patients. Non neoplastic urothelial diseases as well as the absence of mucosal abnormalities were associated with a marked increase in BTA TRAK levels with respect to the control group. Overall sensitivity and specificity was 63 and 63% for BTA TRAK (cut-off 34 U / ml), and 68.3 and 73.4% for urine cytology, respectively. The diagnostic advantage of urine cytology was maintained when patients were stratified by tumor grade. Conclusions: The clinical performance of the BTA TRAK in the detection of primary or recurrent bladder cancer is acceptable and reproducible as shown by similar results with previous reports, although urine cytology performed on three samples showed the highest sensitivity and specificity. 2001 Elsevier Science B.V. All rights reserved. Keywords: BTA TRAK; Urine cytology; Bladder cancer
1. Introduction Diagnostic procedures in patients with bladder
Abbreviations: BTA, bladder tumor antigen; VUC, voided urine cytology; TCC, transitional cell carcinoma; PPV, positive predictive value; NPV, negative predictive value *Corresponding author. Tel.: 139-11-633-6765; fax: 139-11676-052. E-mail address: [email protected] (G. Priolo).
cancer symptoms include urine cytology, cystoscopy with biopsy, and excretory urography, but cystoscopy still represents the gold standard technique for the detection of both primary and recurrent tumors [1,2]. The majority of transitional cell carcinomas (TCC) of the bladder are superficial and thus amenable to conservative treatment by local endoscopic resection [3]. Yet, recurrence rate is very high, ranging from 40 to 85% according to different studies [4,5]. Because of the risk of recurrence, a
0009-8981 / 01 / $ – see front matter 2001 Elsevier Science B.V. All rights reserved. PII: S0009-8981( 00 )00416-2
48
G. Priolo et al. / Clinica Chimica Acta 305 (2001) 47 – 53
regular follow-up is required and patients must be examined at frequent intervals after treatment. The high demand for an endoscopic procedure, which is relatively invasive, has prompted the search for objective noninvasive methods that could reduce the need for cystoscopy in the surveillance protocols. Voided urine cytology (VUC) has played a significant role as a diagnostic tool in the assessment and follow-up of patients with bladder cancer [6], although a standardization of methods is lacking. Moreover, results of the test are highly dependent upon the skill of the cytopathologist, the correct preservation of urine and the number of exfoliated cells in the sample [7]. Despite 80% sensitivity for high grade tumors, urine cytology is a poor indicator of low grade tumors [8]. A number of bladder cancer tumor-related molecular markers [9], such as DNA ploidy, p53 mutations [10], and microsatellite DNA, are still difficult and expensive to perform and require trained personnel, whereas others, such as telomerase [11,12] have shown to be of limited value [13]. On the other hand, some non invasive urine analysis, including home hematuria screening [14], Nuclear Matrix Protein 22 (NMP 22) [15], and BTA [16], have become available for bladder cancer testing [17]. The BTA TRAKE assay is a quantitative enzyme immunoassay to detect a bladder tumor-associated antigen in the urine. This antigen has been recognized as a human complement factor H related protein (hCFHrp) and inhibits the complement pathway to cause cytolysis in cells with a resulting selective advantage for the tumor [18]. BTA TRAK sensitivity has been found to be superior to urine cytology and to the original qualitative BTA test [19]. We compared the sensitivity, the specificity and the diagnostic accuracy of the BTA TRAK test and that of VUC in the detection of primary or recurrent TCC of the bladder in a wide population of high risk patients.
2. Patients and methods Four hundred and seventy-three patients admitted to three Institutions in our hospital were enrolled
over a period of 1 year. These included 194 patients (25 females and 169 males, mean age 66.0 years; group 1) who had been scheduled for an urgent diagnostic cystoscopy in view of a high suspicion of bladder cancer (macroscopic painless hematuria and / or urographic / sonographic pattern suggestive of bladder tumor). The remaining 279, with a previous history of TCC, were undergoing a routine follow-up cystoscopy (42 females and 237 males, mean age 68.2; group 2). These latter had been treated with intravesical chemo- or immunotherapy following standard international protocols, mainly either low or full dose of Bacillus of Calmette–Guerin (BCG). Urine samples were collected by all patients on three consecutive days before the endoscopic procedure and sent for cytology. An additional urine sample was obtained and frozen below 2208C until BTA TRAK testing. Since in our experience urine samples may be stored at 48C for at least 24 h without loss of the antigen, specimens were collected and immediately put on ice or stored in the refrigerator at 48C before sending them to the laboratory. The stability of the marker in urine stored at 2208C is longer than 4–6 weeks. All urine investigations were carried out on an additional sample of 45 healthy volunteers (11 females and 34 males, mean age 58.7 years) who served as controls. Cystoscopy was followed by biopsy of the bladder mucosa in all cases even in the absence of macroscopic evidence of a tumor. Neoplastic lesions were staged according to the TNM classification and graded according to the WHO system. VUC was performed in all patients according to the standard procedures used in the Laboratory of the Institute of Pathology, University of Turin. All three specimens from each case were evaluated by a single experienced cytologist who was unaware of the results of BTA TRAK. For statistical purposes the patient was considered positive for tumor when at least one out of the three specimens showed cytological features of tumor. All cases showing features described as ‘cell atypia’ were categorized as dubious or suspicious results. The BTA TRAKE assay (DiaSorin, Saluggia, Italy) is an immunoenzymatic assay (IEMA) using monoclonal antibodies to bind specifically to bladder tumor antigen in urine. The lowest concentration of
G. Priolo et al. / Clinica Chimica Acta 305 (2001) 47 – 53
BTA detectable was estimated to be 0.75 U / ml. Assay reproducibility tests yielded ,5 and ,10% variation in intra- and inter-series runs, respectively (mean value: 30 U / ml). For this method no interference on results was reported by the manufacturer for a number of possible urine constituents such as hemoglobin (up to 100 mg / dl), red blood cells (up to 104 cells / ml) or by therapeutic agents, such as BCG (up to 20 mg / dl or 53106 CFU / ml), Mytomicin C (up to 10 mg / dl), and ampicillin (up to 600 mg / dl). Findings with hemoglobin and red blood cells were confirmed also by tests carried out in our center through the simulation of various degrees of hematuria. All procedures in this study were performed in accordance with the ethical standards as formulated in the Helsinki Declaration of 1975 (revision of 1996).
