Prostate Cancer: Diagnostic Performance of the PCA3 Urine Test

George Leighton Lee; Albert Dobi; Shiv Srivastava


Nat Rev Urol. 2011;8(3):123-124. 

Abstract and Introduction


The potential of the prostate cancer antigen 3 (PCA3) urine assay to aid prostate cancer diagnosis and minimize unnecessary biopsies has been extensively studied. Results from three recent studies that compared the performance of PCA3 with PSA underscore the advancement and future challenges for this new diagnostic biomarker.


The widely used PSA blood test has had a major role in the early detection of prostate cancer. However, the poor specificity of the PSA test for detecting cancer, or for differentiating indolent cancers from clinically significant ones, has resulted in a large number of unnecessary biopsies and overtreatment of men with prostatic adenocarcinoma. Meeting the challenge of improving the diagnosis and prognosis of men with this malignancy will be critically dependent on the development of more-accurate prostate cancer biomarkers. To address this need, the utility of various PSA molecular forms and kallikreins, as well as prevalent alterations in genes that are associated with prostate cancer (such as GSTP1, PCA3, AMACR, ERG, and gene fusions involving ETS-related genes), are being actively investigated.[1]PCA3 (formerly known as DD3) is a prostate-tissue-specific, noncoding messenger RNA (mRNA) that is selectively overexpressed in virtually all prostate carcinoma specimens compared to normal prostate tissue.[2] These attributes of PCA3 mRNA expression make it a promising prostate-cancer-specific marker. After the initial approach of measuring PCA3 mRNA levels in urine specimens following prostatic massage, a streamlined PCA3 assay was developed (Progensa®; Gen-Probe, San Diego, CA), and is currently being evaluated for its diagnostic performance in urine collected after digital rectal examination (DRE).[3,4] The diagnostic utility of urine PCA3 levels for predicting cancer on repeat biopsy has been previously established.[1] Here, we examine the results of three large studies—involving a total of almost 2,000 European patients and including two multicenter evaluations—that assessed the diagnostic performance of PCA3 in relation to PSA.[5–7]

In a reanalysis of data from a previously reported multicenter study,[8] Ploussard et al.[5] compared the performance characteristics of PCA3 with those of free:total PSA ratio (a low free:total PSA ratio is associated with an increased specificity for prostate cancer detection) in a cohort of over 300 patients with a previous negative biopsy and total serum PSA levels of 2.5–10.0 ng/ml—a 'gray zone' of PSA for predicting prostate cancer. The study examined PCA3 cut-off scores of 25, 30, and 35 in patients grouped according to their free:total PSA ratios (>20%, 10–20% or <10%). PCA3 score >30 was a predictor of positive biopsies, especially in the group of patients with free:total PSA ratios >10%. In a direct comparison, PCA3 score was superior to PSA in terms of specificity and predictive values, but inferior with regard to sensitivity. The authors also noted that PCA3 was not associated with prostate volume or Gleason score, whereas free:total PSA ratio correlated with prostate volume, age and stage.

In another multicenter report, Auprich et al.[6] assessed and validated the accuracy of PCA3-based nomograms in a cohort of patients at risk for prostate cancer. More than 600 patients underwent ≥10-core laterally directed prostate biopsy on the basis of suspicious DRE, persistently elevated PSA level (2.5–10.0 ng/ml) or suspicious histology on previous biopsy. Prostate cancer was detected in 41.1% of patients. Median PCA3 scores were significantly higher in patients with biopsy-detected disease compared to patients with negative biopsies (48 versus 20, P <0.001), which is consistent with the findings of current reports in the literature.[4,8,9] This study highlights the applicability of PCA3-based nomograms in European men and demonstrates the robust performance characteristics of PCA3 within the limitations of early detection and referral scenarios.

In the third study, Roobol et al.[7] examined the performance of PCA3 score as a first-line diagnostic test similar to the current use of PSA measurement. The study evaluated nearly 1,000 men from a prescreened population, 75% of whom underwent sextant biopsy. When comparing the performance characteristics of PCA3 score to PSA level, with a focus on cases with an indication for biopsy, the authors set strict PCA3 score and PSA level cut-offs of ≥10 and ≥3.0 ng/ml, respectively. The study revealed that PCA3 performed marginally better than PSA, but the difference was not statistically significant (P = 0.143). Although there were no strict cut-offs at which PSA and PCA3 had an acceptable balance of sensitivity and specificity, the best balance was achieved with PSA level >2.0 ng/ml and PCA3 score >35. An intriguing finding of this report was the improved capability of PCA3 to identify the higher risk prostate cancer cases, which will likely prompt further investigations.

The overall premise of these studies was to determine whether the measurement of PCA3 scores could spare men from unnecessary biopsies without missing instances of clinically significant cancer. All three articles analyzed PSA, DRE and PCA3 data from patients undergoing initial and/or repeat biopsies, but they differed in terms of how many biopsy cores were collected (sextant biopsy versus >10-core biopsy). In light of the findings reported by Ploussard and colleagues,[5] it seems that a PCA3 score cut-off of >30 achieves a 30% rate of positive biopsies. Consistent with the recent literature,[4] Roobol et al.[7] found that they achieved the best balance of sensitivity and specificity by using 35 as the PCA3 score cut-off;[8,9] however, by using this cut-off as a first-line detection tool, 39% of prostate cancers were missed, along with 26.3% of tumors that were considered "serious", such as Gleason score >4 or stage >pT2a disease.

Another common thread throughout these articles, and the literature as a whole, is that PCA3 has a greater specificity than PSA level and free:total PSA ratio but a slightly lower sensitivity, making PCA3 a good complementary marker to PSA. These reports also concurred that PCA3 scores were independent of prostate volume, which improves its diagnostic accuracy compared to PSA. Overall, area under the curve (AUC) values for the PCA3 urine test were similar in the reports of Auprich et al.[6] (AUC of 0.739 with a PCA3 score cut-off of >35) and Roobol et al.[7] (AUC of 0.635 overall and 0.681 in patients with previous negative biopsies). In summary, the potential of the PCA3 urine assay is promising in some diagnostic settings, and further improvements are expected by combining PCA3 with other molecular markers of prostate cancer.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: