Dr. Peter F. A. Mulders spoke at the 24th International Prostate Cancer Update on Thursday, February 20, 2014 on “MRI and Biomarkers: Targeting Treatment Decisions in Prostate Cancer” In his presentation, Dr. Mulders discusses how to detect the biomarker and how to implement that biomarker in the clinic.



Keywords: biomarkers, multiparametric MRI, PCA3, biopsies, TMPRSS2 ERG, PSA, Gleason score, screening

How to cite: Mulders, Peter F. A. “MRI and Biomarkers: Targeting Treatment Decisions in Prostate Cancer.” January 9, 2015. Accessed Apr 2020. https://grandroundsinurology.com/prostate-cancer-peter-fa-mulders-mri-and-biomarkers/.



MRI and Biomarkers: Targeting Treatment Decisions in Prostate Cancer

You’ve heard a lot of biomarkers and how we should implement it in the clinic, and I am happy that we have some case discussions later on with the panel. And, of course, we have to see how we can balance these new biomarkers here.

Because Alan gave an excellent talk on biomarkers also including PCA3 I can go a little bit quicker through that data. But I also figured out that Alan presented most of the US data, and I have most of the European data and maybe that is also a good balance. So I will address the biomarker issue. It’s not that it is a characteristic only for pathologic processes, but also for normal biological processes, which is important to detect the biomarker and how to implement that biomarker in the clinic.

So the ideal prostate cancer biomarker is a biomarker that comes from the tumor tissue. It should be non-invasive, easy to manage, and that is one thing that you have to implement into your clinical practice because the cost is also important in the care of prostate cancer patients. So it should be able to detect prostate cancer in the early stage, but also? and that is I think the most important issue these days to differentiate between aggressive and non-aggressive cancers.

And I remember when we had a meeting in – – two weeks ago, it was the Winter Forum where it was taped and especially coming, I think, from Canada that maybe in the future we should not treat Gleason 6 prostate cancers anymore so that is another issue that comes up, and hopefully a biomarker will help us in that. And, of course, you have seen the area under the curve with specificity and sensitivity which should be reflected. And this is the dilemma which, especially in Europe, has a big impact actually.

Yes, we know that we can lower with 20% the chance of dying from prostate cancer, but we also know that we have to screen 1,410 patients to prevent one death. And this is really the issue of significance between patients from dying from prostate cancer and patients dying with prostate cancer. So can we better identify the patients? And that is another issue because biopsies of the prostate are not without danger. How can we prevent unnecessary repeat biopsies? Can we identify the candidates? Even for chemo prevention, we’ve heard that – – can be fine-tuned better in chemo prevention with biomarkers. And that is what I will elucidate a little bit on, the screening study. How can we implement new biomarkers in that screening study?

Active surveillance, and this is becoming more and more important that I will mention to you later on, the role of MRI and of biomarkers in the active surveillance issue. Localized, what kind of a treatment also a focal therapy or taking the whole gland out, and in advanced cancer the indication for endocrine therapy and the follow-up of these patients, so how to treat. And I will just mention a little bit of PCA3 because I think it’s an example of how we try to I’m prove the PSA dilemma. Actually it’s a non-coding RNA and it’s overexpressed as was already mentioned in more than 95% of prostate cancers.

And this is the publication in 1999 by Marion Bussemakers when she did her post-doc at Johns Hopkins and came back and detected the DD3 gene. And this is a little bit of how it evolved and hopefully in the future the new biomarkers will evolve much quicker than it was done with PCA3. Here you can see the Johns Hopkins connection and what we always had with our lab in Nijmegen with Bill Isaacs and also the others. And then finally it was detected in the 1999 paper that this differential display was specific for prostate cancer.

But it took some time to detect that it was not coding for a protein, so it cannot be a blood test. And remember at that time I came back from my fellowship at UCLA where I investigated kidney cancer. We had to make a decision to stop the whole program because it wasn’t the protein, and what was the use of a biomarker which is not a protein. And then the non-coding RNA, or the so-called RNA’s, became clear. And in one way or another in a discussion also together with the Canadians, I mentioned the way of a urine test. And remember if you have a question on that, that the PCA3 is a urine test because you need the whole cells to detect PCA3.