2.1. Statistical analysis As urinary BTA data were regarded to be not normally distributed, nonparametric comparisons of central tendency were employed for analysis. The distributions of urinary BTA were compared using the Mann–Whitney U test. A level of P,0.01 was accepted as statistically significant. Sensitivity was calculated as true positive test results / all patients with disease; specificity was calculated as true negative test results / all patients without disease. The receiver operating characteristic (ROC) curves were plotted and the areas under the curves were obtained by the trapezoidal rule [20]. The choice of the cut-offs was based on the best combination of sensitivity and specificity as derived by the analysis of their curve intersection. All data were expressed as median values and range.
3. Results One hundred and forty-seven out of the 194 cases (76%) with high suspicion of bladder cancer (group 1) were found to have a primary bladder TCC, the remaining 47 patients (24%) having a histological response negative for bladder cancer. Tumor recurrence was confirmed in 187 out of the 279 patients (67%) with previous history of bladder cancer
49
(group 2) while 92 (33%) were histologically recurrence free. In Table 1, patients from both groups are listed according to the different stage and grade of the neoplasia as well as according to the histological features of all non-neoplastic conditions. BTA TRAK median values and results from urine cytology (categorized as positive, negative and dubious, respectively) for each group of patients are reported in Table 2. Histology reports with no evidence of neoplasia have been grouped into two main categories named respectively ‘normal mucosa’ when no pathological findings were evidenced by the pathologist and ‘other pathologies’ for all benign conditions of the urothelium. BTA TRAK values were significantly higher in tumor positive cases than in the absence of bladder tumor for both group 1 (P50.0002 and P50.04 with respect to tumor negaTable 1 Patients referred for with a high suspicion of bladder cancer (group 1) and patients during follow-up for bladder TCC (group 2) are categorized according to histology results a n (%) Group 1 Tumor positive Stage Ta T1 T$2 Tx Total Grade 1 2 3 Total Tumor negative
86 (58.5) 29 (19.7) 29 (19.7) 3 (2.1) 147
42 (28.6) 55 (37.4) 50 (34) 147
Group 2
95 (50.8) 34 (18.2) 43 (23) 15 (8) 187
45 (24.1) 70 (37.4) 72 (38.5) 187
n
Normal mucosa Mucosal inflammation Cystic cystitis Granulomatous cystitis Squamous metaplasia Benign prostatic hyperplasia
33 10 3 0 0 1
69 20 1 1 1 0
Total
47
92
a
Tumor positive cases are classified following the different stages (TNM classification) and grades (WHO system) of the neoplasia. Tumor negative cases are grouped in different categories of non-neoplastic lesions.
G. Priolo et al. / Clinica Chimica Acta 305 (2001) 47 – 53
50
Table 2 Median values of BTA TRAK (range) are reported for each group of patients (group 1, patients with high suspicion of bladder cancer; group 2, patients in follow-up for bladder cancer; and group 3, healthy volunteers)a Patient group
Histology
BTA TRAK (U / ml) median (range)
Pb
Cytology positive
Cytology negative
Cytology dubious
n
Group 1
Tumor positive
80.1 (0.1–3000)
0.0001
87
42
18
147
Tumor negative Normal mucosa Other pathologies
21.2 (0.4–3000) 19.8 (0.4–1053) 25.0 (4.3–3000)
NS c
9 7 2
35 23 12
3 3 0
47 33 14
Tumor positive
78.9 (0.7–3000)
0.0001
95
64
28
187
Tumor negative Normal mucosa Other pathologies
25.1 (0.4–3000) 27.0 (0.4–3000) 24.5 (1.6–1005)
NS
11 8 3
67 50 17
14 11 3
92 69 23
0
44
1
45
Group 2
Group 3
9.2 (5.0–59.1)
a
Patients were classified as tumor positive when both cystoscopy and histology were positive. All non-neoplastic lesions (5tumor negative patients) have been grouped into two main categories: normal mucosa and all other benign conditions. b Mann–Whitney U-test. c Not significant.
tive patients with normal mucosa and other pathologies, respectively) and group 2 patients (P5 0.0001 and P50.0038, respectively). Non neoplastic urothelial diseases showed a trend towards a higher expression of the bladder tumor antigen although the difference was not statistically significant either in group 1 or in group 2. BTA TRAK concentrations in urine from patients with no mucosal abnormality (in both groups) were markedly increased compared to the baseline values from the control group. The majority of dubious responses at urine cytolo-
gy were from patients who had tumor positive results during the cystoscopical investigation. However, as shown in Table 2, those few cases with no tumor and ‘dubious’ cytological results belonged mainly to the subgroup with histology finding of ‘normal mucosa’. Sensitivity and specificity for BTA TRAK were evaluated assuming as control cases for group 1 the tumor negative patients and for group 2 the recurrence free patients (Table 3). With the method employed to calculate the cut-off threshold, a comparable BTA TRAK value was obtained for both
Table 3 Sensitivity, specificity, positive predictive value, negative predictive value and diagnostic accuracy for both BTA TRAK and urine cytology are reported either separately for group 1 and 2 or combining both groups a Patient group
Test
Cut-off (U / ml)
Sensitivity (%)
Specificity (%)
PPV
NPV
Diagnostic accuracy
1
BTA TRAK Cytology
35 Dubious as positive Dubious as negative
64 71.4 59.2
64 74.4 80.4
84.5 89.7 90.6
35.7 45.4 38.8
63.4 72.1 64.4
2
BTA TRAK Cytology
33 Dubious as positive Dubious as negative
63 65.8 50.8
63 72.8 88
77.1 83.1 89.6
45.2 51.1 46.8
62.7 68.1 63.1
112
BTA TRAK Cytology
34 Dubious as positive Dubious as negative
63 68.3 54.5
63 73.4 85.6
80.5 86 90
41.5 49 43.9
63 69.8 63.6
a
PPV, positive predictive value; NPV, negative predictive value.
