Read the Full Video Transcript

Zach Klaassen: Hi, my name is Dr. Zach Klaassen. I'm a urological oncologist at the Georgia Cancer Center in Augusta, Georgia. I'm pleased to be joined today by Dr. Colin Pritchard, who is a professor of pathology and laboratory medicine, University of Washington. Thank you for joining us today.

Colin Pritchard: My pleasure to be here.

Zach Klaassen: So today we're going to talk about something a lot about, and I hope you're going to educate our listeners, I know you will, about basically how to interpret a genomic report. And I think at a high level, just sort of give us an overview of who we should be testing when it comes to advanced prostate cancer.

Colin Pritchard: Well, I think you nailed it. Advanced prostate cancer patients. So I think that's really the key group right now where there's clinical utility is in the advanced setting. Certainly in patients with metastatic prostate cancer, there's all kinds of clinical utility, mostly for second line therapy, but still in that advanced setting. There are emerging roles, particularly in high risk localized disease, but much less well established.

Zach Klaassen: So if you look at the NCCN they lay it out very nicely, and I know it's hard to get through the NCCN guidelines at times, but even those unfavorable, intermediate and high risk patients we should be testing as well. So is there actionability even at that level?

Colin Pritchard: Certainly in terms of germline testing in particular, there's a recommendation even in sort of intermediate, unfavorable intermediate. There's actionability there for the family for sure, and I'm aware of what's in the guidelines. I think it's the right thing to do. I mean, to be honest, what you're going to do with that in the setting of localized disease for treatment is much less well established.

Zach Klaassen: More the counseling, the patients and their families?

Colin Pritchard: But certainly they're at higher risk for having germline mutations that are important not only for the patient, you're testing the prostate cancer patient, but for his family members.

Zach Klaassen: Absolutely. How about timing of testing? Obviously diagnosis. Is there other timelines in their prostate cancer journey where we should be considering testing, particularly from a somatic standpoint?

Colin Pritchard: It's going to obviously depend widely on the patient in front of you, but certainly a common scenario we see is at the time of castration-resistant disease, we will get testing. Progression on whatever therapeutic option is. Being tried is a very common timing. A lot of times it comes up because for example, if we take lung cancer where genomic testing has better established, longer lead time, there are now indications for looking for resistance mechanisms and then testing at that point.

And we're excited for that in prostate cancer, there aren't really indications for that yet clinically, although we do see people get interested in, "well, I'm treating with a PARP inhibitor or whatnot, and I want to see what the resistance mechanism is." So I think that will be more in the future too, that we'll be testing in that setting as we learn more what to do with those kinds of resistance mechanisms.

Zach Klaassen: That's great. So going back to the germline versus somatic, is there patients preferably, let's take just sort of your sporadic prostate cancer patients. Let's say they come with a Gleason nine or they show up with metastatic disease. Do they all need germline? Is somatic okay? How do you sort of delineate when we should be doing one or the other versus both?

Colin Pritchard: For somatic testing, the tumor testing, I would say we really need it for sure in metastatic disease. For germline testing, you had it from the time you were born, so there's no timing issue there.

Zach Klaassen: There's no timing, yeah.

Colin Pritchard: There's really just how likely are you going to have the variant? And so in that setting, I think, yeah, certainly high risk localized, I would personally be interested in testing if it was me. I think that's a good idea, but I don't know that we need to be doing somatic profiling at that point. I mean, there's certainly debate about that, but what you're going to do with that information at that stage of disease is much less clear in terms of treatment. Again, there's emerging data and there are some things to think through there, but certainly in the metastatic setting I would argue that both really ought to be, and that is pretty firmly in guidelines now, that both germline and tumor based testing is recommended in the metastatic setting.

Zach Klaassen: Do you have a preference in the metastatic setting, whether it's a bone biopsy, soft tissue, is there better quality with either one?

Colin Pritchard: Yeah, it's a good question. The issue with bone biopsy is that it's technically challenging to get a good specimen for molecular testing. So it's not the biology per se. If we were able to have a very, and some institutions and some organizations have very good protocols, where you can get a high quality bone biopsy for molecular testing. So it kind of depends on that. But for the most part, bone is a challenge. And so if you have a soft tissue site that's easily accessible, there's sort of a practical advantage of that because that's less of a challenge for pathology workflows as well.

