High-Low Frequency Ultrasound Vs. MRI for Prostate Cancer Diagnosis - James S. Wysock

December 16, 2021

In this 2021 Large Urology Group Practice Association Annual Meeting presentation James Wysock discusses imaging of the prostate, high and low-frequency ultrasound versus MRI, and where this area is going in the landscape of prostate cancer management specifically how to improve our screening and how to select men for treatment.


James S. Wysock, MD, MS, Associate Professor, NYU Langone Health, Chief, Bellevue Urology Service, New York City, New York

Discussant: David S. Morris, MD, FACS, Robotic/laparoscopic urologic oncology, Endourology and stone disease, Reconstructive urology, Urology Associates P.C., Nashville, TN

Read the Full Video Transcript

James Wysock: Thanks Dr. Shore. So this is the hot seat presentation. I'm James Wysock from NYU. And I just want to say, thanks for having me here. What we're going to talk about is a little bit about imaging of the prostate high and low frequency ultrasound versus MRI, and to try to keep this moving and to keep it compelling. We'll really focus on where this area is going, but nothing to disclose. So, prostate cancer management has been debated in terms of how to improve our screening and how to select men for treatment. And one of the major impacts that we can have is in identifying the men that would benefit from treatment. So, the imaging of the prostate's really a critical component to that. And a lot of that really relies upon tissue evaluation through an ultrasound guided biopsy.

And as you can see here, ultrasound imaging on a low frequency really does have significant characteristics, which hamper our ability to make appropriate diagnoses. In fact, prostate cancer is the only solid tumor that we diagnose through a random biopsy technique, and that does remain the standard of care. And as we all know, this is the workhorse for our prostate cancer diagnostic paradigm. They're over a million prostate biopsies or about a million performed in the United States every year. And cancer detection rate appears to be steady despite continued added biopsy specimens in the blind biopsy paradigm.

So we're all familiar with this, and we're all familiar with the limitations of the transrectal ultrasound guided biopsy and the low frequency ultrasound has poor image resolution. Can't really see the midline very well. There can be calcifications that block the ultrasound signal, hyper AOIC and hypo COIC tissue characteristics could represent disease, maybe completely normal in a nodular prostate. The false negative rate is unacceptably too high around 30. And even when we do diagnose a tumor, the under grading rate on the blind biopsy is also unacceptably too high. And finally, just putting in 12 biopsy cores has contributed to the over detection of low grade indolent disease, which led to the major debate that took place with PSA screening downgrading.

You know, this was really part of what is the impetus for Dr. Walsh's comment, which I think serves as a bellwether for the development of MRI. And we can debate this and talk about it a little bit later, but we really do need and did need at the time, something more accurate when looking at the prostate. And so enter MRI, this is an axial T2 image, and I think it's pretty clear. You could see that the peripheral zone of this man's prostate. One of my patients, there's a signal drop out in the left mid gland medial portion of the prostate. That's pretty clear, if I gave you that image and said, "okay, why don't we plan to biopsy it? It's about five millimeters off the midline, mid gland." You could approximate where to put some cores and you might actually feel pretty confident that you biopsy that area. Ultimately, did that work?

Well let's take a look, but beyond just that axial T2 image, which gives you anatomical information, gives you a prostate volume assessment. That's probably more accurate than some of the transrectal ultrasound, and also doesn't involve putting the patient through an invasive procedure. So you can get an MRI, understand how big the gland is, also risk assessed. It has two other components in the multi-parametric MRI that give us information as to the tissue characteristics and potentially biomarker capability, if you think about it that way. So first is, diffusion weighted imaging and in diffusion weighted imaging, we're really assessing cellular density. It's an assessment of the restriction of the brownie in motion of protons within the tissue, which is essentially what the MRI is measuring in the first place, but a more densely packed part of the prostate will have restricted diffusion. I like to think of that almost as a digital rectal exam and sort of feeling a nodule.

