Read the Full Video Transcript

Christopher Wallis: Hello, and thank you for joining us for this UroToday Journal Club. Today, we are discussing a recent publication titled, “Association of Inherited Mutations in DNA Repair Genes with Localized Prostate Cancer.” I'm Chris Wallis, an Assistant Professor in the Division of Urology at the University of Toronto. With me is Zach Klaassen, an Assistant Professor in the Division of Urology at the Medical College of Georgia. This is the citation for this recent publication in European Urology led by Dr. Lee.

DNA damage repair is a key cellular function. While rare, germline mutations in DNA repair defects may be found in patients with prostate cancer, they can have significant clinical implications as a result of their influence on the prostate cancer stage, the cancer risk at screening, cancer-specific mortality, as well as treatment response. Due to immediate treatment implications, germline genetic testing has clear clinical benefits, and this is particularly true in the advanced disease setting. However, current clinical guidelines including the NCCN guidelines recommend testing in high-risk localized disease, a space that lacks the strength of evidence that we see in more advanced cases.

As you can see here with three publications, the prevalence of germline DNA repair mutations is relatively well established among patients who have metastatic prostate cancer. However, for those with localized disease, the prevalence is somewhat less clear, and this is further confounded by the fact that there is a substantial under-representation of Black men in the literature. As a result, these authors sought to determine the prevalence of germline DNA repair mutations in a large, diverse academic biobank of patients with a focus on localized prostate cancer. To do this, they used the Penn Medicine BioBank and examined patients who were included between March 10th, 2000, and June 1st, 2019. They further identified patients with prostate cancer on the basis of the Penn Medicine Cancer Registry and ICD9/10 billing codes. All charts flagged as having prostate cancer were subsequently manually reviewed to confirm this diagnosis. In addition to these prostate cancer patients, the authors identified just over 26,000 patients in the Penn Medicine BioBank who did not have prostate cancer and these were included as controls. Among these, just over 3,200 had whole-exome sequencing and could be utilized for this study.  In addition, the authors identified the second cohort of cancer-free men to use as a second cancer control cohort.

Among each prostate cancer patient, the electronic health record was abstracted and natural language processing was used to collect unstructured data. Further, the Cancer Registry data were abstracted manually. In particular, the authors sought to gather the information that is relevant for guiding the provision of genetic testing, including family history, Gleason score, TNM stage, histology, PSA level at diagnosis, and Ashkenazi Jewish ancestry. In addition, they captured details including age, self-reported race, biochemical recurrence, and the development of metastatic disease.

When going through the process of germline sequencing, the authors first began with 2,391 patients with localized prostate cancer. Out of these, 1,666 had available DNA samples. On further review, the clinical record, 23 had de novo metastatic disease, thus leaving 1,643 samples among patients with localized prostate cancer. These samples were then prepared with the QiaSeq amplicon-based sequencing platform and multiplexed and subsequently sequenced using NextSeq550. This provides coverage of all exons of 17 relevant DNA repair genes as well as 827 single nucleotide polymorphisms. Following sequencing, the data were aligned and adapters were trimmed and all samples underwent joint germline variant calling. Following quality control at this step, an additional 47 samples were excluded leaving 1,596 samples available for analysis.

Among these samples, the authors determined the genetic ancestry of the included patients. They dichotomized this and grouped individuals into European and African ancestry subsets.  Among the nearly 1,600 patients, there were only eight who had discordant results between self-reported race and genetic ancestry. These patients were excluded leaving 1,588 patients in the study of which 1,174 were European and 351 were African. The authors then annotated variants across all three cohorts, that is the prostate cancer cohort from the Penn study, the non-prostate cancer cohort, as well as their second non-cancer cohort. These are annotated and variants were included for burden testing where the variant allele fraction was over 30% or was classified as pathogenic or likely pathogenic on the basis of the characteristics listed here.  Among those from the Penn Medicine BioBank, the age-adjusted VAF threshold was used to identify evidence of somatic interference. And these were subsequently dropped. The count of the number of pathogenic mutations was performed for each gene.

They then performed pairwise burden testing between the Penn Medicine BioBank prostate cancer cases and the biobank non-cancer controls as well as the gnomAD non-cancer controls. The mutation rates were compared on the basis of clinical characteristics using Fisher's exact test and all P-values were corrected for multiple comparisons using the Bonferroni technique.

I am now going to hand it over to Zach to walk us through the results of this study.

Zachary Klaassen: Thanks Chris. So, this is table one looking at the demographics, clinical, and family histories of the Penn prostate cancer cohort. And we can see here that the self-reported race was basically about three-quarters white. In terms of the age at Biobank enrollment, the median age was about 72, 73 years for the prostate cancer cohort compared to 61 years for the control arm. In terms of age at diagnosis of prostate cancer, 64 years, there was roughly three to 4% of patients that were under the age of 50, and about one-third of the patients were under the age of 60. With regards to the most common Gleason grade, this was Gleason grade 2, at 42% of the whole cohort and 44% in the sequencing cohort. Looking at the rate of biochemical recurrence, 16% in the cancer cohort and 17% in the sequencing cohort. Looking at the most common T stage, the most common stage was T2 at roughly half of the patients. Most of these patients were N0 at 77% to 80%.

And only a handful of patients had intraductal histology at less than 1%. In terms of family history of cancer, in terms of those with prostate cancer, family relatives, 76% and 75% respectively in the cancer cohort and the sequencing cohort had no family history of cancer. Not surprisingly, the most common major medical condition for both the prostate cancer patients and the controls was heart disease at roughly 60% in the prostate cancer cohort and 36% in the control arm.

This figure looks at the mutation rates of 17 DNA repair genes in localized prostate cancer among men of European descent. And we can see here looking at the way this is broken down on the top is the homologous recombination DNA repair genes, below that is the DNA damaged checkpoint genes and below that is the mismatched DNA repair genes.  And so looking at the cases, which is this dark orange color, we can see that BRCA is roughly about 1%, BRCA2 is 1% and BRCA1 is about 0.7%. When looking at aggressive cases which have slightly more genetic mutations, BRCA2 was almost 2%, BRCA1 was again about 0.7%. And when we moved down here, ATM for these aggressive cases was just over 1% for European descent. A similar figure, looking at men of African descent and generally, we see here that the mutation rate is quite a bit lower in African Americans compared to European descent. Again, we could see that BRCA2 for these aggressive cases is about 0.8%, PALB2, 0.8%, but not much of a signal for any of the other genes with regards to men of African descent.

This table is a case-control of the Penn Prostate Cancer cases compared to the Penn controls and the gnomAD controls among European ancestry individuals. And so, looking at this table overall, the only difference between the Penn Cancer cases compared to the Penn controls was that there was an increased mutation rate among the cases for TP53. When we compare these to the gnomAD controls, we see overall there were more mutations, 3.1%, compared to 4%, which was statistically significant. We see more BRCA2 in the Penn European ancestry cases compared to the controls. And again, we see more of a signal with TP53.

This looks at the mutation rates by prostate cancer clinical and pathological variables. I've put a box around the NCCN localized risk categories. So looking at intermediate, high-risk, and very high-risk, and essentially looking at these columns, this is the BRCA2 mutation rate, essentially all of the other important genes in this next column, and then on the far right, any repair gene defect. So, we're looking at intermediate risk, and we see that the BRCA mutation rate was only 0.5, 1.9 for the other mutations, and 3.1 for any DNA repair gene defect. Where things start to get interesting is when we look at the high-risk and very high-risk. So, 1% BRCA2 mutation rate in the high-risk, 2.5 in the very high-risk and overall 3.7% in any DNA repair gene defect for high-risk criteria and 6.2% for very high-risk.  Just below this line is the N1 diagnosis.  Again, we see more of a mutation rate in these patients, 3.7% for BRCA2, 15% for the other defects, and 15% for any DNA repair gene defect. I will quickly point out a couple of other interesting notes here. So, in terms of Gleason grade group 5, 3.4% BRCA2 mutation, 10% for the other genes, and 11% for any DNA repair defect.

This table looks at the association of mutation carrier status with aggressive clinical features. So, this is broken down by variables on the side as you can see here. The first column is all DNA repair mutations, the middle column is BRCA2, and again on the right is all the other genes other than BRCA2. So, we can see when we compare Gleason 6,7 to Gleason 8 for all DNA repair mutations, this is statistically significant as are those patients with Gleason 9,10 as well as those with N1 disease.  Looking at BRCA2, again Gleason 8-10 versus 6,7 statistically significant for the BRCA2 mutation, 9-10, again, verses 6,7, as well as biochemical recurrence, being more common among patients with BRCA2. A similar trend is seen for all other genes, notably for N1 disease, any gene mutation compared to those in the control arm.

Several discussion points from this important work from the Penn folks, germline mutations in prostate cancers are essential to characterize due to their implications for personalization of treatment. Importantly, other localized prostate cancer studies have lacked patients of African ancestry or included mostly aggressive disease or have lacked controls. And this study had all three of these variables. A current study demonstrated that germline mutation rates were lower in men of African versus European ancestry, 1.4% for African and 4.0% for Europeans. And interestingly, this disparity in prostate cancer outcomes for Black men with localized prostate cancer may not be due to germline mutations in these 17 genes based on this data. the germline mutation rate is lower in localized prostate cancer generally than in metastatic prostate cancer.  Historically, the metastatic rate of germline mutations is roughly 9-15%, whereas the localized rate is less than 6%. And as we saw in this study, the highest rates of germline mutations were in N1 disease at 15%, and those with Gleason 9-10 disease at 9%.

So in conclusion, there was a significantly lower rate of germline mutations for men of African ancestry with localized prostate cancer than for men of European ancestry. This also was associated with lower rates of DNA repair mutations in localized disease. Secondly, there were similar DNA repair mutation rates in localized prostate cancer cases compared to those of controls, and the majority of genes except for those in BRAC2. And ultimately, genetic testing restricted to localized prostate cancer patients with NCCN very high and high-risk disease would capture the majority of clinically actionable mutations in these patients.

Thank you very much. We hope you enjoyed this UroToday Journal Club discussion.