And, finally, in 2012, look at the first 1993 and 2012, finally it is FDA approved and we still have discussions on how to implement it in our clinical practice. And this is the way that PCA3 was better than PSA because it is a 66-fold overexpressed in prostate cancer cells which is much more than PSA. It is a urine test, it is a quantified in the way by a digital rectal exam. The first portion of urine is detected and measured in a qualitative way and you’ve got a PCA3 score.

Initially the place was to prevent unnecessary biopsies, so after the first biopsy. And in general, you can say this is more than 1,500 patients that the sensitivity and the specificity is pretty consistent in all of the tests done. Here you can see the US cohort of patients with the specificity and the sensitivity there, and the European cohort is quite similar with an area under the curve compared with PSA, which is much better. And if you see all of the new biomarkers coming out and if you compare that with PSA, you can always see these kinds of curves.

And the curve should be ideally very steep here, and the question is how steep should it be to have really a place into the clinic. You can see also here that there is a qualitative relationship between the higher PSA level and the percentage of positive biopsies in the US; 27% was positive and in the European study 25% was positive, so it’s quite consistently a qualitative relationship between the absolute number of PCA score and the chance of positive biopsies.

So this is the – – , the clinical indication is to prevent repeat biopsies. And therefore it was approved, but there are more clinical studies of course. If you have a biomarker you try to see if there is a prognostic value. And the prognostic value should be to detect the significant cancers out of the non-significant cancer because that has a real clinical implication. You can see here the hypothesis. If the volume of the tumor is high, the likelihood of a positive PCA3 should be higher, especially with a higher grade. The PCA3 should be high. If that’s the case then it would really have an impact on your clinical practice.

And actually it can determine the significance from the significant cancers. If that is black and white, I think it’s not but there is a relationship. And actually eight studies confirmed this observation. You can see relatively small numbers with the significant cancers, like let’s say the high volume higher grade compared with a low volume low grade that has a really significant difference in the PCA3 level. And here you can also see that the higher the PCA3 is, the more likely you will have a Gleason 7 score. There is some discrepancy in the really high scores with the Gleason 8 and the Gleason 9 score, but PCA3 is not really reflecting well because these cells are so differentiated.

So the demand for screening, it’s easy, non-invasive, and that is the case for PCA3. It’s really significantly more specific. PSA actually, most of our new biomarkers search, you see the area under the curve really better than PSA which is a little bit of a flip of a coin. It’s significantly less of an overdiagnosis so you can prevent repeat biopsies and, hopefully in the future, and that is the discussion that is still ongoing, to determine the significant from the non-significant cancers. Cost effective; it is relatively expensive. In Europe it’s about 200 Euro for the test, which is far more than the PSA so that is a limitation.

So we tried the implement that in the screening study, and we asked the screening people like Fritz Schroder and Monique Roobol to determine if PCA3 will add onto the screening of prostate cancer. And, yes, the conclusion is there that PSA as a first screening test shows improvement in the performance characteristics of detecting prostate cancer. And there is a chance that it may detect the significant cancers out of that. You can see that everything was tested and the positive predictive value, etc, all based on PSA and PCA3 scores. And if you focus on the unnecessary biopsies, you can see to present it as a necessary biopsy goes especially with a higher PSA score as depicted in this slide.

So as a diagnostic tool, it is approved. You can use it with a cutoff of 35, and it can add to the nomograms already available. Like you can see here in the determination of prostate cancer diagnosis with an area under the curve that finally goes up to 0.8 in one which is actually accurate for prostate cancer detection. As a prognostic tool, as I showed you, hopefully it can detect the significant cancers out, but I think we still need validations on this issue.

