Assessing Breast Cancer Risk: Rationale for Multigene Panel Use

Transcript:Claudine Isaacs, MD: There is a study that was published in 2015, and it was really looking at the Invitae multigene panel. And what they did is they had a little over a thousand individuals who already had traditional genetic testing. And they looked, first of all, how their panel compared to what their traditional testing showed. The important thing is, when you’re looking at these multigene panels, if somebody is found to have a BRCA1 or BRCA2 mutation on standard testing, did the panel actually find it as well? And the first thing is that there was very, very high concordance. People who tested positive by traditional testing tested positive on this multigene panel testing. The people who tested negative for deleterious mutations tested negative, and the concordance was close to 100%, 99.8% or so in terms of the findings. So, that suggested that the Invitae panel could accurately figure out whether people had deleterious BRCA1/2 mutations or not.

And then the other thing—because obviously most of the people who had traditional testing had just had one or two genes tested previously—was: what does a multigene panel uncover aside from that? And they found a fair amount of variants. When we say a variant, we mean a different finding from normal. Now, what could that mean? That has lots of different interpretations when you’re doing genetic testing. A variant could be clinically important, it could be a deleterious mutation. A variant could be entirely inconsequential, just an alternate spelling of the gene that has no impact—what’s called a benign polymorphism—or it could be like we’re not sure which is a variant of uncertain significance.

So, there were a lot of variants that they were finding when they do multiplex testing or multigene testing. They classified them and through their techniques, they were able to classify them well, but it also brings up one of the conundrums. One of the issues when we’re thinking about multigene panel testing is that we get more variants of uncertain significance. The more you test and the more things haven’t been tested before, the more questions come up for the time being. And it speaks to the incredible importance of us having publicly available databases, so that those variants can be reclassified over time, and so we’re not left with question marks for too long. Certainly, with our experience with BRCA1 and BRCA2 testing, we turned the clock back 20 years. When we were starting to test, we were finding all these variants in BRCA1 and BRCA2. And the rate of variant detection plummeted over the years because we gained information about which variants were important and which ones were inconsequential.

Harold J. Burstein, MD, PhD: We’re beginning to see studies that address the practical utility of results from multigene panel testing, and there have been a couple of important reports here. Allison Kurian at Stanford reported a couple of years ago on a very interesting group. These were high-risk women. They were all being referred for hereditary breast or ovarian cancer counseling. They had BRCA1 or BRCA2 testing that was negative, and then they did panel testing to look for mutations in other genes. And, in that work, they found that about 15% of women had a mutation in one of these other genes. So, if you take a high-risk group, you might find mutations in about 15 out of 100 women who don’t have BRCA1 or BRCA2 mutations, and that’s probably a high enough number that it warrants more broad panel testing for women once the BRCA1 and BRCA2 genes are negative.

There’s been other work with other assays. The Invitae assay has been studied. They took a less heavily selected group of women, about 1000 patients. They excluded those patients who had known BRCA1 or BRCA2 mutations, and then they looked and they saw what the likelihood was of finding a mutation in these breast cancer patients who didn’t have that dramatic family history, though they had of course been referred for genetic testing. And in their study, about 5% of the cases turned up with a mutation on panel testing.

Now, historically, we’ve been recommending hereditary breast cancer testing when the likelihood seems to be about 10%, maybe drifting down to 5%. So, it suggests that even in relatively unselected groups of women, there’s a critical mass of data emerging for these multigene panels that approximates the kind of frequency we see with BRCA1 testing in similar patients.

Adam M. Brufsky, MD, PhD: The panels that were used in some of these tests are very similar—again, where they differ or what genes are actually sequenced, what’s actually annotated, what’s a variant of unknown significance. There’s a lot of other more minor differences, but to me the biggest difference in all of this, in some of the tests that were already done, had to do with the annotation of the genes. And what they found in this particular study, among others (there’s a number of studies like this now where they’ve actually done multi-panel gene testing) is a number of variants that actually appear to be pathogenic that were not realized in that. I think, those trials were done a couple years ago or those analyses were done a couple of years ago. And, again, our annotation of those genes is much different now than it was before. Therefore, we still have a lot of variants of unknown significance. But, actually, what we have now are a lot of variants that we know are either deleterious or not. And I think that’s probably, to me, the biggest difference in all of these studies that have been done.

But the bottom line is that multigene testing and multigene panels do have their benefits but they do have their drawbacks. I think it’s going to be an informed decision that both the patient and the physician have to make when they ask for one of these tests, and then try to interpret it. Because, on the one hand, we’re going to get a lot more information a lot more quickly. But, on the other hand, again, we’re going to get variants of unknown significance. And I think it’s really important that whatever panel you use, their annotation of the variants of unknown significance needs to be as firm and as solid as it can be.

Transcript Edited for Clarity