So hypercellular regions will restrict the protons, and that goes together into what we call the apparent diffusion coefficient or the ADC image, which is the lower image here on the screen, which is very dark. And so that helps correlate with T2 images is that might have some suspicion as well. And the final component of the multi-parametric MRI is dynamic contrast enhancement. And so we all know that as a tumor grows, it needs blood supply. And so if you give a contrast enhanced or a contrast agent, it'll rapidly be taken up in an area that more blood vessels and then wash out, and that correlates nicely with angiogenesis, put all that together into the PIRADS score, and we get our multi-parametric MRI. Now the promise trial really is in my opinion, one of the most important imaging studies in prostate cancer in the last 20, 30 years, probably. Level 1b evidence, but what's really important about the promise trial is that it was 11 centers, not all of which were expert centers in prostate imaging.

They all used 1.5 Tesla MRI, which we would feel is not the best MRI study. And they showed that around 25% of the men that were in the study had negative MRIs. Those negative MRIs had about a 2% risk of finding Gleason grade group two or higher. And that was on the five millimeter trans peroneal mapping biopsy. So I think that's a very good endpoint. When you think about MRI. First of all, about a quarter of the men who had an MRI avoided biopsy, that helps address the concerns of over detection of indolent disease and finally, any significant disease you should then be able to target, which leads us to the precision trial, which showed that if you targeted those MRI findings, you enriched your diagnosis in the detection of clinically significant cancer, and you decreased the detection of the indolent disease. So, those pieces of information that the MRI provide, really do work to address some of the shortcomings of the ultrasound with low frequency ultrasound biopsy, which I had mentioned earlier.

So a pre biopsy MRI, you get a picture of the MRI. You can sit down with a man say, here's your prostate, here's the volume, but it also tells you some benign features that can help guide treatment decisions such as, how big is the gland? What's the PSA density? Is there a median lobe? And you can start even talking about other aspects of their prostate health beyond risk prostate cancer. It helps you identify targets for biopsy that can be achieved through transperineal, transrectal. Anterior lesions can be seen better that may have not been seen or detected in the past. Posterior midline lesions can be detected as I had shown in that MRI earlier, and it has opened the door to the target biopsies, and finally it opens and sets the stage for or targeted therapy.

So look, that's all well and good, but the hot seat here is what if we had better ultrasound would better ultrasound be potentially a game changer as well? So there is transrectal micro. Most of the ultrasound we use is around 6 to 12 megahertz, right? And that's what we're all used to looking at. But 29 megahertz ultrasound is what we call micro ultrasound. And as we know with ultrasound technology, you increase the frequency, you increase tissue resolution. And here is a picture up, hit the upper right here of a micro ultrasound image of the prostate and outlined in red is an area that is consistent with a prostate tumor, 300% improvement in tissue resolution, four times the density of ultrasound crystals and the ExactVu company provides you a PRIMUS system. And that's essentially a scoring system to the PIRADS score. You can use this as similar to a liker score as the PRIMUS system goes up, so does your risk, your relative risk of having significant prostate cancer in those regions.

Now here's a slide that the ExactVu company provided me, which compares low frequency ultrasound on the left middle image set is 29 micro ultrasound. And then on the right is the multi-parametric MRI. Now, in this particular case, it's a Gleason grade group three prostate cancer. It's in the patient's left sided peripheral zone, and it's best seen on the micro ultrasound and it was not seen on the MRI. The purpose of this is to demonstrate that the conventional ultrasound or the low frequency ultrasound doesn't really show you the same degree of tissue heterogeneity and characteristics that the micro ultrasound can provide. So micro ultrasound versus low frequency ultrasound, seems that it would be a pretty straightforward comparison, and it essentially was. This is a study looking at disease detection between the two techniques and they showed that the detection of significant cancer was improved with micro ultrasound versus low frequency ultrasound.

