Andrea Miyahira: Hi, everyone. Thank you for joining us today. I'm Andrea Miyahira at the Prostate Cancer Foundation. With me today is Dr. Brandon Mahal, an assistant professor at the University of Miami. He'll be discussing his group's recent paper, Comprehensive Genomic Profiling and Treatment Patterns Across Ancestries in Advanced Prostate Cancer: A Large Scale Retrospective Analysis, that was published in Lancet Digital Health. Thanks so much for joining me today, Dr. Mahal.

Brandon Mahal: Thanks, again, for the introduction. And thanks for inviting us to give a talk on our publication, again, that was recently published in June 2023. So, a little bit of background on my study. It is well-established that prostate cancer incidence and associated mortality vary widely by race and genomic ancestry. And that men of African ancestry, specifically, experience the greatest burden of disease. These differences are likely due to the interplay of socioeconomic factors, environmental exposures, and biological or epigenetic phenomena. Furthermore, precision oncology studies have severely underrepresented men of African ancestry, thereby limiting our comprehensive understanding of prostate cancer disparities. Therefore, in the largest cohort reported to date of just under 12,000 patients with comprehensive genomic profiling and a subset of over 1200 patients with real world genomic and clinical outcome data, we sought to characterize the genomic landscape, comprehensive genomic profile utilization and treatment patterns, based on genomic ancestry in an effort to better understand disparities in advanced prostate cancer.

So getting a little bit into the cohorts. Again, there were 11,741 patients in the large scale genomic cohort that included over 1400 men with African ancestry, which is approximately 12% of the population. Again, with advanced prostate cancer and comprehensive genomic profile testing from which genomic alterations were identified. And then there was also this clinical genomic database of just over 1200 patients where the CGP utilization and treatment patterns were analyzed. Notably, the genomic ancestry was predicted using a SNP-based approach using a thousand genomes data. For the purposes of this study, focused ancestry specific interrogation of the genomic landscape between patients of European and African ancestry was performed in this cohort. But we also have supplemental data on examining patients with Asian ancestry. So here's the overall map of the mutation landscape, including the most commonly mutated genes in European and African ancestry individuals, which highlights a potentially ancestry specific mutational landscape pattern.

I'll better highlight these patterns in the coming slides. Specifically, the mutational landscape across ancestry demonstrated differences in the depicted genes, notably TP53, PTEN, TMPRSS2:ERG. In the African ancestry patients there were notably lower frequencies of the alterations mentioned, and there was a higher frequency of MYC, SPOP, CDK12, KMT2D, CCND1, and HGF. I'll highlight the specific differences in the following slide. So this plot to the left highlights whether gene alterations were depleted or enriched in African ancestry individuals. With genes that were depleted, depicted to the left and represented by gray dots. And genes that were enriched to the right and represented by orange dots. The size of the dots represents the prevalence of alteration. The X axis represents the odds ratio comparing African to European ancestry individuals. And the Y axis represents the P-value. With anything above the dotted lines depicted representing genes with significant differences in alteration frequency by ancestry.

We used the Fisher's exact test for this test of enrichment, and the P-values were notably adjusted for multiple testing. The threshold of the dotted line is drawn at 0.05. So again, the genes that were significantly depleted in African ancestry individuals were TP53, PTEN, and TMPRSS2:ERG. While those that were enriched included SPOP, CDK12, CCND1, KMT2D, HGF, and MYC. Now when summing actionable genes and DNA damage response genes as depicted in this figure, we found no significant difference in the frequency of alterations across ancestry, and either actionable genes or DNA damage response genes except for BRAF. The BRAF is notably the one gene that was in the actionable gene panel that was different by ancestry. Otherwise, when summing all genes together, we did not find any difference, nor by individual genes.

So now switching gears to highlight the findings from the clinical genomic database. Please note that overall this patient population, as I mentioned, is smaller with just over 1200 patients, and it's quite heterogeneous. And therefore the findings must be interpreted within these limitations. So this slide specifically illustrates that African ancestry individuals receive comprehensive genomic profiling later in their treatment course when compared to European ancestry individuals. Specifically after a median of two lines of therapy compared with one line of therapy in European ancestry individuals. And this difference could ultimately impact the observed mutational landscape by ancestry. I'm not showing the figure here, but we did see a survival difference in this retrospective cohort when analyzing survival by timing of CGP. Where men who got CGP earlier in their care tended to have a longer overall survival with an increase by 10 months.