Zach Klaassen: Less painful for the patient too, with a soft tissue as well.

Colin Pritchard: Very good, yeah. But in terms of what's better biologically, I don't know that one's a winner and I don't know that we have great data to know if there was a winner on that question right now.

Zach Klaassen: Let's talk about assays. So there's a bunch of commercially available assays. I'd like to maybe highlight which ones you sort of preferably will use. You have some of the assays you've developed at your own institutions. Maybe talk about both of those topics, if you can.

Colin Pritchard: Sure. So the reality right now is that the vast majority of this testing is done by commercial reference labs. I mean, it's expensive, it's not well reimbursed. And for those reasons and a variety of other reasons, also very technically challenging to bring up requires a lot of highly trained personnel. So for all those reasons, there aren't too many institutions that have robust in-house testing programs. So absolutely true. There's lots of good commercial lab options, and I'm not going to speak to specific labs or make any recommendations in this format, but there are a lot of reputable labs that are offering high quality testing.

That said, there are also some limitations to the commercial labs. And one of the biggest limitations I would say is that because of the nature of doing a reference lab test, I like to say throwing it over the fence. So you have a specimen and you throw it over the fence and then they do something over here and then they throw the result back over the fence. That's not really the best medicine. And increasingly as we need to know detailed clinical history, an oncologic history to be able to interpret those genomic results, it's going to be more and more difficult to have that model be the best medicine. So that's the inherent challenge.

Zach Klaassen: And in-house you're able to keep that all in one spot. And it is a lot less back and forth, so to speak?

Colin Pritchard: Yeah, and that's the advantage. I mean it's just analogous to having in-house radiology or in-house traditional anatomic pathology. So for all the same reasons that we like to have our radiologists on site so we can talk to radiologists and explain what the clinical scenario is and have that kind of medical expert interpreting the results who you may know personally and your colleague. For all those same reasons, it's I think very advantageous if possible to have in-house testing programs.

And we've been very lucky at the University of Washington, Fred Hutchinson Cancer Center to be able to have a robust in-house program for over 10 years. We started our NGS program in 2011 and have a large support from the prostate program to keep that going, which has been great. So the advantages of that is, again, we can have molecular tumor boards, which is a format where our molecular pathologists can interact directly with urologic and medical oncologists to talk about prostate cancer patients. And then we can do all kinds of tweaks to our assays to really meet the needs of our patient population.

Zach Klaassen: That feedback is back and forth in terms of tweaking it. That's great.

Colin Pritchard: And then I think you alluded to it too, and maybe we'll get into this side. The other thing you can do is push the envelope in terms of what's possible with testing. For example, paired tumor germline testing. Some other academic institutions have done a great job with that as well. The commercial labs are lagging a little bit behind on that. Some of the commercial labs are starting to take that approach.

Zach Klaassen: So looking at the report, so say there's somebody watching today, they're just getting into testing, they're on the bandwagon, they know who to test, what's the easiest way to interpret it? What are they going to see when they get the report back from a lab?

Colin Pritchard: So if we focus on the commercial lab setting, mostly it's going to be a highly glossy report. It's going to look good.

Zach Klaassen: It's going to look nice.

Colin Pritchard: It's going to look nice.

Zach Klaassen: Show it to the patient.

Colin Pritchard: So that's good. And typically, most of these reports we'll try to have a summary page that just like a summary table that just summarizes the really key finding with maybe the gene, the mutation, and then what do you do about it? What are the FDA approved drugs that may be eligible for it? So that would be the typical way a normal report looks. The other, in addition to mutations, these reports will almost always report microsatellite instability because that's an FD approved biomarker, tumor agnostic and tumor mutation burden.

And then more recently, some of the reports, including ours, will have a so-called HRD score or LOH score. You'll see these kinds of terminologies and that's kind of a surrogate for BRCAness. It's sort of like what I like to call that is it's the MSI of BRCA. So MSI is a surrogate for mismatch repair. This is a surrogate for homologous recombination, DNA repair, which is the BRCA gene pathway.

Zach Klaassen: So just like the level of MSH may predict response to pembro, this LOH sort of responsiveness may theoretically say a PARP inhibitor may be more apropos.