What I'd like to point out, however, is that it's primarily showing that it's improved in the regions between grade group two and grade group three, which is where you would expect to see the biggest difference. Ultimately the grade group fours and fives are fairly large, usually, and more obvious, even on the low frequency ultrasound. And Laurence Klotz published this review of a small meta analysis, essentially meta analysis of micro ultrasound data. And he put out the compelling title. Can you actually replace the MRI and I think that's what something we can talk about in a bit, but essentially summarizing the data that's out there on micro ultrasound, so far, the test characteristics may even look a little bit better than what was shown in promise trial.

Now, what about combining micro ultrasound with an MRI paradigm? And that, I think a very interesting avenue for exploration. These are a couple studies. One thing we do know from using low frequency ultrasound at the time of an MRI guided biopsy is that if you see a hypoechoic lesion in the region of the MRI, finding at the time of a fusion or to targeted biopsy, it actually predicts and correlates that you'll find disease there.

So, when you see that it's actually a nice piece of feedback that you're targeting is good. There are many registration errors that can occur when you do an MRI ultrasound fusion biopsy. But if you get a hypo correlate and say, "Hey, that's where it is." Your confidence improves significantly. So we've shown that at NYU and our data as well, but what this study looked at was 159 men who had both the MRI fusion biopsy with low frequency, but also with micro ultrasound and what the bar graph shows there is as you increase the modalities and you have findings on the micro ultrasound that correlate you actually continue to improve the detection of the significant disease versus the lower rates of insignificant disease.

So I think it does speak to if you had a tool with ultrasound that could give you more confidence in where you're looking and whether it correlates with your MRI target or not, you will improve your diagnostic paradigm. And how about in the equivocal lesion? This is an interesting space. You get an MRI, you have a PIRADS 3, you may have a high 4k score or a high PSA density and you put the ultrasound probe on, do your fusion biopsy. And the ultrasound shows you no hypoechoic lesion. So you start to worry, "Hey, am I registered correctly? Are there numerous points at which I'm making a targeting error?" If you use a micro ultrasound in this environment, what they showed was that you actually improved your detection of cancer in those settings, but further you could risk stratify men and potentially avoid a biopsy. So there was interestingly around a 25% rate in this setting where the micro ultrasound was negative. And I just highlight that, that goes back to the same rate of negative MRI in the promised trial, and also the same rate of negative MRI in the precision trial.

So it's an interesting corollary there, but it's another area where micro ultrasound may have some significant advantages. So micro ultrasound fits into our workflow paradigm nicely. If we had an ultrasound that we could rely upon, that is, as I mentioned earlier, the workhorse for the prostate cancer paradigm, in terms of diagnosis. So I think there is some real value to that. It does take the MRI out of the picture. In some concepts, you might be able to say, "look, you have an elevated PSA. We'll just go straight to micro ultrasound." And it simplifies some of the recommendations. So you don't have to have that intermediate step.

So it's a hot seat. So we'll turn it over, maybe some questions at this point, but I think that one of the end points I'd like to leave you with, is that MRI gives us a lot more information in the micro ultrasound, despite micro ultrasounds value, where I think the combination would be, or where I think the future would go as sort of a combination, but I wouldn't want to lose that three dimensional component that the MRI can provide.

What the image on the right there is. Basically a three dimensional model that I developed with one of our researchers where the tumor is in green, but we've actually modeled the cryo probes based on the known isotherms that the cryoablation probes can provide. And we can then develop a three dimensional treatment plan for a patient. Micro ultrasound can help us make sure we're looking in the right place, but it can't really help us at that stage. So I think that's something to keep in mind. And then on the bottom there, there's some MRI imaging follow up after an of a partial gland ablation. So micro ultrasound has some components there that I think will help and as we move forward, coupling it with MRI, I think is the way to go. Thank you.

David Morris Discussant: Thank you very much. I figured I'd start quickly with some audience participation groups that currently are doing MRI fusion biopsies. Can I see just kind of a show of hands? That's kind of what I expected near a hundred percent. Anybody here doing micro ultrasound? We have a handful. Okay. So just a couple as an outsider, not doing micro ultrasound, but certainly doing MRI a long time. What's the reproducibility for the Primus scoring? Has there been any studies looking at kind of inter observer or variability on that?