G. Priolo et al. / Clinica Chimica Acta 305 (2001) 47 – 53
groups of patients (35 U / ml for group 1 and 33 for group 2), yielding 64 and 63% sensitivity and specificity as well as 63.4 and 62.7% accuracy, respectively. The same parameters were therefore determined also for all 473 patients (group 1 and group 2 combined, cut-off 34 U / ml), using as control cases all tumor free patients from both groups. ROC curve analysis showed a good diagnostic profile for BTA TRAK test. Sensitivity of urine cytology was determined for each group assuming all dubious results either as positive cases or as negative cases. In Table 3, sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and diagnostic accuracy are reported for each group and for both group combined. According to the grade of the bladder TCC, BTA TRAK sensitivity was 46, 58 and 81.1% for G1, G2 and G3 tumors, respectively, when calculated on the whole sample of patients (group 1 and group 2 combined). In the same analysis cytology resulted in 52.9, 66.1 and 84.4% sensitivity, when dubious cases were considered as positive, and in 41.4, 50 and 69.7% sensitivity, when dubious cases were considered as negative.
4. Discussion In the present report the diagnostic ability of BTA TRAK to detect a bladder TCC has been assessed in two clinical settings: history suggestive of bladder cancer, where a noninvasive diagnostic procedure may lead to an immediate and accurate cystoscopy, and in the follow-up of superficial bladder cancer patients, when the ability of an urine test to detect a tumor recurrence could be an useful tool for the selection of cases to be submitted for control cystoscopies. In this scenario urine cytology still plays an important role although it has a poor sensitivity to pick up low grade TCC [21]. Following the successful introduction of the BTA stat qualitative assay in the diagnosis of bladder cancer, a new quantitative assay detecting the same bladder antigen has been recently developed. Early results with the BTA TRAK assay have so far been promising. The
51
quantitative determination of the Bladder Tumor Antigen in the urine, named as BTA TRAK assay test, has shown higher sensitivity than the qualitative BTA stat test in the detection of a bladder cancer [22]. Conversely, urine cytology sensitivity is much lower in the same authors experience, never exceeding 44% [12,19,22]. The results from our large series seem not to be in contrast with those of two previous multicenter trials. Ellis [19] reported an overall sensitivity of the quantitative assay for the bladder tumor antigen of 72% using a cut-off of 14 U / ml and Thomas [23] found a BTA TRAK ability to detect a bladder cancer of 66%. In the present study the overall BTA TRAK sensitivity was 63%. A selection bias may be advocated to explain this lower value. The enrollment of patients with a high suspicion of bladder cancer (group 1) resulted in a higher positivity rate of bladder TCC following diagnostic cystoscopy. Similarly, the recurrence rate in follow-up patients (group 2) was very high (67%), pointing out that the diagnostic marker was assessed in a high risk bladder cancer group. A consistent part of the tumor free patients from both groups showed histological features of a benign pathology of the bladder. It should be assumed that other coexisting situations, like hematuria, among those with histologically normal urothelium could have interfered with the immunoassay. As a result, mean BTA TRAK values were much higher in tumor free patients than in the control group. This latter group was not included in our calculations as the investigation is not routinely addressed to asymptomatic patients. We believe our sample reflects more closely the population to which the test is applied. A value of 14 U / ml has been employed as a standard cut-off value in several reports on BTA TRAK [19,22,23]. This value is far lower the mean BTA concentration in tumor negative patients from our two groups but it is comparable with the mean value of BTA in the control group. Assuming the literature cut-off of 14 U / ml would result only in a mild increase in the sensitivity (87%) with a dramatic decrease of the specificity to 32%. Our cut-off was determined as the bladder tumor antigen concentration allowing the same sensitivity and specificity. The cut-off differences between the two groups was not significant and this finding allowed
52
G. Priolo et al. / Clinica Chimica Acta 305 (2001) 47 – 53
grouping all patients for the determination of a single sensitivity value. According to our data, urine cytology had the highest sensitivity and specificity power, particularly when the dubious results were assumed to be positive for malignant transitional cells. Cytological analysis was carried out on three consecutive samples and all the specimens were read by an expert cytopathologist. This could have contributed to the striking difference observed in comparison with other reports where urine cytology was performed on one single sample and assessed by several technicians [12,24]. Urine cytology was superior to BTA TRAK in the detection of primary or recurrent bladder cancer. The diagnostic advantage was maintained also when patients were stratified by tumor grade. Our study confirms the diagnostic utility of the BTA TRAK in the detection of a primary or a recurrent bladder cancer. An overall 70% sensitivity for urine cytology is attractive but probably highly dependent on an investigation carried out on three samples, which is time consuming and quite expensive, and on the availability of an expert cytopathologist, a situation difficult to maintain in many institutions. The diagnostic accuracy of BTA TRAK, although a bit lower than urine cytology, is acceptable and easily reproducible in different Institutions as shown by the substantial similar results with previous reports. Unfortunately, none of the two urine tests seems able to modify the diagnostic role of cystoscopy so far.
