Comparison of cytokeratin 20 RNA and angiogenin in voided urine samples as diagnostic tools for bladder carcinoma

Background: We evaluated the diagnostic efficacy of urinary angiogenin (ANG) and cytokeratin 20 (CK-20) mRNA in comparison with voided urine cytology in the detection of bladder cancer patients. Objectives and methods: A total of 97 Egyptian patients provided a single voided urine sample for ANG, CK-20 and cytology before cystoscopy. Of the 97 cases, 63 were histologically diagnosed as bladder cancer; 33 with transitional cell carcinoma (TCC) and 30 with squamous cell carcinoma (SCC), whereas the remaining 34 had benign urological disorders. A group of 46 healthy volunteers were also included in this study. Voided urine was centrifuged and the supernatant was used for estimation of ANG by EIA and confirmed by Western blotting (WB). The urine sediment was used for cytology and RNA extraction. CK-20 RNA was detected by RT-PCR. Results: The best cutoff value for ANG was calculated by a ROC curve as 322.7 ng/mg protein. The median urinary ANG level in bladder carcinoma, benign urological disorders and healthy volunteer groups was: 802.7, 425 and 33 pg/mig protein, respectively. The positivity rate for urinary CK-20 mRNA of the control, benign and malignant groups was 0%, 2.9% and 82.3%, respectively (P = 0.000); while the rates for ANG were 11.6%, 54.8% and 75.4%, respectively (P = 0.000). There was no significant difference in positivity rates of CK-20 and ANG with respect to sex, smoking, schistosomiasis, urine cytology, tumor grade, tumor stage, hematuria or pus cells. The overall sensitivity and specificity were 71.4% and 90% for voided urine cytology, 75.4% and 70.3% for ANG, and 82.3% and 98.8% for CK-20. Combined sensitivity of voided urine cytology with ANG and CK-20 together (98.2%) was higher than either the combined sensitivity of voided urine cytology with ANG (96.5%) or with CK-20 (91.6%) or than that of the biomarker alone. We demonstrated significant positive correlation between CK-20 positivity with age (P = 0.043) and nodal involvement (P = 0.037); however, there was no significant correlation between CK-20 and ANG with the other clinicopathological parameters. Conclusions: Our data indicate that CK-20 and ANG in voided urine had higher sensitivities compared to voided urine cytology. However, when specificity was considered, CK-20 alone had superior sensitivity and specificity compared to ANG and voided urine cytology. (C) 2004 The Canadian Society of Clinical Chemists. All rights reserved.