Read the Full Video Transcript

Charles Ryan: Hello. Today I'm joined by Dr. Simpa Salami, who is an Assistant Professor of Urology at The University of Michigan. Dr. Salami has been working on the development of a urine-based multiplexed assay for the early detection of aggressive prostate cancer, and I had the pleasure of speaking with him a couple of years ago about MRI-based techniques for identifying prostate cancer, and now he's taken this work a little bit further, a lot further, and their findings are now published in European Urology. Simpa, great to see you again, and thank you for joining us.

Simpa Salami: Great to see you, Dr. Ryan. I'm very excited to be here with you to discuss our work today.

Charles Ryan: Yeah, it's really exciting, and I think it gets to a significant clinical need. Tell us a little bit about, before we talk about your new assay, the background challenge of detecting aggressive prostate cancer.

Simpa Salami: In general, prostate cancer is a very heterogeneous disease. It's multifocal. It's been estimated that prostate cancer often occurs in multiple areas of the prostate in about 60% to 80% of patients. And the strategy for detecting aggressive prostate cancer based upon elevated PSA or an abnormal prostate exam is to do a prostate biopsy, more recently, one would get an MRI to identify regions of interest that are targeted. And in addition to that, sample the rest of the prostate. But even then, we still missed about 10% to 15% of high-grade prostate cancers. So I think that there is a need to develop tools that can really capture all of the events that are happening in the entire prostate.

Charles Ryan: And before we talk about the urine-based approach that you are developing, the sequence normally for a man who is considered to be at risk for prostate cancer is to do a biopsy and then an MRI, or an MRI than a biopsy. What's the sort of standard sequence that is happening now in the world?

Simpa Salami: About five to 10 years ago, the standard sequence is, a patient has an elevated PSA, you proceed straight to a standard 12 core random biopsy under ultrasound guidance. But more recently, MRI has been shown to improve the detection of aggressive prostate cancer. But of course, the technology is still evolving, especially in community practices, such that it is not necessarily the standard of care in most places in the United States.

The other issue that we have encountered clinically is insurance coverage, where it may not be covered, in which case you still will have to proceed to a prostate biopsy, and only if it is negative would you get insurance companies to cover an MRI in that situation?

Charles Ryan: So it's interesting because you have a situation where, I don't know what the cost of a biopsy versus an MRI is, but I can't imagine that either one is particularly inexpensive and that you have a number of patients, or actually they're not patients if they haven't been diagnosed yet, a number of subjects who are essentially going to go through both procedures.

Simpa Salami: That is correct.

Charles Ryan: Okay. So if only we had something we could test in the urine that might help us along. So tell us about the evolution of your assay and how it got to be to the point where it is.

Simpa Salami: Perhaps I should start out by congratulating the team. This is really a team effort, a work that was led by Andi Cani, one of our, at the time, was a Ph.D. student and now a post-doc in one of our labs, and also Dr. Scott Tomlins, who really was my mentor and who supervised this work early on before he moved on to his company. To say, this work really has been an effort that has been led by multiple people, and I'm privileged to represent them today.

This work builds upon earlier work by [inaudible 00:03:44] and Scott Tomlins and others where they developed a urine-based assay that incorporates serum PSA with urine PCA3 and TMPRSS2:ERG in a test called the Michigan Prostate Score. And more recently, this has been rebranded as the My Prostate Score or MPS.

What this assay does is help detect aggressive prostate cancer. So our current work builds upon that because we believe that, given the heterogeneity of prostate cancer, that there are some molecular alterations that may be different from patient to patient and increasing the capacity to detect these alterations can also improve our ability to pick up high-grade prostate cancer. So we started out by looking at our prior work in prostate cancer tissue. We identified 306 transcripts or alterations that we think are relevant in prostate cancer, and based on that, we narrowed down our initial pool of transcripts to 84 transcripts that we can reliably measure in urine. And then we assembled 109 patients with high-grade prostate cancer versus benign or low-grade prostate cancer as a cohort for both development and validation of this assay.

Charles Ryan: So it's essentially an 84 gene score. Is that correct? 84 genes that you are looking at in the urine transcripts?

Simpa Salami: Correct. We looked at 84 transcripts. So the gene pool is about 60 because there are some genes that we've targeted with multiple transcripts, and then the final score or model that we developed to predict hybrid prostate cancer is comprised of 15 transcripts.

Charles Ryan: Okay. And your definition of high-grade prostate cancer or your definition of aggressive prostate cancer, I should say?

Simpa Salami: It is Grade Group 3 prostate cancer, or Gleason 4 + 3 and higher. I should also point out that because this was the initial development of this assay, we removed patients with Grade Group 2 prostate cancer, such that we compared Grade Group 3 and higher versus Grade Group 1 and benign so that we can pick up extremes of differences in transcripts expression.

Charles Ryan: And are the transcripts, if we dive in deep as to what are the genes that you are looking at, are the genes that are known to be associated with the biology of more aggressive prostate cancer?

Simpa Salami: That is correct. 60% of the transcripts that we have in our assay are TMPRSS2:ERG fusion isoforms. We also have PCA3, we have SHARP1, we have TDRD1, we have HOXB13 and DRH1, and ERG, which have been shown in some of the urine assays to be important in predicting aggressive prostate cancer.

