Genetic Evaluation of Hereditary Prostate Cancer - Sanjeev Kaul

November 15, 2018

(Length of Presentation: 27 min)

Sanjeev Kaul discusses hereditary prostate cancer (HPC) and why HPC is important, which patients should be tested based on the NCCN guidelines and what tests are available.  Also, presented is the process for clinic testing for HPC and the future directions in this field.


Sanjeev Kaul, MD, M.CH, Comprehensive Urology, Associate Professor of Urology, Oakland University and William Beaumont Hospital, Royal Oak, Michigan

Read the Full Video Transcript

Sanjeev Kaul: So we'll be talking about the genetic evaluation of hereditary prostate cancer. And in terms of the objectives, it's a who, why, what and when. What is hereditary prostate cancer, why is genetic evaluation of this condition important? Essentially, who should be tested, what tests are available, how can we incorporate this into a general urology clinic? And then some questions about what are the pros and cons of this right now? We don't have all the information as of now, it's a rapidly evolving field. And finally, what are the future directions and where do we hope to go in the future?

In terms of general definitions you have a general population risk, which is just a chance mutation, occurs in anyone, usually in older patients and more run of the mill type of cancers. Then you have familial risk, which is a clustering of cancers in the family, much more than what could be explained by the general population. And that's usually a combination of environmental factors and familial factors because everybody in the family is living in the same location. And then finally, you have hereditary cancer, which is cancers that are much more often in a family. They're usually seen in a younger patient population and also more aggressive or unusual forms of cancers.

And the difference between the general and the familial and hereditary cancer is mostly in terms of where the mutation occurs. So in the general population and in the familial cancers, it's a somatic mutation, occurs only in the organ at risk and obviously, within the cancer, the rest of the organ is normal. And in a hereditary cancer, it's a germline mutation to begin with. So it's something that stayed in each and every cell in the body. And more importantly, it's passed on down the generation. So it has implications not only in that patient but also in the family.

As urologists we know about hereditary and familial prostate cancer, it's been there for a long time. And those of us who manage cancers always ask the patients, do you have a family history? Anybody on the father's side, your brother, grandfather, so on and so forth. So the earliest criteria were the Johns Hopkins criteria, which were three or more first-degree family members with prostate cancer across three generations or at least two cases that are younger.

This has evolved over a period of time and now a couple of years back or last year we had the GEM study that was spearheaded by Jefferson and what they did was they looked at a prospective evaluation of multigene testing in these patients with a significant family history. So there are 200 men, 125 with prostate cancer. The rest who had no cancers and what they found when they evaluated these patients with multigene testing was that the number or the frequency or prevalence of mutations varied broadly. So in people with no prostate cancer, only 2.73% had a pathogenic or a malignant mutation. Overall, in the entire group, it was 5% but as you saw patients who had more aggressive cancer, Gleason 8 and above, or they had locally advanced cancer, the prevalence of these germline mutations was much higher, all the way up to 10%, 8-10%.

What is also important is that a variation of uncertain significance, which is a mutation, but one that we didn't know the meaning of was seen in 35% of patients. So this was the first study run by a urology group that told us that, yes, this is something that's important and that's something that we should look more closely at.

This is the Pritchard study that is published in New England Journal Of Medicine that looked at the germline mutations in metastatic prostate cancer patients unselected. These are not patients who had a family history. But again, if you look at the difference, almost 12% of these patients, irrespective of a significant family history, had germline mutations in a metastatic setting. And patients with localized prostate cancer, that number fell down to the single digits, approximately 5%.

There was a paper that was presented at AUA a couple of years back that looked at germline mutations, specifically BRCA mutations in African-American patients as compared to Caucasians. And what they found were that these mutations are much more common in African-American patients. And when you compare the BRCA patients with controls, patients who did have a BRCA mutation in this group had a much higher incidence of metastasis and also your time to metastasis was shorter. So what they explored was that these germline mutations may be a way to explain why African-Americans have more aggressive cancer and have poorer outcomes as compared to other racial groups.

So this is again another study, the IMPACT study that looked at the difference in clinical-pathological parameters between BRCA positive mutated and non-mutated patients. And they found that the PSA was higher in the BRCA mutated group and the positive predictive value of a biopsy was much higher in the BRCA mutated group. So again, if you have a germline mutation specifically BRCA in this study, you're more likely to have a higher PSA and you're more likely to have a positive biopsy.

