Re: Scott A. Tomlins, John R. Day, Robert J. Lonigro, et al. Urine TMPRSS2:ERG Plus PCA3 for Individualized Prostate Cancer Risk Assessment. Eur Urol. In press.


The recent study by Tomlins et al [1] provides detailed information about the ability of the urinary marker TMPRSS2:ERG (T2:ERG) in combination with serum prostate-specific antigen (PSA) and the US Food and Drug Administration (FDA)–approved urinary prostate cancer antigen 3 (PCA3) gene, by using the so-called Mi-Prostate Score (MiPS), to improve prostate cancer (PCa) and highgrade PCa detection. However, as already briefly discussed by Tomlins et al, it is of extraordinary practical interest to know how the clinical validity of this new marker combination MiPS is to be judged in comparison with the Prostate Health Index (Phi), as another FDA approved marker [2], or in combinations with this PSA derivative. For this purpose, we used the given MiPS formula and recalculated our own data (area under the receiver operating characteristic curve [AUC] and decision curve analysis [DCA]) from2013, whenwe first compared T2:ERG, PCA3, PSA, and Phi [2]. The data from our 246 patients (PCa, n = 110; no evidence of malignancy [NEM], n = 136) indicated that the highest AUC was shown not by MiPS (AUC 0.748) or MiPS to detect high-grade PCa (AUC 0.747) but rather by our artificial neural network (ANN; AUC 0.757) and the logistic regression analysis (AUC 0.752) that included PCA3 and Phi. The differences were not significant (p values from 0.61 to 0.87). Neither MiPS value could significantly improve Phi (p = 0.10 and p = 0.11) or PCA3 alone (p = 0.45 and p = 0.51). In addition, the data of the DCA showed comparable results for MiPS, PCA3, and Phi alone (Fig. 1A). However, a distinct net benefit in risk stratification of PCa patients was evident if patients were biopsied based on the ANN model that included PCA3 and Phi in comparison with MiPS or PCA3alone (Fig. 1B). One reason for this improvednetbenefit could be that Phi behaves as an orthogonalmarker [3] versus PCA3 and T2:ERG in PCa patients (rs = 0.040 and rs = 0.111; p = 0.678 and p = 0.248) and NEM patients (rs = 0.024 and rs = 0.007; p = 0.780 and p = 0.934), whereas PCA3 and T2:ERG correlate with each other (rs = 0.344; p < 0.001). Moreover, the clinical potential of T2:ERGseems tobeapriori limited, as it occurs in only 50% of all PCa patients [4], whereas Phi shows diagnostic advantages with its particular ability to detect aggressive tumors in comparison with PCA3 [5]. In conclusion, it would be worthwhile to validate the combination PCA3 and Phi regarding their real potential for diagnosis and risk stratification of PCa patients in new studies. E U RO P E AN URO LOG Y 6 8 ( 2 0 1 5 ) e 1 0 6 – e 1 0 7

DOI: 10.1016/j.eururo.2015.07.028

1 Figure or Table


Citations per Year

775 Citations

Semantic Scholar estimates that this publication has 775 citations based on the available data.

See our FAQ for additional information.

Cite this paper

@article{Stephan2015ReSA, title={Re: Scott A. Tomlins, John R. Day, Robert J. Lonigro, et al. Urine TMPRSS2:ERG Plus PCA3 for Individualized Prostate Cancer Risk Assessment. Eur Urol. In press.}, author={Carsten Stephan and Henning Cammann and Klaus Jung}, journal={European urology}, year={2015}, volume={68 5}, pages={e106-7} }