The Modern Role of Genomics & Imaging in Prostate Cancer Diagnosis & Treatment

[Dr. Aditya Bagrodia]
I think when we talk about surveillance, it's hard not to make it a little bit of a PSA screening talk in diagnostics, but when patients are coming in, let's just say, elevated PSA referral, in your ideal world, what transpires between an elevated PSA and say, for instance, a biopsy?

[Dr. Ashley Ross]
That's actually a great point and a great starting point. I think that where this dovetails in, as you were kind of leading to, is we're trying to avoid the diagnosis of what would have been classified as NCCN, very low-risk prostate cancer, or let's just say, small amounts of low-grade prostate cancer. In many ways, we're trying to avoid that entirely. That's a disease that obviously should only be surveyed as a management option, but it'd be better just not to diagnose this and not to give patients anxiety over it.

How do you avoid that? The AOA just came out with their new guideline statement, and they maybe weren't firm enough. I mean, honestly, in every situation possible, you should be doing something else between the PSA and a biopsy, and that something else should eventually lead you to an MRI prior to the biopsy, as long as the gentleman doesn't have some contraindication, like a pacemaker that's not compatible.

At my institution, we have access to the Prostate Health Index. Not everybody has that, but as a biomarker, I tend to like it. It has a much better sensitivity and specificity than PSA alone, and it's dirt cheap, so it's really of low cost. If you don't have that, there's a bunch of other tests on the market that can help sort of determine, is this PSA due to a large prostate inflammation, or is it really prostate cancer-related? Now, you may or may not take that step of this interim biomarker. For me, like I said, Prostate Health Index, under 27, I don't do anything else. I just watch the guy. You know, you could use percent-free. Percent-free over 25 is a good sign. However, that only captures a small fraction of the population. Other people use things like MIPS or 4K score or other tests out there, including exome and DX, but then I think that you want to do a MRI.

Now, MRI is where there's really, like quite frankly, level one evidence that it's going to help us, and it's going to help us by reducing biopsies by over 30%, maybe even up to 50%, and if we do a biopsy, it's going to help us with its positive predictive value and targeting. In my shop, I tend to do all my biopsies with MRI guidance. Essentially, ideally, the patient would have an elevated PSA, maybe an interim test, but then go to an MRI. If the MRI is clean, PI-RADS 1 or 2, defer a biopsy. There's a lot of people with a PI-RADS 3 score that could defer the biopsy. Then some PI-RADS 3, most of your PI-RADS 4 and 5, because remember, we should only be PSA testing people who we think need it, like 10-year survival, et cetera, you're going to want to do a biopsy to figure out what's going on.

I think that that, to your point, sorry to be long-winded, you're going to limit the detection of clinically insignificant disease, and you're not going to miss clinically significant disease or just miss a small fraction of the men that might have clinically significant disease that you'll probably pick up the next year as you follow them.

[Dr. Aditya Bagrodia]
That's perfect. I think it's going to get increasingly more interesting and complex as some of these secondary prostate cancer screening tests and their results, which maybe led to an MRI and a diagnosis of an indolent prostate cancer, may factor into all of the bits of information, which of course, we're going to cover here extensively as we think about surveillance, surveillance regimens, and counselings for the likelihood of coming off of surveillance. I think bottom line, ideal world, PSA, MRI, MRI ultrasound fusion biopsy, whether that's transperineal, transrectal, cognitive, or actual fusion, in-bore, that's kind of the best case scenario, and I would agree.

[Dr. Ashley Ross]
I should note, there's one thing, you know, people have asked me, what's the limitation? What's keeping us from that? I think that there's a few major things, and I know it's not the topic of this podcast, but since we have a listening audience, and something that I think I have to double down on myself, now that I really see the problem: MRIs have a few problems in prostate land. They can be expensive. They can be hard on the patient because they're about an hour if you do with contrast. There can be access issues, and again, with contrast, that makes it an hour. Then finally, there can be quality issues, where the quality MRI is not good enough for a fusion, the quality of the read or the MRI sequences itself is not good enough for an adequate interpretation, or the reader is not qualified.