Again, this was in a retrospective cohort and likely reflects differences in access to care. Furthermore, we noted that the overall proportion of men receiving targeted therapy was not significantly different by ancestry. And that's also what's depicted in this figure. However, when analyzing the same cohort and looking at the probability of receiving clinical study drug, this is after receiving comprehensive genomic profiling. This slide highlights that overall African ancestry men were less likely to receive clinical trial study drug. With only 11% of African ancestry individuals receiving study drug versus 30% of European ancestry individuals. This difference could ultimately impact the observed mutational landscape by ancestry. Notably, these differences were observed in both the community and academic setting. As you can see, we've stratified these analyses. Please note that these specific findings are in very small cohorts. In particular, there were only five African ancestry patients treated at academic centers. And therefore must be interpreted within these limitations.

So overall, there were 11,741 prostate cancer patients with advanced prostate cancer who had comprehensive genomic profiling as a part of their routine clinical care, and who were evaluated for their genomic landscape. The analysis of these patients that included approximately 12% of individuals with African ancestry, revealed largely similar rates of alterations in genes with therapy implications across ancestry. So we conclude that intrinsic biological differences are unlikely to be a major driver of ancestry-based disparities, specifically in men with advanced prostate cancer. When looking at the over 1200 prostate cancer patients in the clinical genomic database, at the trends in CGP testing and utilization and treatment patterns, we found that men with African ancestry were less likely to receive CGP testing earlier in their treatment course and less likely to be treated on clinical trials. These types of differences may potentially impact the genomic landscape outcomes and ultimately disparities. And we argue that equitable use of comprehensive genomic profiling, clinical trial enrollment and subsequent precision medicine treatment pathways could lead to a major reduction in disparities.

I'd like to acknowledge all of the collaborators on this study. In particular, the co-first authors Smruthy Sivakumar and Jessica Lee. As well as my co-senior author, Jeffrey Venstrom. All of the Foundation Medicine team who performed a heavy lift and collaborators who were at the University of Miami Sylvester Comprehensive Cancer Center and abroad. Thank you for your attention.

Andrea Miyahira: Thank you, Brandon, for sharing that with us. So certain genomic alterations such as AR that are clinically important are acquired during disease progression and not very prevalent in early stage localized disease. So did you compare genomic alterations between ancestries among stage or prior treatment that matched cohorts, specifically when the CGP was done on that stage of tissues?

Brandon Mahal: Yeah, thank you for the question. It's a really important question to address. Two points I want to make about the cohorts. So we used two different cohorts. There was one large cohort that was nearly 12,000 patients. And those patients, we only had comprehensive genomic profiling data on those patients. We know that they had advanced prostate cancer, but we did not have information on their treatment types. Then there was the clinical genomic cohort, which again included patients who had comprehensive genomic profile testing. And these patients also had treatment information. And so in the larger cohort where we make most of our conclusions about the genomic analyses, unfortunately we were limited by not having the complete treatment data information.

And then in the smaller cohort, unfortunately, that cohort is so small that it was difficult to match by prior treatment types. So I would say that this is a limitation of the study. But certainly something that we've extrapolated from the findings is that although there weren't major differences in types of treatments received, there was definitely a difference in the time that CGP was performed. The comprehensive genomic profile testing was performed, and there was a difference in receipt of clinical trial drug.

So ultimately we do conclude that those differences in the timing of the genomic testing, and differences in clinical trial enrollment, could ultimately impact the genomic landscape. So we do think that there's probably some differences in there about when those mutations were acquired in disease progression that do give some of those ancestry specific differences that we saw in the genomic landscape.

Andrea Miyahira: Okay, thank you for that. Some of these other alterations are not actionable, but they are associated with clinical outcomes and treatment responses. Such as, for instance, p53 and PTEN loss. We know those indicate much more aggressive disease while SPOP mutations are associated with less aggressive disease typically. So were you able to compare frequencies of alterations with these sort of known clinical outcome associations across the ancestries?

Brandon Mahal: Yeah, so we were able to examine the specific genes that we had in the panel, including those TP53, PTEN, or SPOP. And what we found is not a consistent story. So we found that TP53 and PTEN were depleted in African ancestry individuals, whereas others like SPOP were enriched. And so ultimately we didn't have a genomic signature necessarily that was to predict for prognosis of aggressive disease that we were able to examine. But we did examine those individual genes and those three stood out in particular. We did look at genes that predicted for response, again, such as AR and DDR genes, which in a prior study in a much smaller cohort and in other studies from other independent groups, have seen differences specifically by race.

