Use of MRI to Risk Stratify Patients with Prostate Cancer- Simpa Salami

June 19, 2019

Simpa Salami, MD discusses the role of MRI in risk stratifying patients with prostate cancer.  MRI has revolutionized the ability to detect aggressive prostate cancer but it is still possible that patients who have a negative MRI still have aggressive cancer. Up to 10-15% of patients may have aggressive prostate cancer despite negative findings on MRI.  Dr. Salami addresses why these patients may have negative MRI findings and his research suggests that it is due to genomic differences of the tumors and in particular the tumor architecture.  Tumors that are more well organized tend to be missed on the MRI.  Dr. Salami's take-home message is that MRI alone should not be used to risk stratify patients.  For patients that have a negative MRI, with rising PSA, and other finding such as a positive digital rectal exam, they should undergo a biopsy and not rely solely on the MRI. 


Simpa Salami, MD, MBBS, MPH, Department of Urology, Hofstra Northwell School of Medicine

Charles J. Ryan, MD, the President and Chief Executive Officer of The Prostate Cancer Foundation (PCF), the world’s leading philanthropic organization dedicated to funding life-saving prostate cancer research. Charles J. Ryan is an internationally recognized genitourinary (GU) oncologist with expertise in the biology and treatment of advanced prostate cancer. Dr. Ryan joined the PCF from the University of Minnesota, Minneapolis, where he served as Director of the Hematology, Oncology, and Transplantation Division in the Department of Medicine. He also served as Associate Director for Clinical Research in the Masonic Cancer Center and held the B.J. Kennedy Chair in Clinical Medical Oncology.

Read the Full Video Transcript

Charles Ryan: Joining me today is Simpa Salami, Assistant Professor of Urology at the University of Michigan. 

Dr. Salami, thank you for joining us. You have a really interesting presentation here. Tell us about what you're talking about at EAU.

Simpa Salami: I will be speaking about the use of MRI to risk stratify patients with prostate cancer. The question that I'm trying to address is the fact that you don't see certain cancers on the MRI of the prostate, does that mean that they are not biologically important?

Charles Ryan: So MRI has become a very highly used tool now for staging and detection of localized prostate cancer. Are you suggesting that perhaps urologists are missing some tumors by doing MRIs and that it may not be as useful as we originally had thought?

Simpa Salami: I think the MRI has really revolutionized the ability to detect aggressive prostate cancer. Personally, in my practice, I routinely use MRI before I proceed to do a prostate biopsy. However, what I'm trying to say is that in patients with negative MRI, it's still possible that they have aggressive prostate cancer, and this has been estimated to be about 10 to 15%, in some series even up to 20%. So the question is should we really be depending on MRI alone in making the decision to put a patient on surveillance or to treat a patient?

Charles Ryan: I see, so you're going to miss some high-grade tumors in 10 to 15% of the cases in which you are relying on MRI tests to make treatment decisions.

Simpa Salami: That is correct.

Charles Ryan: And is it that the MRIs are completely negative and there's high-grade disease there, or that there's high-grade disease outside of the space where you see disease on an existing MRI?

Simpa Salami: I think that there are areas in the prostate where sometimes there's actually high-grade disease in the prostate that the MRI does not pick up. This can be due to a number of factors. It could be due to the quality of the MRI or the expertise or experience of the radiologist interpreting the MRI, and so if we were to simply rely on the MRI alone, those patients may go with aggressive disease undetected. So what my findings have indicated is that the visibility or invisibility of MRI is dependent on how disorganized the cells are.

Charles Ryan: I see.

Simpa Salami: It is very closely related to the grade of the disease, and as we know, the more aggressive a cancer is, the more likely it is to be visible on MRI. What my findings have shown is that genes that are involved in cell organization and structure under expressed in lesions that are visible on MRI, in lesions that are invisible are particularly more well-organized.

Charles Ryan: So they have more normal appearing cellular architecture, which enhances, is this true, enhances the magnetic resonance and produces the image.

Simpa Salami: That is correct.

Charles Ryan: Okay. So is that a biologically different tumor from one that has down-regulated or downward expression of these genes that is visible on MRI? In other words, if you looked at a Gleason 8 that's positive on an MRI and a Gleason 8 that's not positive on an MRI, are you suggesting that on the genomic level, these are different tumors essentially?

Simpa Salami: When you take grade for grade, a Gleason 7 prostate cancer that is visible and Gleason 7 prostate cancer that is invisible on MRI, that there's no difference in the biological trajectory of those tumors, suggesting the fact that you do not see a cancer on MRI does not mean that it is biologically indolent.