References [1] Lessing JA. Bladder cancer: early diagnosis and evaluation of biological potential. A review of newer methods. J Urol 1978;120:1–55. [2] Raghavan D, Shipley WU, Garnick MB, Russell PJ, Richie JP. Biology and management of bladder cancer. New Engl J Med 1990;322:1129–38. [3] Mostofi FK, Sobin LH, Torloni H. Histological typing of urinary bladder tumors. WHO International Histological Classification of Tumors, Geneva: World Health Organization, 1973, pp. 9–35. [4] Heney NM, Ahmed S, Flanagan MJ, Frable W, Corder MP, Haferman MD, Hawkins IR. Superficial bladder cancer: progression and recurrence. J Urol 1983;135:1083–6.
[5] Thrasher JB, Crawford ED. Current management of invasive and metastatic transitional cell carcinoma of the bladder. J Urol 1993;149:957–72. [6] Zein T, Wajsman Z, Englander LS, Gamarra M, Lopez C, Huben RP, Pontes J. Evaluation of bladder washings and urine cytology in the diagnosis of bladder cancer and its correlation with selected biopses of the bladder mucosa. J Urol 1984;132:670–8. [7] Murphy WM, Soloway MS, Jukkola AF, Crabtree WN, Ford KS. Urinary cytology and bladder cancer. The cellular features of transitional cell neoplasms. Cancer 1984;53:1555–65. [8] Gregoire M, Fradet Y, Meyer F. Diagnostic accuracy of urine cytology and deoxyribonucleic acid flow cytometry and cytology on bladder washings during follow-up for bladder tumors. J Urol 1997;157:1660–4. [9] Burchardt M, Burchardt T, Shabsigh A, De La Taille A, Benson MC, Sawczuk I. Current concepts in biomarker technology for bladder cancers. Clin Chem 2000;46:595– 605. [10] Herr HW, Bajorin DF, Scher HI, Cordon-Cardo C, Reuter VE. Can p53 help select patients with invasive bladder cancer for bladder preservation? J Urol 1999;161:20–2. [11] Rahat MA, Lahat N, Gazawi H, Resnick MB, Sova Y, Ben-Ari G et al. Telomerase activity in patients with transitional cell carcinoma: a preliminary study. Cancer 1999;85:919–24. [12] Ramakumar S, Bhuiyan J, Besse JA, Roberts SG, Wollan PC, Blute ML, O’Kane DJ. Comparison of screening methods in the detection of bladder cancer. J Urol 1999;161:388–94. [13] Arai Y, Yajima T, Yagihashi A, Kobayashi D, Kameshima H, Sasaki M et al. Limitations of urinary telomerase activity measurement in urothelial cancer. Clin Chim Acta 2000;296:35–44. [14] Messing EM, Young TB, Hunt VB, Newton MA, Bram LL, Vaillancourt A et al. Hematuria home screening: repeat testing results. J Urol 1995;154:57–61. [15] Soloway MS, Briggman JV, Carpinito GA, Chodak GW, Church PA, Lamm DL et al. Use of a new tumor marker NMP22 in the detection of occult or rapidly recurring transitional cell carcinoma of the urinary tract following surgical treatment. J Urol 1996;156:363–7. [16] Sarosdy MF, deVere White RW, Soloway MS, Sheinfeld J, Hudson MA, Schellhammer PF et al. Results of a multicenter trial using the BTA test to monitor for and diagnose recurrent bladder cancer. J Urol 1995;154:379–83. ¨ [17] Lokeshwar VB, Obek C, Soloway MS, Block NL. Tumorassociated hyaluronic acid: a new sensitive and specific urine marker for bladder cancer. Cancer Res 1997;57:773–7. [18] Kinders R, Jones T, Root R, Murchison H, Bruce C, Williams L, Hass GM. Human bladder tumor antigen is a member of the RCA (regulators of complement activation) gene family. J Urol 1997;157(suppl):28–36. [19] the Multi Center Study Group, Ellis WJ, Blumenstein BA, Ishak LM, Enfield DL. Clinical evaluation of the BTA TRAK assay and comparison to voided urine cytology and the Bard BTA test in patients with recurrent bladder tumors. Urology 1997;50:882–7.
G. Priolo et al. / Clinica Chimica Acta 305 (2001) 47 – 53 [20] Hanley JA, McNeil BJ. A method of comparing the areas under receiver operating characteristic curves derived from the same cases. Radiology 1983;148:839–43. [21] Murphy WM, Crabtree WN, Jukkola AF. The diagnostic value of urine versus bladder washing in patients with bladder cancer. J Urol 1981;126:320–2. [22] Irani J, Desgrandchamp F, Millet C, Toubert ME, Bon D, Aubert J et al. BTA stat and BTA trak: A comparative evaluation of urine testing for the diagnosis of transitional cell carcinoma of the bladder. Eur Urol 1999;35:89–92.
53
[23] Thomas L, Leyh H, Marberger M, Bombardieri F, Bassi P, Pagano F et al. Multicenter trial of the quantitative BTA TRAK assay in the detection of bladder cancer. Clin Chem 1999;45:472–7. [24] Del Nero A, Esposito N, Curro’ A, Blasoni A, Montanari E, Mangiarotti B et al. Evaluation of urinary level of NMP 22 as a diagnostic marker for stage pTa-pT1 bladder cancer; comparison with urinary cytology and BTA test. Eur Urol 1999;35:93–7.