I think this year, I always try to have at least one or two missions in my back pockets, something I want to push forward. I think this year, one of those missions is really lobbying, whichever way we can, right now, it's a grassroots thing through your BackTable urology, that the AUA get together with the American College of Radiology, and say, you know what, we need a way to certify readers, certify centers, to do prostate MRIs, so we can have some consistency. Then the second part is, I think we really need to move towards bi-parametric MRI like Europe has done for screening. That means drop the contrast. That way, there's no nurse needed in the MRI suite. Some places require radiologists there when you have the contrast from contrast reactions. No nurse, 10-minute scan, easy in, easy out, increases the capacity thus by about fivefold. I know that's off topic, but I think that that's one of the things that we should all be thinking about in our head is, do we really need to be doing triphasic multi-parametric MRI?

[Dr. Aditya Bagrodia]
I think everything you mentioned, and on top of that, I'm not young, and I'm not old, but for practices that may be a little bit more mature, getting an MRI for a PSA, interpreting that, having the infrastructure to actually act on that is not a foregone conclusion. Now you're taking a bread-and-butter urologic diagnosis, like an elevated PSA, and you're referring that out. I don't mean to sound like people have nefarious interests there, it's just adoption of a new technology. You and I might not think of it that way, but it could be for the other dude, just like doing Haifu or robots, et cetera. I guess we talked about the ideal, and then there's going to be a very real patient who comes in with a 12-core biopsy and they've got some type of prostate cancer.

[Dr. Ashley Ross]
Right.

[Dr. Aditya Bagrodia]
You know, it's that exact scenario. Maybe we'll run through some common clinical scenarios to frame this a little bit, but just a couple of thoughts. I mean, absolutely. Sometimes it takes somebody, something additional to convey the aggressiveness of this and something objective that's not read by a pathologist or sampling error by a urologist, et cetera, like a genomic classifier can be useful. On the flip side, I have these highly educated, super healthy people in La Jolla that are like, "Doc, it's cancer. I want it out," and I'm trying to walk them off the ledge and saying, “hey, Mr. So-and-so, I understand your concerns and this is going to be a dynamic process for the first time you heard cancer, you're freaking out and you don't want to die, but over time, you're going to understand the quality of life implications, et cetera, but let's be deliberate. Let's get some additional information and maybe this will help ease some of your fears.” I've found that if I'm going to get a genomic classifier, that it's incredibly helpful to run through what the output is going to look like and actually run through what my thresholds are to do something. Otherwise, it becomes extremely abstract and confusing to try to sit with the man and say, here's your likelihood of metastasizing, dying, et cetera. Can you talk to us a little bit about how that works for you?