Another paper, which was the Castro paper, actually provided a lot of information about how significant the BRCA mutation is in terms of significant disease and even the outcomes of these patients. So there are approximately 2,000 prostate cancer patients and they divided this group into BRCA mutation-positive and negative and they looked at overall survival, cancer-specific survival, metastatic-free survival and cancer-specific survival in the metastatic population. And what they found was that no matter which parameter you looked at, people who had a BRCA mutation did poorly. In fact, they found that 19% of these patients died in five years, almost 25% developed metastasis soon. And on average, these patients live seven years less than in the non-mutated group. So you can go on and on and you're going to look at these papers that compare these germline mutation patients to the ones that don't and there's just a consistency of more aggressive disease and poorer outcomes.

This is a paper that looked at the Gleason difference between the two and you find that people who have BRCA and certain other mutations are more likely to have Gleason 8 and above, more likely to have locally advanced disease T3 and node-positive disease and also their outcomes in a Kaplan Meier curve are poor, both in terms of cancer-specific survival and metastatic-free survival.

So, based on all this data, we had NCCN that looked at modifying your criteria of testing these patients for germline mutations. A couple of other slides. This is the PARP inhibitors. We've already spoken about that earlier that PARP inhibitors in mismatched repair mutated patients had a much better outcome, a more robust response and a higher response as compared to ADT and chemotherapy.

The Philadelphia Prostate Cancer Consensus was run by Lenny Gomella from Jefferson and this was a gathering of 70 odd experts in the field and they looked at all the data that is available out there and they wanted to put forward some consensus about what the urologist should do in this situation.

And what they found was that the common genes that are affected in hereditary prostate cancer, BRCA1, BRCA2, ATM, CHEK2, HOXB13, and that depending on which gene is mutated, that we should incorporate certain treatment algorithms or screening algorithms to identify these patients and identify the cancer sooner and also treat them in a particular way.

So BRCA2, all the risk categories should, whether it's low risk, intermediate or high-risk patients. ATM and BRCA2 in high-risk patients and ATM and BRCA in metastatic CRPC patients.

The important genes, BRCA1 and BRCA2, are uniform. Each and every study shows that these genes are the most important in hereditary cancer and that patients who have BRCA2 mutations and BRCA1 have a higher incidence of cancer, more aggressive cancer, and poorer outcome.

Lynch syndrome, a group of genes that are the mismatch repair genes also has been shown to be important. They cause a higher prevalence of prostate cancer, a higher likelihood of developing prostate cancer in those groups of patients and they also have a higher risk of developing upper tract TCC. And the recommendation right now is that those patients should be screened more aggressively starting at age 40. HOXB13, similar situation, a higher risk of developing prostate cancer in their lifetime and also a higher risk of developing more advanced prostate cancer diagnosis. CHEK2 and ATM are the two other common genes that had a high prevalence of the mutation in the Pritchard study as well.

So how does this impact the care of our patients? Somebody who does have a germline mutation, they are at high risk for developing additional cancers. So if you have a BRCA2 mutation, your risk of pancreatic cancer is extremely high compared to the general population. Your risk of melanoma and male breast cancer is extremely high as compared to the general population. If you have MSH repair genes, that's Lynch syndrome, you can have colon cancer, you can have upper tract TCC that you need to be screened for.

Screening for prostate cancer in this group of patients, especially if they are unaffected, you should have a higher index of suspicion and screen these patients more carefully as compared to the general population.

And then there are certain genes that can drive us to a specific treatment. Dr. Harm spoke about the PD1 inhibitors in mismatched repair genes. You've got BRCA1 that can be patients that can be treated with PARP inhibitors that have a much better response and a more durable response, compared to chemotherapy and ADT in metastatic prostate cancer patients. So there are multiple treatments that are available that are individualized to these groups of patients.

So as a result of all this data, NCCN, last year in 2017, included family history and testing of these patients for germline mutations in their NCCN guidelines. And those guidelines this year have been updated to include testing for germline mutations in all patients with metastatic prostate cancer, with all patients with Gleason 8 to 10 prostate cancer. Patients with 4+3 Gleason and a PSA about 20 and then for patients with very low, low and favorable intermediate-risk, they should be tested if they have a strong family history. And a strong family history includes one other family member who has prostate or pancreatic cancer or breast cancer at less than 50 or two family members with prostate or pancreatic or breast or ovarian cancer at any age.

So there are strict guidelines that are laid down by NCCN in the 2018 NCCN guidelines to help us to decide which patients need to be tested. NCCN also included some guidelines for multigene testing as opposed to single-gene testing because most of the genes that are affected, BRCA1 or BRCA2, make up almost 50% of all the mutated genes. Most people were doing single gene testing because it's cheaper. NCCN came up with their guidelines and said that multigene testing is actually more efficient and better than doing single-gene testing. And that commercial tests are not available and there are commercial mainstream tests and also direct to consumer tests and we'll talk about that a little bit, that they differ in the genes that they analyze and that careful consideration should be made before you order the test.

