Over the past 20 years a large number of prostate cancer (PCa) markers have been proposed. A combination of PCA3 and TMPRSS2-ERG is among them. The real-time RT-PCRbased method for РСА3 and TMPRSS2-ERG determination in urine and prostate tissue is being developed at Russian Scientific-research Institute of Urology. Purposely to assessment of its diagnostic value, two independent samples of prostate tissue (n=78) and urine (n=56) specimens were investigated. The blood level of PSA, pathomorphological findings, PCA3 and TMPRSS2-ERG urinary and tissue expression were assessed. Samples were divided into groups, according to pathomorphological findings: PCa and non-cancerous prostatic changes (BPH and chronic prostatitis).
Prostate tissue РСА3 expression was significantly higher in PCa patients (Ме РСА3 in tissue 3,1 (1,2:4,9)) compared to BPH and chronic prostatitis patients (Ме РСА3 in tissue 9,1 (6,9:10,4)) (р< 0,01). Urinary РСА3 demonstrated the same picture: Ме of РСА3 in urine was -0,04 (-0,6:2,1) in PCa patients and 4,4 (1,3:7,4) in patients with BPH and chronic prostatitis (р< 0,01). TMPRSS2-ERG tissue expression was detected in 7 (35,0%) patients with PCa. Expression of urinary TMPRSS2-ERGwas found in 4 (14,8%) PCa patients.
We determined a cut-off score for РСА3 in prostate tissue and in urine, using ROCanalysis: ≤5,7 and ≤ 3,3 (diagnostic accuracy 88,0% and 71,4%, sensitivity 95,0% and 96,3%, specificity 86,2% and 51,7%), respectively. Area under the curve (AUC) was 0,952 for tissue РСА3 and 0,768 for urinary РСА3(р< 0,0001). Therefore РСА3 is a high quality marker. TMPRSS2-ERG diagnostic efficacy was lower compared to РСА3: AUC for TMPRSS2-ERG was 0,675 in tissue and 0,574 in urine, respectively.
The method for РСА3 and TMPRSS2-ERG determination in urine and prostate tissue showed good diagnostic performance, and it could be useful for the improvement of early PCa diagnosis, particularly it could lead to optimizing costs by lowering amount of unnecessary prostate biopsies.
Attachment | Size |
---|---|
Download | 382.2 KB |