Colin Pritchard: Correct, but it's difficult to do well and it's more emerging. MSI as a surrogate for mismatch of repair deficiency is much better established as a mutation signature.

Zach Klaassen: Just to sort of wrap up, I'd like you to talk some about the challenges and maybe some of the inaccuracies and even some of the disparities that may come with genetic testing.

Colin Pritchard: Yeah, I'm glad you asked that question. So with these, and again, I'll focus on the commercial reports because a lot of the challenges, again, I think are solved if you have onsite testing and access to an expert pathologist, you can talk these things through with, especially in tumor board. But if we take the situation that is most common right now, at least in the United States where it's commercial assay, there's not a lot of discussion back and forth. One of the big problems is these summary tables are mostly what a clinician has to use to make those treatment decisions. And they're great in that they're a summary, but they lack subtlety. And so I'll give you a couple examples where one can go wrong. So let's take that example of a mismatch repair deficient prostate cancer, MSI high. That tends to have lots and lots of mutations.

What happens in the biology of that tumor is with microsatellite instability, you get subclonal like little small amounts of tumor cells that get these little frameshift mutations that's expected, including in genes like BRCA1 and BRCA2. It's actually very common to have mutations in BRCA1 and BRCA2 in the setting of mismatch repair deficiency. But those mutations do not predict PARP inhibitor.

So they're analytically there. They really are in the tumor. So when you take an algorithmic approach like a commercial lab does, and you just say, "I saw a BRCA2 mutation, use olaparib or use whatever, a PARP inhibitor", that is often wrong in the setting of MSI high because even though it's true that BRCA predicts PARP in this case it's the clonal. So the subtlety there is the key. And that's the kind of thing that if you have a local molecular tumor board, even with only commercial lab, your pathologists or other experts can talk that through with you.

Zach Klaassen: You can tease that out. Yeah.

Colin Pritchard: That's I think the number one common scenario, and I think the one that has the highest risk, because it is so often acted on incorrectly. And then another one that I'll highlight is false positive tumor mutation burden. So we just heard a talk at the Prostate Cancer Foundation meeting today that we're sitting at as we film this video that works that we weren't involved with, but from Dr. Nasser and colleagues describing how genetic ancestry, specifically non-European ancestry, particularly African ancestry and East Asian ancestry, can falsely elevate your tumor mutation burden.

And that's not a subtle problem. That's a problem where to the point where sometimes you can't even show that the biomarker works in certain populations. And the reason for this disparity is because non-European populations are underrepresented in databases that are used to subtract out germline variants.

So if you're doing a tumor only test, which most of the genomic testing right now is tumor only in the commercial space, you're going to have sort of false positive elevation of germline variants that are misclassified as tumor mutations in the setting of patients that are non-European ancestry. So that's an example of a disparity that has really, really practical implications for interpreting genomic reports.

So again, that's just a couple of examples and there's many, many others. And my recommendation to people watching this who don't have access to onsite testing is to really try to forge a relationship with a molecular pathologist or equivalent expert. Even if you don't have access at your own institution or your own private practice, try to find someone you can talk to so that you can consult, especially if something doesn't feel right on the report to you.

Zach Klaassen: That's a great comment. I was going to ask you what do we do in those situations, right? Because it's great to know about it, but it makes sense. I mean, this is clearly for practicing oncologists and neurologists. This is not our area of expertise. We rely on you guys, but when we see that report, it's hard not to take it as sort of the Bible of like, "This is what we're going to do." But I guess use your gut feeling if it doesn't feel right or doesn't seem right, just reach out to somebody hopefully in your institution or somebody that you can connect with to maybe get those nuances sorted out.

Colin Pritchard: That would be my recommendation, and hopefully we will get past this sort of time where this is a new field and I think organically as things get more complex and also we need more genomic testing to guide our treatment, I think organically there'll be more and more what I call the distributed model, where you have more and more onsite testing and more and more onsite experts just like we've done with radiology, just like we've done with traditional pathology.

Zach Klaassen: Excellent. Great conversation. Any last minute points or things we haven't highlighted that you want to touch on?

Colin Pritchard: This has been a lot of fun. I think we've covered all the high points.

Zach Klaassen: Well, I appreciate your time. Thanks so much.

Colin Pritchard: Thank you.