TMPRSS2 was determined. Why should we add new markers on it, and that is because of the limitations of your initial marker. And this is the same as what we are currently doing in our lab now. TMPRSS was explained now. We did a study with a panel, PCA3, and TMPRSS2 ERG. And you can see here that especially the sensitivity increases with the combination of TMPRSS2 and PCA3. We published that in the European Urology in a prospective multicenter way, the combination of PCA3 and TMPRSS2 ERG. And you can see again the area under the curve improves when you add TMPRSS2 ERG to PCA3. It’s not statistically significant, is it clinically relevant, that is something which we should determine now and in the future. We are evaluating new markers, so it goes with a fast screening method that we are trying to add on to the performance of PCA3 and TMPRSS2 ERG with multiple markers. We call it the Quattro. We submitted this for publication and it is quite straightforward.

When we obtain mRNA of frozen tissue then we validate independently with normal cancerous tissue, and finally we come up with a panel of candidate genes. The four genes we detected here were implemented in a validation study of 358 patients.And you can see that with the multiple panel the accuracy and the negative predictive value increases with hopefully especially in the Gleason 7 and more prostate cancers. And this is once again an example that we really, these days, focus on the significant versus the non-significant cancers because why should we detect a Gleason 6 prostate cancer, and that is the discussion that we can have later on in the panel. So the panel is superior to PCA3 alone. These are the curves. Once again, the PSA curve and the improvement with these 4-gene panels as we have are now going up to more than the 0.8 area under the curve.

So, the conclusion here is, yes, we have new markers. They are significantly better than PSA, but we can add upon that by adding TMPRSS2 ERG gene fusion, but also these new markers to increase the accuracy of our biomarker panel to detect significant cancers out of the non-significant cancers. This is an overview that we published if you want to have a look in Clinical Biochemistry which is all of the PCA3 tests quite consistently with the sensitivity. The combination study is done, this is the study I presented to you, and also quite consistently in the area under the curve with the combination.

So, once again, we have new biomarkers, but are we able to implement these biomarkers in our clinic because one other thing is happening now, and especially most of the patients we see in our clinic do have an MRI. It’s not the MRI, it’s the multiparametric MRI. If you see all of the tests done with the biomarkers because of ultrasound biopsies as an endpoint, and there is a significant difference between an MRI biopsy compared with an ultrasound biopsy. We will have that later on with fusions, but I think the multiparametric MRI really has an impact on our clinical decision of what to do with biomarkers, what to do with screening, etc, etc. Multiparametric MRI is not the MRI, so it’s not that you have to say, well, my patient had an MRI.

No, the multiparametric MRI is really significantly better in its performance than the normal MRI independent of the number of tests that you have there. So this is a 3-Tesla multiparametric MRI. And I really was surprised on the radiologist with the yellow bands who really tries to tell what kind of a Gleason score he sees and in what kind of an area. Actually if you talk about screening, his goal is to screen every patient, every man over 40 with a multiparametric MRI. That will see what the discussion is and what we had with PSA and PCA3.

So it’s really a different ballgame where we are entering now with MRI. MRI is really excellent in functional soft tissue, contrasts with the multiparametric MRI and really adds tot performance of MRI. It’s able to localize the tumor not only for your biopsy, but also finally for your therapy. That is another thing that is happening. We treat more focally in our prostate cancer patients. If you see the sensitivity and the specificity of multiparametric MRI with the aggressiveness of the cancer and really surprise in his data, it has to be validated prospectively on what is currently done. And that is also what I figured out in my clinical practice. With biomarkers and MRI, I really feel more safe in surveillance protocols, even without biopsies.

We did an investigation because we had a lot of patients with PCA3, we had a lot of patients with multiparametric MRI. We decided to have an investigation done by one of our PhD’s to see if MRI and PCA3, and that was the discussion of what we had with the other barons because she said we have to see if PCA3 is still necessary in a patient who had received an MRI. But you can see here that there is collation between the detection of significant cancers with MRI and the PCA3 score. Actually the patients with a significant cancer had a PCA3, on MRI, had a PCA3 of 52, and the patients with a normal MRI had a PCA3 of 21 so there is a correlation definitely.