Bladder tumor antigen assay as compared to voided urine cytology in the diagnosis of bladder cancer a, b a a Gabriella Priolo *, Paolo Gontero , Giovanna Martinasso , Giulio Mengozzi , Andrea Formiconi b , Gianni Pelucelli b , Andrea Zitella b , Giovanni Casetta b , Laura Viberti c , Giuseppe Aimo a , Alessandro Tizzani b a
Laboratory of Clinical Chemistry, San Giovanni Battista Hospital, Corso Bramante 88, 10126 Turin, Italy b Department of Urology, University of Turin, Turin, Italy c Institute of Pathology, University of Turin, Turin, Italy Received 31 May 2000; received in revised form 7 November 2000; accepted 20 November 2000
Abstract Background: The study was aimed at comparing the diagnostic accuracy of the quantitative bladder tumor antigen (BTA) TRAKE immunoassay with exfoliative urine cytology in the detection of primary and recurrent bladder cancer. Methods: The analysis was carried out on 194 high risk patients undergoing a diagnostic cystoscopy, 279 patients with previous history of transitional cell carcinoma awaiting a follow-up cystoscopy, and 45 healthy controls. Urine cytology was performed by a skilled cytopathologist on three consecutive samples. Results: BTA TRAK values resulted significantly higher in tumor positive cases than in absence of bladder tumor for both groups of patients. Non neoplastic urothelial diseases as well as the absence of mucosal abnormalities were associated with a marked increase in BTA TRAK levels with respect to the control group. Overall sensitivity and specificity was 63 and 63% for BTA TRAK (cut-off 34 U / ml), and 68.3 and 73.4% for urine cytology, respectively. The diagnostic advantage of urine cytology was maintained when patients were stratified by tumor grade. Conclusions: The clinical performance of the BTA TRAK in the detection of primary or recurrent bladder cancer is acceptable and reproducible as shown by similar results with previous reports, although urine cytology performed on three samples showed the highest sensitivity and specificity. 2001 Elsevier Science B.V. All rights reserved. Keywords: BTA TRAK; Urine cytology; Bladder cancer
1. Introduction Diagnostic procedures in patients with bladder
Abbreviations: BTA, bladder tumor antigen; VUC, voided urine cytology; TCC, transitional cell carcinoma; PPV, positive predictive value; NPV, negative predictive value *Corresponding author. Tel.: 139-11-633-6765; fax: 139-11676-052. E-mail address: [email protected] (G. Priolo).
cancer symptoms include urine cytology, cystoscopy with biopsy, and excretory urography, but cystoscopy still represents the gold standard technique for the detection of both primary and recurrent tumors [1,2]. The majority of transitional cell carcinomas (TCC) of the bladder are superficial and thus amenable to conservative treatment by local endoscopic resection [3]. Yet, recurrence rate is very high, ranging from 40 to 85% according to different studies [4,5]. Because of the risk of recurrence, a
0009-8981 / 01 / $ – see front matter 2001 Elsevier Science B.V. All rights reserved. PII: S0009-8981( 00 )00416-2
48
G. Priolo et al. / Clinica Chimica Acta 305 (2001) 47 – 53
regular follow-up is required and patients must be examined at frequent intervals after treatment. The high demand for an endoscopic procedure, which is relatively invasive, has prompted the search for objective noninvasive methods that could reduce the need for cystoscopy in the surveillance protocols. Voided urine cytology (VUC) has played a significant role as a diagnostic tool in the assessment and follow-up of patients with bladder cancer [6], although a standardization of methods is lacking. Moreover, results of the test are highly dependent upon the skill of the cytopathologist, the correct preservation of urine and the number of exfoliated cells in the sample [7]. Despite 80% sensitivity for high grade tumors, urine cytology is a poor indicator of low grade tumors [8]. A number of bladder cancer tumor-related molecular markers [9], such as DNA ploidy, p53 mutations [10], and microsatellite DNA, are still difficult and expensive to perform and require trained personnel, whereas others, such as telomerase [11,12] have shown to be of limited value [13]. On the other hand, some non invasive urine analysis, including home hematuria screening [14], Nuclear Matrix Protein 22 (NMP 22) [15], and BTA [16], have become available for bladder cancer testing [17]. The BTA TRAKE assay is a quantitative enzyme immunoassay to detect a bladder tumor-associated antigen in the urine. This antigen has been recognized as a human complement factor H related protein (hCFHrp) and inhibits the complement pathway to cause cytolysis in cells with a resulting selective advantage for the tumor [18]. BTA TRAK sensitivity has been found to be superior to urine cytology and to the original qualitative BTA test [19]. We compared the sensitivity, the specificity and the diagnostic accuracy of the BTA TRAK test and that of VUC in the detection of primary or recurrent TCC of the bladder in a wide population of high risk patients.
2. Patients and methods Four hundred and seventy-three patients admitted to three Institutions in our hospital were enrolled
over a period of 1 year. These included 194 patients (25 females and 169 males, mean age 66.0 years; group 1) who had been scheduled for an urgent diagnostic cystoscopy in view of a high suspicion of bladder cancer (macroscopic painless hematuria and / or urographic / sonographic pattern suggestive of bladder tumor). The remaining 279, with a previous history of TCC, were undergoing a routine follow-up cystoscopy (42 females and 237 males, mean age 68.2; group 2). These latter had been treated with intravesical chemo- or immunotherapy following standard international protocols, mainly either low or full dose of Bacillus of Calmette–Guerin (BCG). Urine samples were collected by all patients on three consecutive days before the endoscopic procedure and sent for cytology. An additional urine sample was obtained and frozen below 2208C until BTA TRAK testing. Since in our experience urine samples may be stored at 48C for at least 24 h without loss of the antigen, specimens were collected and immediately put on ice or stored in the refrigerator at 48C before sending them to the laboratory. The stability of the marker in urine stored at 2208C is longer than 4–6 weeks. All urine investigations were carried out on an additional sample of 45 healthy volunteers (11 females and 34 males, mean age 58.7 years) who served as controls. Cystoscopy was followed by biopsy of the bladder mucosa in all cases even in the absence of macroscopic evidence of a tumor. Neoplastic lesions were staged according to the TNM classification and graded according to the WHO system. VUC was performed in all patients according to the standard procedures used in the Laboratory of the Institute of Pathology, University of Turin. All three specimens from each case were evaluated by a single experienced cytologist who was unaware of the results of BTA TRAK. For statistical purposes the patient was considered positive for tumor when at least one out of the three specimens showed cytological features of tumor. All cases showing features described as ‘cell atypia’ were categorized as dubious or suspicious results. The BTA TRAKE assay (DiaSorin, Saluggia, Italy) is an immunoenzymatic assay (IEMA) using monoclonal antibodies to bind specifically to bladder tumor antigen in urine. The lowest concentration of
G. Priolo et al. / Clinica Chimica Acta 305 (2001) 47 – 53
BTA detectable was estimated to be 0.75 U / ml. Assay reproducibility tests yielded ,5 and ,10% variation in intra- and inter-series runs, respectively (mean value: 30 U / ml). For this method no interference on results was reported by the manufacturer for a number of possible urine constituents such as hemoglobin (up to 100 mg / dl), red blood cells (up to 104 cells / ml) or by therapeutic agents, such as BCG (up to 20 mg / dl or 53106 CFU / ml), Mytomicin C (up to 10 mg / dl), and ampicillin (up to 600 mg / dl). Findings with hemoglobin and red blood cells were confirmed also by tests carried out in our center through the simulation of various degrees of hematuria. All procedures in this study were performed in accordance with the ethical standards as formulated in the Helsinki Declaration of 1975 (revision of 1996).