They also said that the likelihood of finding radiation of uncertain significance, which is an issue, is higher if you do multigene testing, but your probability of picking up a mutation is also higher across multiple genes. And the multigene test includes moderate-risk genes, which are the BRCA1, BRCA2, ATM, CHEK2, PALB15, HOXB13. And that certain genes have clear cut criteria, whereas other genes have more limited data on their impact.

And they strongly recommended the use of genetic counseling either before or after. And that's again a controversy or an issue whether we should send all these patients to genetic counselors and have them do the testing or should we as urologists? These are our patients, they see us almost as primary care physicians, should we do the testing and then send only the appropriate patients to a genetic counselor.

So their guidelines for a genetic counselor are here that anyone with three or more of these cancers, anyone with metastatic prostate cancer or anyone with a positive mutation on multigene testing. Now there are only 3000 genetic counselors in the country and almost 80 to 90% of them don't even deal with oncology. They deal with prenatal genetic counseling.

So you've got a very limited number of genetic counselors that are available. Most of them are available in an urban setting and as you go outside the main cities, there are very few genetic counselors that are available. And if you start sending each and every one of these patients to them, they very frequently get overwhelmed.

So I think it's a better strategy for us to order the tests. We have the guidelines, we are seeing the patients and then the appropriate patients that do have a mutation, and that's going to be somewhere between five to 20% of these patients, only those should we send on to genetic counselors and that's just a more efficient way of doing it.

There are multiple tests out there. Some by established commercial laboratories, some by direct to consumer. You can go on Amazon and type hereditary gene testing and you'll have six different tests that are available that you can buy off Amazon. That's how prevalent it is. It keeps changing. You've got new companies that keep coming in, offering these tests for a limited amount of money.

There's a difference between these tests though. Some offer an included genetic counseling support, so if you buy the test and you do happen to have a positive test or even a negative test and you want to talk to a genetic counselor, a telephone consultation is included in the test. Others don't. As the NCCN guidelines said, there are differences in which genes are done. Some have just three genes that they evaluate. Some have six genes, some have 30 genes, some have 90 genes. So there's a lot of variation out there in terms of what you're going to get when you pick one of these tests.

How do we incorporate this into clinical practice? I think for sure it should be incorporated. I think it's there. The NCCN guidelines are there and it's very easy to incorporate this into a regular urology clinical practice. I think the most important thing in incorporating hereditary gene testing is capturing an accurate family history and that can be so variable.

We have a couple of studies that were done. There was a study, I'm just going to step forward one slide. So this was a study that was done by Dr. Henderson's group and over a period of a month, they captured family history from each and every patient that came to one of their offices. And what they found was that the overall prevalence of significant family history in that population was almost 30%. And even in women, it was 34%. It was 25% in men. And these are just all-comers, not somebody specifically who comes to the office with cancer. So clearly, significant family history is there, you can get that in a significant proportion of your patients.

Once you've done a good family history ... Now, we did a study in our own practice. When we did a regular family history evaluation, which was something that the patient either filled out in the waiting room or something that the MA obtained from the patient, we found 17% of our patients had a significant family history. But then when we resorted to intensive questioning that was either done by the physician or by a research coordinator that called the patient and tried to get more information, that prevalence of significant family history increased to 35%. So you can capture double the number of patients and everything depends on the robustness and the accuracy of your family history.

Once you have that and you have the NCCN guidelines to help you who should be tested, the rest of it is pretty simple. It's either a blood test or a spit test. There's even a company that just does a buccal swab to evaluate it. And then once you get that done, you get the results typically in a matter of a week or two weeks. And the results identify which mutations are present, whether they are benign mutations or clinically significant mutations or variants of uncertain significance. So this is just a process flow which is pretty straight forward. I think the most important thing is just capturing a good family history.

A couple of issues with gene testing and especially these mutations. One of the problems we face in this space is what is a positive result? Just identifying a mutation on one of these genes doesn't necessarily mean that that mutation is clinically significant. You're going to have mutations across the entire BRCA1 or BRCA2 gene, but not all mutations are going to code for a protein and not all coding mutations will actually affect the function of the gene. So when you do this test, a positive result is a combination of the analytical accuracy of the test, which is how well does that test identify these mutations in the genes, which is pretty straightforward.

I mean you have the coding of the gene and you analyze it and if there's any change, that's a mutation. But I think what's more important is the interpretive accuracy. What does that mutation actually mean? Is that mutation associated with a higher incidence of cancer or more aggressive cancer? Or is that mutation just a benign mutation that just happens to happen, but it's not going to affect the function of that gene in any way? So I think that's the most important parameter.