This study is different because we use genomic ancestry. But in this large scale study, again, the largest of its kind in advanced prostate cancer, those genes that predicted for response to therapy such as AR and DDR genes didn't show any significant difference. But there was one that stood out, and it stood out in a lot of studies. And so this has been validated in other independent groups, and that was BRAF. So BRAF alterations were enriched in African ancestry individuals. So this could definitely be the basis for future clinical trials, although the reasons for the association are not definite. And we did not, again, develop or look at specific genomic signatures for this particular study. But that could be an opportunity for future studies as well.

Andrea Miyahira: Thank you. I think one observation that struck me was that genomic alterations were similar between African and Asian ancestry patients, despite them representing like the highest and lowest disease burden populations respectively. So what do you think this tells us about the contributions, or rather lack thereof, of genomics to racial disparities?

Brandon Mahal: Yeah, thank you for that question as well. And so, there's a lot of data on our more comprehensive analyses of the genomics by different ancestries in the supplemental material. So I would encourage anybody who's watching to go take a look at that. And that's where you can find most of the information about your question.

So we did find in that supplemental material that there were ancestry-specific genomic landscapes, even by African versus Asian ancestry. However, there were no differences as you noted in genes that were targetable. So I would say specifically this means that men with advanced prostate cancer... When you take men with advanced prostate cancer across ancestry, and you look at potential differences across ancestry, there's unlikely to be major differences in tumors once disease becomes advanced by ancestry. And so I would say that when you see disparities reported in men with advanced disease, I would say that particular scenarios, we have enough information from this study and other studies to suggest that cause is other than genomics. Such as access to care, systemic racism, structural barriers, differences in really, like I said, access to care, I think would be the major driver of disparities in advanced care.

But I would say that it's also important to make note that genetics may track with ancestry and ultimately contribute to differences in the risk of developing prostate cancer in the first place. So when you just look at men out there and say, "What's the overall risk of developing prostate cancer?" There's definitely a genetic predisposition that drives the incidence of prostate cancer. And ultimately developing prostate cancer is the biggest driver of prostate cancer death. And so there's still definitely good reason to study genomics in prostate cancer, but maybe it's better to focus those efforts on the differences in incidence of disease rather than looking at differences in genomics when disease is already developed and matched by stage or aggressiveness. And then the other piece is that there may definitely be gene environment interactions that influence outcomes. Different comorbidities that interact with the environment and ultimately the underlying genome to drive some of these differences. So I still think there's a big, big opportunity and room to understand genomics and how they contribute to prostate cancer disparities. But I think we should maybe shift our focus to where we're looking.

Andrea Miyahira: Yeah, thank you for that. So your data do support that access to care and systemic racism are significant contributors or probably like the major contributor to racial disparities in outcomes. So do you have any recommendations for actionable next steps for the clinical community to take to overcome disparities?

Brandon Mahal: Yeah, absolutely. So I would share some of what we're doing at our institution here at the University of Miami Sylvester. We're using a community-based approach, and this would not be an approach that's necessarily one-size-fits-all. Because I think, especially for cancer centers, the catchment area that we serve is, they're very different across the country, across the state, across regions within the state. And so it's very important to have community partnerships first between our treatment centers, our healthcare centers, our cancer centers, and the community. And then once that's established, I think it's important to understand the catchment area that we serve individually. And study the disease burden alongside social determinants of health, ancestry, genomics, and that would hopefully help us tease apart the relative contributions to increased mortality. So it's not only important to understand and address disparities, I think that's the most obvious thing we can draw from it.

But it's also important for us to gather information that we haven't been gathering, such that we're better able to understand the risk factors for disease in general. So I think there's a lot of environmental influences, a lot of genomic genetic influences, and the interaction between the two that we just have not been looking at. Because we haven't been engaged with those communities who are underrepresented. And I think that ultimately this is best done through community-based approaches that again, vary from region to region. And so for example, here at the University of Miami, what we're doing is we're studying incidence in mortality of prostate cancer by neighborhood level and by disaggregated race/ethnicity, in an effort to better understand what neighborhoods and communities are most impacted by prostate cancer. Then we go directly to those communities either by partnering with local pharmacies or using mobile clinical units.

And we go out and we do targeted screening efforts. We give information about prostate cancer, but we also consent patients to research in a natural history cohort. And I think these types of efforts are going to be the way forward. And ultimately, it's the way to go and gather information about what's driving disease, but also directly impact the communities who are most at risk. And feed that information that you get right back to the community, and improve access to screening and treatment while doing the research. So again, I would argue for a community-based approach that varies from region to region and center to center, as providing the most potential high impact in addressing disparities.

Andrea Miyahira: Thank you so much for joining me today, Dr. Mahal.

Brandon Mahal: Of course. Thank you. Thank you for having me.