2.1. Statistical analysis As urinary BTA data were regarded to be not normally distributed, nonparametric comparisons of central tendency were employed for analysis. The distributions of urinary BTA were compared using the Mann–Whitney U test. A level of P,0.01 was accepted as statistically significant. Sensitivity was calculated as true positive test results / all patients with disease; specificity was calculated as true negative test results / all patients without disease. The receiver operating characteristic (ROC) curves were plotted and the areas under the curves were obtained by the trapezoidal rule [20]. The choice of the cut-offs was based on the best combination of sensitivity and specificity as derived by the analysis of their curve intersection. All data were expressed as median values and range.
3. Results One hundred and forty-seven out of the 194 cases (76%) with high suspicion of bladder cancer (group 1) were found to have a primary bladder TCC, the remaining 47 patients (24%) having a histological response negative for bladder cancer. Tumor recurrence was confirmed in 187 out of the 279 patients (67%) with previous history of bladder cancer
49
(group 2) while 92 (33%) were histologically recurrence free. In Table 1, patients from both groups are listed according to the different stage and grade of the neoplasia as well as according to the histological features of all non-neoplastic conditions. BTA TRAK median values and results from urine cytology (categorized as positive, negative and dubious, respectively) for each group of patients are reported in Table 2. Histology reports with no evidence of neoplasia have been grouped into two main categories named respectively ‘normal mucosa’ when no pathological findings were evidenced by the pathologist and ‘other pathologies’ for all benign conditions of the urothelium. BTA TRAK values were significantly higher in tumor positive cases than in the absence of bladder tumor for both group 1 (P50.0002 and P50.04 with respect to tumor negaTable 1 Patients referred for with a high suspicion of bladder cancer (group 1) and patients during follow-up for bladder TCC (group 2) are categorized according to histology results a n (%) Group 1 Tumor positive Stage Ta T1 T$2 Tx Total Grade 1 2 3 Total Tumor negative
86 (58.5) 29 (19.7) 29 (19.7) 3 (2.1) 147
42 (28.6) 55 (37.4) 50 (34) 147
Group 2
95 (50.8) 34 (18.2) 43 (23) 15 (8) 187
45 (24.1) 70 (37.4) 72 (38.5) 187
n
Normal mucosa Mucosal inflammation Cystic cystitis Granulomatous cystitis Squamous metaplasia Benign prostatic hyperplasia
33 10 3 0 0 1
69 20 1 1 1 0
Total
47
92
a
Tumor positive cases are classified following the different stages (TNM classification) and grades (WHO system) of the neoplasia. Tumor negative cases are grouped in different categories of non-neoplastic lesions.
G. Priolo et al. / Clinica Chimica Acta 305 (2001) 47 – 53
50
Table 2 Median values of BTA TRAK (range) are reported for each group of patients (group 1, patients with high suspicion of bladder cancer; group 2, patients in follow-up for bladder cancer; and group 3, healthy volunteers)a Patient group
Histology
BTA TRAK (U / ml) median (range)
Pb
Cytology positive
Cytology negative
Cytology dubious
n
Group 1
Tumor positive
80.1 (0.1–3000)
0.0001
87
42
18
147
Tumor negative Normal mucosa Other pathologies
21.2 (0.4–3000) 19.8 (0.4–1053) 25.0 (4.3–3000)
NS c
9 7 2
35 23 12
3 3 0
47 33 14
Tumor positive
78.9 (0.7–3000)
0.0001
95
64
28
187
Tumor negative Normal mucosa Other pathologies
25.1 (0.4–3000) 27.0 (0.4–3000) 24.5 (1.6–1005)
NS
11 8 3
67 50 17
14 11 3
92 69 23
0
44
1
45
Group 2
Group 3
9.2 (5.0–59.1)
a
Patients were classified as tumor positive when both cystoscopy and histology were positive. All non-neoplastic lesions (5tumor negative patients) have been grouped into two main categories: normal mucosa and all other benign conditions. b Mann–Whitney U-test. c Not significant.