And when you combine both of them, that's when you get an accurate test. So there were a couple of papers or there's a lot of papers, but there were a couple of significant papers that looked at the differences in these gene mutations amongst the different labs out there and they've classified these labs into established commercial laboratories and then direct to consumer tests like 23andMe® and Ancestry® and so on and so forth.

And they found that there's a significant difference in the accuracy of identifying these mutations amongst these labs. So this was one study that looked at 1,900 patients and they found 600 odd genetic gradients. Again, 37%, which is pretty consistent with the GEM study that we saw earlier, were variants of uncertain significance.

But they found a 26% discordance between two labs. So you have one lab that found a mutation and said it was clinically significant. And then the same sample looked at by another lab said that that was not significant radiation. And in 11% of those patients, the difference was between a radiation of uncertain significance and a pathogenic variation. So if you look at that and you think to yourself that the pathogenic variation, especially in a certain group of patients with breast cancer or ovarian cancer, can drive treatment decisions and a bilateral mastectomy or oophorectomy, that's a pretty significant discordance.

This was another paper that looked at false-positive results in a comparison between direct to consumer genetic tests as compared to the established companies out there and they found that there was a 40% false-positive rate when the tests were done by the direct to consumer labs as compared to an established commercial lab. And then they also looked at the different genes that were affected in this false positive. And you can see it's almost all the genes. And so the difference is, again, how to identify which mutations are clinically significant and which ones are not.

In terms of unanswered questions, what are the pros of doing genetic testing? Clearly, prostate cancer, almost a quarter of prostate cancer or 30% of your prostate cancer patients are going to meet testing criteria. So it is a legitimate problem that you're going to see in your office. It's important because you're going to surveil these people for secondary cancers. You're going to identify at-risk family members. These are all autosomal dominant mutations. So they are transmitted to future generations 50% of the time. So it's important to identify if one of the progenies or even siblings have a mutation.

And then, if you do identify a mutation, it has several advantages for that patient. You can administer more accurate treatment in terms of avoiding active surveillance in somebody who has a mutation versus treatment. You can avoid radiation in people who have mismatch repair gene mutations because they're not able to repair the mutations from radiation. They are at a high risk of secondary malignancies. And then also, depending on whether they have a BRCA mutation or an MSH mutation, those patients are more accurately or better treated with PARP inhibitors or PD1 inhibitors.

What are the cons? One of the issues with prostate cancer genes and one of the earlier talks today spoke about how the genes that we see are ... There are a lot of genes that have a very limited penetrance. So as compared to breast or ovarian cancer where the penetrance of these genes is 80%, 90% and your risk of having cancer is extremely high. Most of the prostate cancer germline genes have a 20 to 40% penetration. So even the impact of those genes is not as much.

Only about four to 20% of these patients will have a malignant mutation or a deleterious mutation. And the cost of testing has been an issue. So you're going to test a patient and then you've got cascade testing because you're going to now identify family members.

There is one point to note though that multigene testing is only required for the index patient. Once you've identified which gene is mutated, the remaining family members don't need to get multigene testing. They can only be tested for that single-gene mutation. And that's a much cheaper test to do.

In terms of unknowns. We test unaffected individuals and that's a part of the hereditary breast and ovarian cancer NCCN guideline. Where even if you see a patient in the office who comes to you for a prostate cancer check and they're 50 years old or BPH and they're 60 years old and you get a family history and they have two or three family members and they fit the strong family history criteria. And one of the family members has been checked and has a mutation, that unaffected patient can be checked for a germline mutation.

And then what are the implications for health insurance? Are we going to get into a "get a card" type of situation where we have gene information in everybody and you're excluded first from health insurance or life insurance?

There is an act of Congress that prevents health insurance companies from using this information to deny you health insurance. And so that cannot be done. You're going to get your health insurance but that same protection is not available for life insurance. So you could be denied life insurance or your life insurance could be much higher if you have that information available.

What do you do with these variants of uncertain significance? It's a constantly changing situation. What is a VUS now? A year or two or 10 years down the line could be much better defined in terms of being benign or malignant. And then one of the problems we do face is how do we educate the family through all this genetic maze and how do we get everybody involved?

As a part of the music consortium, we are looking to create a registry for these hereditary gene testing across the state of Michigan. And some of our goals are to get a statewide registry to try to correlate germline testing with actual tumor testing or tumor gene testing and then correlate both of those with the eventual outcome of the patient in terms of active surveillance, surgery, radiation. And also try to optimize the panel for multigene testing.

And I think one of the goals would be to have a panel of these gene tests that can, just like a PSA at age 40, that can be done on men at age 40 and that gives us a good idea or a very accurate estimate of who is likely to develop prostate cancer and certainly significant prostate cancer in their lifetime. So I think that's one of the goals for using this gene testing as well. Thank you.