Andrea Miyahira: Hi, everyone. I'm Andrea Miyahira at the Prostate Cancer Foundation. I'm excited to be joined by Dr. Kara Maxwell of the University of Pennsylvania and Corporal Michael Crescenz VA Medical Center. She will share her recent paper, Somatic Tumor Next-Generation Sequencing in US Veterans with Metastatic Prostate Cancer, published in JAMA Network Open. Dr. Maxwell, thanks for joining us.

Kara Maxwell: Thank you so much for having me. So, thank you so much to UroToday and, of course, the Prostate Cancer Foundation for funding this study. I'm excited to talk to you about the results.

So, I think that most of us are well familiar with the data that African-American men, self-identified Black men in the United States have a higher chance of being diagnosed with prostate cancer and a higher chance of dying from prostate cancer. And so one of the major things that we, as our group, are interested in is trying to understand these disparities. And to try to tease apart, can we figure out contributions of different components, be it structural racism, socioeconomic status, access to healthcare, genomic findings, and germline findings? And so, the purpose of this study was really to use what is the largest dataset of sequencing data comparing self-identified Black men and self-identified white men's tumors.

So, this is data from the National Precision Oncology Program within the VA. So as folks are probably aware, veterans have likely an increased risk of prostate cancer. And given the age of our veteran population, prostate cancer is the most common cancer that we see among United States veterans. And so approximately 6,500 patients at the time that we did this data freeze had undergone next-generation sequencing with the Foundation Medicine tumor testing, either of their primary tumor, one of their metastases, or a liquid biopsy.

We wanted to specifically compare self-identified non-Hispanic Black men to self-identified non-Hispanic white men. So, we excluded patients who either did not have their race or ethnicity listed or were not in one of those categories. That was about 1,000 patients. And then we excluded patients that didn't actually have metastatic disease. So occasionally, as one may imagine, some tumors get sequenced, even if they may not be metastatic. And we really wanted to look at the tumors that came from patients who actually had metastatic disease. And so that ended up with a cohort of about 5,015 patients, 36% of whom self-identified as non-Hispanic Black and 64% as non-Hispanic white.

You can see the breakdown of the specimens sequenced. And this is one thing that I really do want to point out, and because the VA does reflect, I think, what's happening a lot in community non-academic practices as well, about half the time the primary tumor is what is sequenced. Even though these patients are metastatic, often a metastatic biopsy is not obtained or it's not sufficient for sequencing. And so about half the time in this cohort, the doctors went back to the primary tumor for sequencing.

The metastatic biopsy was actually only about 20%. That was the minority of specimens. And then when a primary couldn't be found or a metastatic lesion wasn't biopsied, what happens in our clinical practice within the VA, according to our pathways, is that we then consider sending liquid biopsy from the plasma. So that ended up being about 33% of the samples.

So first, just looking clinically, as we would have expected, compared to non-Hispanic white veterans, the non-Hispanic Black veterans were younger at prostate cancer diagnosis. They had significantly higher PSAs at the time of diagnosis. There was less Agent Orange exposure in our self-identified Black veterans; that likely reflects the period of service. And then finally, our self-identified Black veterans were more likely to reside in neighborhoods with a higher area deprivation index.

So again, this study was really to look at the genomics. And so, these are a few plots that look at the top about 10 to 12 genes that are mutated in prostate cancer across these different specimen types. And I'm showing this visually because what we really end up seeing is that the top mutated genes really are very similar in the primary tumors, the metastases, and the liquid biopsies. They're just at slightly different rates depending on either the analyte type or if the specimen is from a self-identified Black or white patient.

And so for example, as we would expect, TMPRSS2-ERG alterations are seen at a higher rate in self-identified Black men. That's a finding that's been seen before. TMPRSS2-ERG is higher in white patients. TP53 was also found higher in primary tumors from white patients. And then when we think about primary tumor versus metastasis, of course we see AR mutations more in the metastasis than in the primary, and also in the liquid biopsy space as well. I think it's also important to point out that there are some differences in distribution, of course, in liquid biopsy that may not actually be reflective of the underlying prostate cancer, and may be more reflective of clonal hematopoiesis.

So then taking these data, we wanted to group the genes into pathways, in addition to looking at individual genes, to see if there were any differences in the rates of alterations between self-identified patients. And so, we did find a few significant hits here that are listed. Most findings, as you can see here, were higher in self-identified white patients. For example, alterations in tumor suppressor genes, the AR axis, and the AKT pathway. But then immunotherapy targets were actually higher in self-identified Black patients. Now, this is unadjusted analysis.

Then if we look at the alteration frequencies, actually considering the analyte, and on the left, the way you read this graph is this is the percent change between self-identified Blacks and whites. So the pathways on the left where the bars go to the left, that means the alteration rate is lower in Black patients, compared to the right, which is higher in Black patients. And so again, you can see that these immunotherapy targets, for example, across different analyte types are found at a higher rate in self-identified Black patients. But you can certainly see that the analyte type really matters.