tive patients with normal mucosa and other pathologies, respectively) and group 2 patients (P5 0.0001 and P50.0038, respectively). Non neoplastic urothelial diseases showed a trend towards a higher expression of the bladder tumor antigen although the difference was not statistically significant either in group 1 or in group 2. BTA TRAK concentrations in urine from patients with no mucosal abnormality (in both groups) were markedly increased compared to the baseline values from the control group. The majority of dubious responses at urine cytolo-
gy were from patients who had tumor positive results during the cystoscopical investigation. However, as shown in Table 2, those few cases with no tumor and ‘dubious’ cytological results belonged mainly to the subgroup with histology finding of ‘normal mucosa’. Sensitivity and specificity for BTA TRAK were evaluated assuming as control cases for group 1 the tumor negative patients and for group 2 the recurrence free patients (Table 3). With the method employed to calculate the cut-off threshold, a comparable BTA TRAK value was obtained for both
Table 3 Sensitivity, specificity, positive predictive value, negative predictive value and diagnostic accuracy for both BTA TRAK and urine cytology are reported either separately for group 1 and 2 or combining both groups a Patient group
Test
Cut-off (U / ml)
Sensitivity (%)
Specificity (%)
PPV
NPV
Diagnostic accuracy
1
BTA TRAK Cytology
35 Dubious as positive Dubious as negative
64 71.4 59.2
64 74.4 80.4
84.5 89.7 90.6
35.7 45.4 38.8
63.4 72.1 64.4
2
BTA TRAK Cytology
33 Dubious as positive Dubious as negative
63 65.8 50.8
63 72.8 88
77.1 83.1 89.6
45.2 51.1 46.8
62.7 68.1 63.1
112
BTA TRAK Cytology
34 Dubious as positive Dubious as negative
63 68.3 54.5
63 73.4 85.6
80.5 86 90
41.5 49 43.9
63 69.8 63.6
a
PPV, positive predictive value; NPV, negative predictive value.
G. Priolo et al. / Clinica Chimica Acta 305 (2001) 47 – 53
groups of patients (35 U / ml for group 1 and 33 for group 2), yielding 64 and 63% sensitivity and specificity as well as 63.4 and 62.7% accuracy, respectively. The same parameters were therefore determined also for all 473 patients (group 1 and group 2 combined, cut-off 34 U / ml), using as control cases all tumor free patients from both groups. ROC curve analysis showed a good diagnostic profile for BTA TRAK test. Sensitivity of urine cytology was determined for each group assuming all dubious results either as positive cases or as negative cases. In Table 3, sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and diagnostic accuracy are reported for each group and for both group combined. According to the grade of the bladder TCC, BTA TRAK sensitivity was 46, 58 and 81.1% for G1, G2 and G3 tumors, respectively, when calculated on the whole sample of patients (group 1 and group 2 combined). In the same analysis cytology resulted in 52.9, 66.1 and 84.4% sensitivity, when dubious cases were considered as positive, and in 41.4, 50 and 69.7% sensitivity, when dubious cases were considered as negative.
4. Discussion In the present report the diagnostic ability of BTA TRAK to detect a bladder TCC has been assessed in two clinical settings: history suggestive of bladder cancer, where a noninvasive diagnostic procedure may lead to an immediate and accurate cystoscopy, and in the follow-up of superficial bladder cancer patients, when the ability of an urine test to detect a tumor recurrence could be an useful tool for the selection of cases to be submitted for control cystoscopies. In this scenario urine cytology still plays an important role although it has a poor sensitivity to pick up low grade TCC [21]. Following the successful introduction of the BTA stat qualitative assay in the diagnosis of bladder cancer, a new quantitative assay detecting the same bladder antigen has been recently developed. Early results with the BTA TRAK assay have so far been promising. The
51
quantitative determination of the Bladder Tumor Antigen in the urine, named as BTA TRAK assay test, has shown higher sensitivity than the qualitative BTA stat test in the detection of a bladder cancer [22]. Conversely, urine cytology sensitivity is much lower in the same authors experience, never exceeding 44% [12,19,22]. The results from our large series seem not to be in contrast with those of two previous multicenter trials. Ellis [19] reported an overall sensitivity of the quantitative assay for the bladder tumor antigen of 72% using a cut-off of 14 U / ml and Thomas [23] found a BTA TRAK ability to detect a bladder cancer of 66%. In the present study the overall BTA TRAK sensitivity was 63%. A selection bias may be advocated to explain this lower value. The enrollment of patients with a high suspicion of bladder cancer (group 1) resulted in a higher positivity rate of bladder TCC following diagnostic cystoscopy. Similarly, the recurrence rate in follow-up patients (group 2) was very high (67%), pointing out that the diagnostic marker was assessed in a high risk bladder cancer group. A consistent part of the tumor free patients from both groups showed histological features of a benign pathology of the bladder. It should be assumed that other coexisting situations, like hematuria, among those with histologically normal urothelium could have interfered with the immunoassay. As a result, mean BTA TRAK values were much higher in tumor free patients than in the control group. This latter group was not included in our calculations as the investigation is not routinely addressed to asymptomatic patients. We believe our sample reflects more closely the population to which the test is applied. A value of 14 U / ml has been employed as a standard cut-off value in several reports on BTA TRAK [19,22,23]. This value is far lower the mean BTA concentration in tumor negative patients from our two groups but it is comparable with the mean value of BTA in the control group. Assuming the literature cut-off of 14 U / ml would result only in a mild increase in the sensitivity (87%) with a dramatic decrease of the specificity to 32%. Our cut-off was determined as the bladder tumor antigen concentration allowing the same sensitivity and specificity. The cut-off differences between the two groups was not significant and this finding allowed
52
G. Priolo et al. / Clinica Chimica Acta 305 (2001) 47 – 53
grouping all patients for the determination of a single sensitivity value. According to our data, urine cytology had the highest sensitivity and specificity power, particularly when the dubious results were assumed to be positive for malignant transitional cells. Cytological analysis was carried out on three consecutive samples and all the specimens were read by an expert cytopathologist. This could have contributed to the striking difference observed in comparison with other reports where urine cytology was performed on one single sample and assessed by several technicians [12,24]. Urine cytology was superior to BTA TRAK in the detection of primary or recurrent bladder cancer. The diagnostic advantage was maintained also when patients were stratified by tumor grade. Our study confirms the diagnostic utility of the BTA TRAK in the detection of a primary or a recurrent bladder cancer. An overall 70% sensitivity for urine cytology is attractive but probably highly dependent on an investigation carried out on three samples, which is time consuming and quite expensive, and on the availability of an expert cytopathologist, a situation difficult to maintain in many institutions. The diagnostic accuracy of BTA TRAK, although a bit lower than urine cytology, is acceptable and easily reproducible in different Institutions as shown by the substantial similar results with previous reports. Unfortunately, none of the two urine tests seems able to modify the diagnostic role of cystoscopy so far.