Charles Ryan: And in predicting a higher Gleason score, poor clinical outcome, I guess those all kind of are in one sort of package, right? In terms of, when a urologist says aggressive prostate cancer, what we're really referring to is the likelihood of recurrence after local therapy of some kind or another, correct?

Simpa Salami: That is correct. Yes. I think a Gleason score really is our short-term endpoint, if you will, for determining aggressive prostate cancer.

Charles Ryan: Yeah. I guess that's what I'm getting at, which is, you can study the biology of the disease and its relation to a high Gleason score, which a high Gleason score is your surrogate marker, really, for aggressive disease, as opposed to doing a study, waiting 15 years, and seeing who recurs and how that bears out in terms of the actual clinical aggressiveness in those particular cases.

Simpa Salami: That is true.

Charles Ryan: Yeah. So tell us a little bit about the sensitivity and the ability of this urine assay to essentially add to the pre-existing models.

Simpa Salami: So the preexisting model that we have is the Michigan Prostate Score that comprises serum PSA with TMPRSS2:ERG and PCA3. And our goal is to be able to improve upon the performance of that model. So in the current study, the performance of that model in predicting high-grade prostate cancer was an AUC of 0.72. And then the current model has an AUC in the training cohort of 0.9 and in the validation cohort of 0.82. So we thought that was a pretty good performance, but again, we have to keep in mind that this is the initial development, it's a biased cohort because we eliminated patients with Grade Group 2 disease, so it would be nice to validate this assay in a prospectively collected cohort.

Charles Ryan: Right, right. Very good. And this is really nicely displayed in the paper, the training set and the validation set, and the AUC. And if I'm reading this correctly, what it suggests is you can compare the validation set, for example, to serum PSA, where serum PSA has an AUC of 0.69 in your data and the year-end UPSeq has 0.82. So quite a bit better than just a PSA and really better than all of the other tests that you've done, including the Michigan Prostate Score in your previous studies.

Simpa Salami: Correct. Yeah. So this assay really has great potential, I think.

Charles Ryan: So what do you tell the patient out there who has been worked up for a high PSA or the urologist out there, or another practitioner, maybe the primary care doctor who's wondering, is this going to help solve the dilemma of detecting and diagnosing and routing for the correct treatment, patients with high-grade disease?

Simpa Salami: I would tell them that this assay is promising. Certainly, a lot of work still needs to be done, but when we compare where we are today to where we were 20 years ago, I think we've made significant improvement. And really with evolving technology, both in terms of techniques, tools, and methods while analyzing data, especially using machine learning, I think that we are making progress. I see this assay as something that will be very useful in helping improve patient selection, patients who need to undergo their first biopsy or repeat biopsy. Another potential application of this assay will be in patients who are considering active surveillance to determine whether active surveillance will be the way to go. And certainly, even for monitoring patients on active surveillance for prostate cancer, you will not necessarily have to biopsy a patient every year or every other year if we are able to demonstrate prospectively that this assay can help identify patients that need to undergo biopsy.

Charles Ryan: And so do you have ongoing studies that are taking this test to the next level?

Simpa Salami: Yes, we do. We currently have an NIH SPORE project in prostate cancer where the goal is to enroll patients on active surveillance for prostate cancer. Urine has been collected every six months after their prostate exam. At this, patients will have biopsies in their second and their fourth year, and then we will be able to analyze the capacity of this assay to detect high-grade disease in that patient population.

Another interest in the application of this assay that we are also studying in that project is in patients who have not been diagnosed with prostate cancer, but they have alterations in their genome that put them at risk of developing prostate cancer, like BRCA1 or BRCA2 or the Lynch syndrome mutations. So if these patients have no other risk factor for prostate cancer, other than having this genetic mutation, we are collecting urine from them, and we're also testing this assay and its capacity to predict the presence of prostate cancer.

Charles Ryan: So you have an active surveillance cohort study where patients have already had their biopsy, and they have elected to undergo active surveillance, and you have a high-risk familial genetic cohort, perhaps. And is that accruing as well now, currently, or is that planned?

Simpa Salami: That is accruing currently. Actually, one of our med students, [Michael Sasson 00:11:49], who led that work, will be starting his urology residency on Wednesday, he recently published the initial report, 45 patients in The Gold Journal, just the clinical characteristics of the first 45 patients. And the next step is to run this assay in the urine that is collected from that cohort.

Charles Ryan: Oh, that's a pretty good medical student project. So congratulations on this work. It's a great paper. It's a really interesting technological step towards solving a really important problem, which I know urologists face day in and day out, which is trying to thread that needle between detecting the aggressive disease that needs the aggressive treatment and not over-treating the nonaggressive disease that doesn't. So we welcome this new technology. I wish you the best in your studies moving forward to try to validate and export this interesting technology. Thank you for joining us.

Simpa Salami: Thank you very much, and thanks for having me, and I look forward to talking to you again soon.

Charles Ryan: All right. Me too. Dr. Simpa Salami from The University of Michigan.