And so what we then ended up doing is we looked using a multivariate logistic regression model, controlling for not only the analyte type here, but other clinical features. Then we can really tease out which alterations are really associated with race specifically when we control for the rest of these features. And we can see that a lot of these still remain significant, such that when you control for clinical features and analyte type, for example, TP53, PTEN, and other tumor suppressor genes are found at lower rates in self-identified Black patients, immunotherapy targets do still remain significantly higher in Black patients.

And so just to finish this out, what we wanted to start to look at was, could we then also look at these alterations to see if they were associated with survival? And so this is a Cox proportional hazards model looking at survival just specifically related to, do you have a mutation in this gene versus do you not? And what we ended up finding is that out of a lot of these things, most things fell out, but TP53 and other tumor suppressor alterations were associated with worse survival, but this was similar. They were associated with a higher risk of death regardless of self-identified race.

And so the conclusions from the study are that there are differences in alteration frequencies between self-identified Black and white men. But there are not really huge differences. Some of these may point to potential differences in the rate at which targeted therapy might be found, for example, immunotherapy targets. But we don't really see any very significantly large differences between the biology of these tumors that would really point us towards what might be different between self-identified Black and white men, as far as their prostate cancer outcomes. And it's just really important that precision oncology is a really critical tool for identifying actionable alterations, regardless of race or ethnicity.

Andrea Miyahira: Well, thank you so much, Dr. Maxwell, for sharing this study with us. So, what were the most important findings from this study for improving patient management decisions? And what are your take-home messages for the importance of performing genomics and genetic studies in diverse populations?

Kara Maxwell: Yeah, it's interesting. I've talked about this with a lot of people who have been a little surprised when we talk about the fact that sometimes there are some views out there that certain alterations might not be found as frequently in certain racial or ethnic groups as others. I think particularly in the germline space, there was some early data that suggested that self-identified Black individuals had lower rates of certain germline mutations, for example, BRCA1 or BRCA2. And that kind of percolated around. And I think that our data, along with a lot of other data that's out there, really shows that these alteration rates are probably not that different between racial and ethnic categories. And so, when we think about how we're counseling patients, when we're doing this testing, what is the likelihood we might find something that could be targeted? I think really self-identified race or ethnicity really shouldn't play a role there.

I think the other really important thing is thinking about the analyte that you're sending. And that there really are pretty significant differences, particularly in plasma. And that based on either the time at which that test is sent or technical limitations, plasma tests may not find alterations that you might be looking for to help treat your patient. And getting a metastatic biopsy, or even just trying to find the primary, particularly some of the clonal alterations that we know stay with the tumor across its biology, is probably a better bet than sending a plasma, for example, when somebody is responding well to treatment.

Andrea Miyahira: Thank you. And what are your next steps in these studies?

Kara Maxwell: Yeah, I think that really what we're most excited about is the beauty of this dataset in having very detailed clinical information with the numbers that continue to increase. Because the National Precision Oncology Program is sequencing hundreds of specimens every month, we're able to now get into a more granular analysis of different alterations' association with therapy response and survival.

And one of the major things that we're really excited about working on is being able to classify tumors into risk categories, very akin to what we see in leukemia where we have risk categories based on different genomic features and in a number of other cancers. And we really believe that this dataset's going to be able to help us with that. There are a lot of folks out there that have shown that tumor suppressor alterations are associated with worse survival. So, can we start to actually put numbers on that so that we can use that in risk stratification and then help with clinical trial design to figure out who are the bad actor tumors? I think that's really one thing that we're excited about.

And then the other thing is, going on what I mentioned earlier about plasma, trying to figure out when liquid biopsy is really going to be the most useful test. I think that a lot of times, particularly for genes in the homologous recombination pathway like ATM, a lot of the studies have been very confusing. Because it's unclear if a plasma sample is used, if that alteration is clonal hematopoiesis versus a tumor driver. And I think that we're going to be able to tease that out, or I know that's one of the things we're working on to tease out. When is the specimen that you're sending, if you have to send a liquid biopsy, when's the best time to send it and what is that alteration? Does that have different predictive value when, for example, the tumor burden is very high versus when the disease is under control?

And then also potentially really encouraging that getting primary tissue is probably the right way to go. I think that there are, in general, certain alterations that are seen in the primary that then have a similar effect in the metastasis. But then there are other alterations that do not, and may have a better reflection of the disease that's actually active and aggressive at the time that you're making therapy decisions. So, those are some of the things that we're looking at in the future with this dataset.

Andrea Miyahira: Okay. Well, thank you so much, Dr. Maxwell, for sharing this. And congratulations on this study.