References [1] Lessing JA. Bladder cancer: early diagnosis and evaluation of biological potential. A review of newer methods. J Urol 1978;120:1–55. [2] Raghavan D, Shipley WU, Garnick MB, Russell PJ, Richie JP. Biology and management of bladder cancer. New Engl J Med 1990;322:1129–38. [3] Mostofi FK, Sobin LH, Torloni H. Histological typing of urinary bladder tumors. WHO International Histological Classification of Tumors, Geneva: World Health Organization, 1973, pp. 9–35. [4] Heney NM, Ahmed S, Flanagan MJ, Frable W, Corder MP, Haferman MD, Hawkins IR. Superficial bladder cancer: progression and recurrence. J Urol 1983;135:1083–6.
[5] Thrasher JB, Crawford ED. Current management of invasive and metastatic transitional cell carcinoma of the bladder. J Urol 1993;149:957–72. [6] Zein T, Wajsman Z, Englander LS, Gamarra M, Lopez C, Huben RP, Pontes J. Evaluation of bladder washings and urine cytology in the diagnosis of bladder cancer and its correlation with selected biopses of the bladder mucosa. J Urol 1984;132:670–8. [7] Murphy WM, Soloway MS, Jukkola AF, Crabtree WN, Ford KS. Urinary cytology and bladder cancer. The cellular features of transitional cell neoplasms. Cancer 1984;53:1555–65. [8] Gregoire M, Fradet Y, Meyer F. Diagnostic accuracy of urine cytology and deoxyribonucleic acid flow cytometry and cytology on bladder washings during follow-up for bladder tumors. J Urol 1997;157:1660–4. [9] Burchardt M, Burchardt T, Shabsigh A, De La Taille A, Benson MC, Sawczuk I. Current concepts in biomarker technology for bladder cancers. Clin Chem 2000;46:595– 605. [10] Herr HW, Bajorin DF, Scher HI, Cordon-Cardo C, Reuter VE. Can p53 help select patients with invasive bladder cancer for bladder preservation? J Urol 1999;161:20–2. [11] Rahat MA, Lahat N, Gazawi H, Resnick MB, Sova Y, Ben-Ari G et al. Telomerase activity in patients with transitional cell carcinoma: a preliminary study. Cancer 1999;85:919–24. [12] Ramakumar S, Bhuiyan J, Besse JA, Roberts SG, Wollan PC, Blute ML, O’Kane DJ. Comparison of screening methods in the detection of bladder cancer. J Urol 1999;161:388–94. [13] Arai Y, Yajima T, Yagihashi A, Kobayashi D, Kameshima H, Sasaki M et al. Limitations of urinary telomerase activity measurement in urothelial cancer. Clin Chim Acta 2000;296:35–44. [14] Messing EM, Young TB, Hunt VB, Newton MA, Bram LL, Vaillancourt A et al. Hematuria home screening: repeat testing results. J Urol 1995;154:57–61. [15] Soloway MS, Briggman JV, Carpinito GA, Chodak GW, Church PA, Lamm DL et al. Use of a new tumor marker NMP22 in the detection of occult or rapidly recurring transitional cell carcinoma of the urinary tract following surgical treatment. J Urol 1996;156:363–7. [16] Sarosdy MF, deVere White RW, Soloway MS, Sheinfeld J, Hudson MA, Schellhammer PF et al. Results of a multicenter trial using the BTA test to monitor for and diagnose recurrent bladder cancer. J Urol 1995;154:379–83. ¨ [17] Lokeshwar VB, Obek C, Soloway MS, Block NL. Tumorassociated hyaluronic acid: a new sensitive and specific urine marker for bladder cancer. Cancer Res 1997;57:773–7. [18] Kinders R, Jones T, Root R, Murchison H, Bruce C, Williams L, Hass GM. Human bladder tumor antigen is a member of the RCA (regulators of complement activation) gene family. J Urol 1997;157(suppl):28–36. [19] the Multi Center Study Group, Ellis WJ, Blumenstein BA, Ishak LM, Enfield DL. Clinical evaluation of the BTA TRAK assay and comparison to voided urine cytology and the Bard BTA test in patients with recurrent bladder tumors. Urology 1997;50:882–7.
G. Priolo et al. / Clinica Chimica Acta 305 (2001) 47 – 53 [20] Hanley JA, McNeil BJ. A method of comparing the areas under receiver operating characteristic curves derived from the same cases. Radiology 1983;148:839–43. [21] Murphy WM, Crabtree WN, Jukkola AF. The diagnostic value of urine versus bladder washing in patients with bladder cancer. J Urol 1981;126:320–2. [22] Irani J, Desgrandchamp F, Millet C, Toubert ME, Bon D, Aubert J et al. BTA stat and BTA trak: A comparative evaluation of urine testing for the diagnosis of transitional cell carcinoma of the bladder. Eur Urol 1999;35:89–92.
53
[23] Thomas L, Leyh H, Marberger M, Bombardieri F, Bassi P, Pagano F et al. Multicenter trial of the quantitative BTA TRAK assay in the detection of bladder cancer. Clin Chem 1999;45:472–7. [24] Del Nero A, Esposito N, Curro’ A, Blasoni A, Montanari E, Mangiarotti B et al. Evaluation of urinary level of NMP 22 as a diagnostic marker for stage pTa-pT1 bladder cancer; comparison with urinary cytology and BTA test. Eur Urol 1999;35:93–7.