PROSPECTIVE EVALUATION OF GENETIC ABNORMALITIES AND TELOMERASE EXPRESSION IN EXFOLIATED URINARY CELLS FOR BLADDER CANCER DETECTION

Abstract

To evaluate alternative procedures to cytoscopic examination we prospectively compared noninvasive procedures for detecting bladder cancer namely cytology, loss of heterozygosity (LOH), microsatellite instability and human telomerase catalytic subunit reverse transcriptase (hTERT) messenger (m) RNA detection. Specificity and cutoff values were established in the blood and urine sediment of 50 controls. Sensitivity was analyzed in the urine and tissue samples of 50 patients with bladder cancer. The diagnosis was established by cystoscopic and histological examination. Genomic alterations were studied using a panel of 24 microsatellite markers to detect LOH events, while 3 additional mononucleotide repeats were analyzed for microsatellite instability detection. Telomerase expression was detected in urinary cells by nested RT-polymerase chain reaction amplification of hTERT mRNA. All techniques were compared by cytological examination. Sensitivity and specificity were 31% and 100% for cytological testing, 96% and 100% for LOH, and 75% and 69% for RT-polymerase chain reaction of hTERT, respectively. No alteration was detected on microsatellite instability analysis in urine or tumor tissue cells. Using only the 5 markers most strongly associated with bladder cancer selected by logistic regression analysis, namely ABL1, IFNa, D9S12, MJD58 and D18S364, LOH test sensitivity slightly decreased to 90%. Urinary LOH analysis was the most sensitive and specific method for bladder cancer detection and it appeared less dependent on urine sediment quality. The logistic regression score may be an interesting complement to cystoscopy. The specificity of hTERT mRNA detection was incomplete since false-positives were observed in 31% of cases. Absent microsatellite instability in our cohort showed that these genomic alterations are not present at the early step of bladder cancer.