[Dr. Ashley Ross]
Honestly, as I mentioned, I use Decipher as my genomic classifier. I don't usually even show them the report. I mean, there's some rich information on there, but they just have to understand at a high level, you know? What I tell them is if they're Gleason grade group 1, I tell them that what we're looking for is really, you're a good candidate for surveillance. If that Decipher comes back low risk, I tell them it's going to come back low, intermediate, or high. If it comes back low risk, we know you should stay on surveillance and maybe we can actually space out your biopsies. If it comes back high risk, you're in this minority population and we have to really be hawkish if we're going to follow you. Most of my Gleason grade group 1 guys, I follow. I tell them that, look, you will likely progress, but I usually will follow them up to a year. For some of them, that's the guys that I actually take to radical: the rare people in my practice with Gleason grade group 1 or the Decipher high Gleason grade group 1, and usually, they have one other yellow flag. An intermediate risk, you know, is middle of the road. If I look at the Northwestern data and I look at what's the frequency of people I would never want to follow, the pathology at radical is 4, 3, 7, or T3b. For intermediate Decipher, then, I mean, low-intermediate than high, it goes from about 15% to 30% to 40%, so intermediate actually has some real risk. Regardless, I've already decided to myself, if this guy has a low-risk Decipher, and they're even up to Gleason grade group 2, favorable intermediate-risk disease, I'm going to tell them that surveillance is a really viable option. If they have low-risk cancer, like Gleason grade group 1, and the Decipher is low, then basically, you shouldn't do this with the data, but you can do this with the sort of talking head stuff, so I pull out whatever I need to do. Like, I tell them, look, you got low-grade disease. We found it so early, it probably doesn't want treatment, and I tell them one of two things. What I used to tell them is, look, I took an oath, and the oath was to do no harm. Me doing anything to you right now, when I have no idea if your cancer is ever going to get worse, is doing harm. I can guarantee you it's going to harm your sexual function in some way. Even my people who have "perfect erections," if I really drill down on it, that is it exactly like it was before surgery, they'll say, well, no, it's a little bit off or something like that, and I'll know it's a little bit of harm. Or even if something, God forbid, happens to them, it's a little bit of harm. I tell them I can't do it. It's against my Hippocratic Oath. Or now, Dr. Eganer, Kupferberg, and others have given us this debate about, is Gleason grade group 1 even cancer. I tell the guy, hey, guess what? There's this debate: should we even be calling this cancer or is it a precancerous lesion? Just to help them put some buzzwords in. They'll go home, if it's like you said, that educated La Jolla guy, they'll Google something, they'll get on that New York Times article, and they'll be like, oh, yeah, this isn't that bad. The reality too, if I explain to them the curves that we were talking about, if they're on that curve C, which is that they can go their whole life and never need treatment, there's no way that we want to pass up that opportunity.

On the flip side, if I really see those warning flags there, then I'm telling them we have to maybe have to do treatment. In some ways, like I said, for that low-grade disease guy, it can be on the other end where you just know the patient's going to be in a bad shape. To your point, I think that pulling out your cards, this genomics really helping us, and it's not just that they're more objective, they are, it's, you know, actually, in almost every study, they are superior to the Gleason grade, which makes sense, right? You're looking at the phenotype from Gleason, even if you have Epstein reading it, he's looking at the phenotype, the genomics is looking at the molecular biology, and particularly Decipher that started out with the genome-wide signature, and then drilled down into what really is driving metastasis or bad potential. It makes sense. The reason that you're against genomics, not you, I mean, one might be against genomics, is this idea of cost. Is there a cost to the patient, cost of the system? If it can help you collapse decisional uncertainty, make it such that your first move is your best move, you're actually going to save the patient costs, save the system cost, and save the patient's life in some instances.

Again, a compound answer, but I agree that genomics helps you on the bookends, on two sides, it helps you de-intensify and understand cadence of surveillance, meaning how often you're following, and intensify. One other thing I should say, which is I've heard some providers say, well, I know what to do, and I don't need the genomics to help me. If we're talking about special tools, genomics, MRI, other stuff, they're like, I know what to do, I don't need the genomics, it's a rare case I don't know what to do. That is like, I won't say any French, but that's a bunch of hooey, right? For me, the approach I take is, I want to get all the information I can on you and your cancer so we can make our best decision. I don't know unless I get the information, I don't know. Now, I'm a little bit on the side of I think, in my practice, like I said, I like to collapse decisional uncertainty, so I sometimes will be an over-tester. For example, PET PSMA, I know it's not the conversation, but for staging, I use PET PSMA on my unfavorable risk, intermediate risk and high risk, and above. Because I want to know if I'm going to go in for that surgery, is it going to be curative or not? Some of my colleagues say, look, I'm going to do surgery on this guy no matter what. I don't need to stage him with something that's going to show me that they're not curable by surgery. I guess so, but for me, I'd rather give the guy a realistic expectation of what's going to happen to him before I make the decision but everyone varies in these ways. Actually, this comes up to like, there's no standard in how we qualify people for surveillance, how we do the surveillance, what takes someone off surveillance. Actually, like I mentioned, for the trials, I think that it's, again, just to plug myself, I'm really trying to put together something for a trial that we can do nationally, where one of the key elements is that we're going to have a committee decide if someone's really progressed on surveillance or not so that the patient can understand it's not just them and their provider.