James Wysock:  Yeah, I think that we don't have a lot of data on exactly inter observer of variability, but what I will say is that the study that Pavlovich in BJUI showed a very interesting phenomenon with the Primus scoring. So they look looked at five centers and were comparing micro ultrasound versus low frequency ultrasound. And what they demonstrated was about midway through the study, they had to retrain a lot of the users and kind of reorient the primus. So I think there's still a significant learning curve on the primus, and that can impact the accuracy of the test. But look, we're all used to low frequency ultrasound, so it's going to be a learning curve. And if we were all adopting it and using it much more frequently, we'll get much better at it. And that speaks to something. Then eventually you could get to inter observer variability, you've got to get in a level of expertise before you can get to that point. And so I think that there is a challenge there. I know that we are interested in that, developing that and studying that going forward.

David Morris Discussant: I honestly, I think most of the students need to get better at low frequency ultrasound basically anyway, but we know where the prostate is. We're not really looking for targeted lesions. So, well next, just quick question, in terms of guidelines, they've really accepted MRI and starting them as part of AS programs. Do you see that coming down the road potentially for micro ultrasound? Is it something for surveillance where you think it might eventually incorporate? Because I'd like to see that if we're going to try to adopt it and justify the cost.

James Wysock: Correct. Yeah. So look, why do we incorporate MRI into the guidelines? I mean, the guidelines are really aimed at trying to identify who's going to benefit from treatment and who's going to be best put on surveillance. Why is there a problem there? Well the low frequency ultrasound misdiagnosis men. So MRI is a way to try to ensure that you are getting the correct diagnosis. You're grading and staging the man correctly. If micro ultrasound data falls into that category, there's no reason why it wouldn't be part of that. How long will it be before it's in? I mean, MRI's still not even fully in the guidelines. So it's a while.

Tom Reichshof: Tom Reichshof with IMP in New York. So I just want to amplify that point too, is that, being in New York, patients go to a lot of different places for their MRIs. And we've had to, in the beginning, we had to study, we looked at the radiology center and our positive rate and we actually saw a big variability. NYU is one of the places we'll use that's approved by us, but really we have to keep it just a few Radiologists because a lot of Radiologists will read stuff and they think they know what they're doing and they're completely off. We get those reread. And then it becomes also with micro ultrasound, how many of us have trained with micro ultrasound? Probably zero. So, I would say, and from our experience with four of us do all our biopsies and we each done over a thousand fusion biopsies. It's just past our seventh anniversary of doing fusion biopsies.

So, we have a lot of experience. We quality the hell out of the whole process from the radiology on. So I would say that for bigger practices, you might want to find a few people who do a lot of them, as opposed to any Urologists can do a prostate biopsy, only if they can. And make sure the radiology is up to snuff, and look at your numbers because if you're not getting good positive rates from a certain center, either they need to learn more or you need to use somebody else. You have NYU, great people in the radiology, you have some years done tons. And so you have a very ideal situation for this study, but I think in the community might be a little bit different.

James Wysock: Well, absolutely. That became right apparent when we started to adopt about 10 years ago, the targeted biopsy strategy, right? So at that time, we did start to restrict who was doing it. What we've seen, I think is an interesting phenomenon is, you get Radiologists who become experts in it and then they may leave and then they may evolve into other roles. And then you bring in new Radiologists who, each time that happens, you have to essentially reassess your data. And it is a QI project really, and quality control project.

And it is challenge because we'll go through waves where we're like, "well, look, these are some significant misses on the MRI based off what we detected on the biopsy results." And so that's a challenge, but promised trials is still encouraging there. Like I said, it's 1.5 Tesla, 11 centers, and they're not all experts. And they still showed a significant improvement over just the blind biopsy. So I would still not abandon it. Look at your own MRIs too, right? I mean, hopefully people who are ordering it, are looking at, you gain a lot of information by looking at your MRIs, seeing the prostate, seeing the anatomy, seeing what the volume looks like, see where the lobe is. It helps more than just the biopsy.

email news signup