Charles Ryan: Interesting. So let's call it the MRI invisibility is not a marker of even higher-grade disease or anything like that, it's just a separate factor in the cancer, but an important one to know about in terms of our reliance on the MRI.

Simpa Salami: That's true.

Charles Ryan: Are there any other genomic differences? Is it only these cellular architecture genes, or are there any other androgen receptor, PSA, other things like that that might show up on a genomic analysis?

Simpa Salami: Particularly the approach that we use, we are able to approximate or derive the commercially available tissue-based prognostic biomarkers like Polaris, Oncotype DX, and Decipher, and what we found was that lesions that are visible and lesions that are invisible, there's essentially no difference in their scores. 

When we sub-stratify the Oncotype DX score into the different modules, the androgen receptor module, the cellular organization module, and the stromal module, what we found was that the only difference was in the cellular organization module, again supporting the fact that the reason why some of these lesions are visible is because of the underexpression of the cellular organization genes.

Charles Ryan: So how should the urologists who are listening use your data in their day-to-day practice? What should they take home from this?

Simpa Salami: I think that the take-home message is that we should not rely on MRI alone. I strongly believe that we should use MRI to risk stratify patients. I think that every patient if possible should get an MRI before biopsy, and in patients with negative MRI, we should not simply forego biopsy. I think that we need to continue to look for strategies or measures that can help us to detect those lesions that are invisible on MRI.

Charles Ryan: I want to go back and have you just describe technically how you did your study. Did you have MRIs performed on patients with known prostate cancer and then follow that with a biopsy? Or how did you get your data of genomics in MRI-visible versus invisible prostates?

Simpa Salami: We have three different cohorts. We collaborated with Cedars-Sinai as well as Genome DX. The first cohort was from the University of Michigan. We assembled 10 patients who had undergone radical prostatectomy. We compared the radical prostatectomy specimen with the pre-surgery MRI, so we really can have a true definition of what was visible and invisible on the MRI.

What we then did was to acquire tissue from the visible and invisible lesions. We subjected those tissue to DNA and RNA next-generation sequencing, and then based on the sequencing data, we were able to develop a model that is based on nine genes. This model had an area under the curve of 0.894 predicting MRI visibility. To evaluate the performance of this model in predicting MRI visibility, we collaborated with investigators from Cedars-Sinai who also had 16 patients with MRI visible and invisible lesions, and we were able to predict MRI visibility using our model in their cohort with an area under the curve of 0.88 and a specificity of 100%, indicating that the model that we had developed has a fair degree of accuracy.

We then took that model and applied that to a large data set from Genome DX. They had sequencing data available from patients who had radical prostatectomy and long-term follow-up. They had biochemical recurrence, distant metastasis-free survival, and overall survival ... excuse me, prostate cancer-specific survival, and what we found was that between lesions that were predicted to be visible and invisible, there was no difference in each of the three different outcome measures that we looked at.

Charles Ryan: Interesting, interesting. So a lot of patients and people being worked up for prostate cancer watch our site and look at some of these videos, and what would be the message you would send to patients regarding these findings about the limitations of MRI in light of what they may be going through as they make decisions for their own care?

Simpa Salami: The message is that MRI is great in helping to detect aggressive prostate cancer. It has about a 90% specificity. However, in patients with negative MRI, about 10 to 15% of them will have aggressive prostate cancer, and so if there are other indicators of a disease, for example, if they were to have a rising PSA or abnormal DRE, those should be indication to do a systematic or a saturation biopsy to find out where the disease is in the prostate.

Charles Ryan: So should a patient who is on active surveillance and may have a relatively indolent appearing MRI, should patients have repeat MRIs over time who are on active surveillance.

Simpa Salami: There is a study from the NIH that indicated that really you don't see a lot of changes in MRI in patients on active surveillance. The time period between the MRIs that were acquired was I think about two years. It is unlikely that you see a lot of changes, but I would believe that combining MRI and periodic biopsies to determine what to monitor patients on active surveillance will be useful.

Charles Ryan: Yeah. Is there any way that you will be able to look at your patients who had MRI invisible prostate cancer, we'll call it that, who have subsequent MRIs to see if those areas become more visible? Does this evolve over time?

Simpa Salami: It is unclear at this time. I do not have the data to be able to figure that out.

Charles Ryan: Well, there's your next trial.

Simpa Salami: That's correct.

Charles Ryan: Well, it's a pleasure talking to you. I've learned a lot, actually. This is quite an interesting area to be learning about, not only the genomics of prostate cancer, but MRI and how they come together, and how this can help patients, so congratulations on your presentation and a pleasure talking to you.

Simpa Salami: Thank you so much. Thanks for your time.