And because you have two diagnostic tools it’s now the discussion of how to implement that in your clinical practice. Should you start with an MRI or should you start with PCA3 to detect if the MRI is still necessary. So this is another study done as a more randomized study. This study is where you can see that the MRI, the multiparametric MRI, increases the sensitivity of PCA3. This is a study with two groups; one group just under TRUS guided biopsies, and the other group had multiparametric MRI and biopsies to see the performance. It’s not really spectacular but it is significant and better in the patients who had a multiparametric MRI in the detection of prostate cancer.

So what is the take home message? We are in a spectacular time with all new biomarkers these days. They should be implemented and I am happy to see that if you present a new marker you need a validation study. The validation study is to see if you can implement it in clinical practice. But there is still an unmet need because we are confronted in the time that we use biomarkers in the overtreatment issue, and especially the discussion of the Gleason 6 as it may be provocative. A Gleason 6 can still be named a prostate cancer.

The MRI, especially the multiparametric MRI, implementation is really crucial and maybe we should go over all of the old studies and, once again, to treat the whole gland compared with focal therapy is another issue which will have an impact on the role of biomarkers. So the personalized medicine in the 21st century is key. Personalized medicine with biomarkers, with MRI, but also with treatment options.


Auprich M1, Chun FK, Ward JF, et al.Critical assessment of preoperative urinary prostate cancer antigen 3 on the accuracy of prostate cancer staging. Eur Urol. 2011 Jan;59(1):96-105. http://www.ncbi.nlm.nih.gov/pubmed/20980098

Biomarkers Definitions Working Group. Biomarkers and surrogate endpoints: preferred definitions and conceptual framework. Clin Pharmacol Ther. 2001 Mar;69(3):89-95. http://www.ncbi.nlm.nih.gov/pubmed/11240971

Busetto GM, De Berardinis E, Sciarra A, et al.Prostate cancer gene 3 and multiparametric magnetic resonance can reduce unnecessary biopsies: decision curve analysis to evaluate predictive models. Urology. 2013 Dec;82(6):1355-60. http://www.ncbi.nlm.nih.gov/pubmed/19304372

Chun FK, de la Taille A, van Poppel H, et al. Prostate cancer gene 3 (PCA3): development and internal validation of a novel biopsy nomogram. Eur Urol. 2009 Oct;56(4):659-67. http://www.ncbi.nlm.nih.gov/pubmed/24080222

Deras IL1, Aubin SM, Blase A, et al. PCA3: a molecular urine assay for predicting prostate biopsy outcome. J Urol. 2008 Apr;179(4):1587-92. http://www.ncbi.nlm.nih.gov/pubmed/18295257

Dijkstra S, Mulders PF, Schalken JA.Clinical use of novel urine and blood based prostate cancer biomarkers: a review. Clin Biochem. 2014 Jul;47(10-11):889-96. http://www.ncbi.nlm.nih.gov/pubmed/24177197

Groskopf J, Aubin SM, Deras IL, et al.  APTIMA PCA3 molecular urine test: development of a method to aid in the diagnosis of prostate cancer. Clin Chem. 2006 Jun;52(6):1089-95. http://www.ncbi.nlm.nih.gov/pubmed/16627561

Hambrock T, Somford DM, Hoeks C, et al. Magnetic resonance imaging guided prostate biopsy in men with repeat negative biopsies and increasedprostate specific antigen. J Urol. 2010 Feb;183(2):520-7. http://www.ncbi.nlm.nih.gov/pubmed/20006859

Hansen J, Auprich M, Ahyai SA, et al. Initial prostate biopsy: development and internal validation of a biopsy-specific nomogram based on the prostate cancer antigen 3 assay. Eur Urol. 2013 Feb;63(2):201-9. http://www.ncbi.nlm.nih.gov/pubmed/22854248