Case 3: Molecular Testing in Recurrent Ovarian Cancer

EXPERT PERSPECTIVE VIRTUAL TUMOR BOARD

Robert L. Coleman, MD:Greg, maybe you can tell me a little bit about what the difference between somatic and germline is? And then, also, what is next-generation sequencing?

Gregory Riedlinger, MD, PhD:A lot of people don’t like the termnext-generation sequencing. You’ll hear it called second-generation sequencing, and there’s potentially third- and fourth-generation sequencing. Some people like the termmassively parallel sequencing. You can do DNA or RNA, but mainly DNA, and make a fragment of what you’re going to sequence, but many millions of fragments. And so that’s why it’s the massive amplification. It has the power, compared with traditional Sanger sequencing, to sequence millions of reactions at the same time. So you can sequence with some of the commercial assays. Most people aren’t necessarily doing the entire exome or the genome. There obviously are some companies that do that, but a targeted panel in which you can look at like 320 genes or things like that on MSK-IMPACT or Foundation Medicine tests. As the cost of this has decreased, and then companies like Foundation Medicine, Inc, have gotten FDA approval for a lot of the advanced cancers, we’re seeing more and more of it being used routinely.

And so the important thing to know with somatic testing is that it’s really done on the tumor. Most of these companies, when they’re sequencing the tumor, it’s from an FFPE [formalin fixed-paraffin embedded] block. If it’s ever insufficient, it’s typically because there’s not enough tumor in the block compared with background lymphocytes or a stromal tissue. But you’ll see a cutoff typically of about 20% for that type of sequencing, to make sure you’re actually sequencing the tumor, which is the somatic as opposed to what would be in the germline. And with most germline testing for solid tumors, that’s typically going to be performed on blood. Where it’s really becoming useful is when you’re seeing a lot of treatments that are not necessarily as much derived from tissue of origin but underlying molecular alterations.

With the treatment withBRCA1andBRCA2, obviously those aren’t the only genes involved in the homologous recombination repair. You have other genes, likePALB2andRAD51and things like that, which could be targeted with similar therapies, potentially. But there are many others. With larger-panel sequencing, where it’s not necessarily even exome but where you’re looking at 320 genes, you can pick up a lot of other things. If the patient didn’t have mismatch repair testing, you can actually see that signature that we do see more, or the microsatellite instability that we see more often in the endometrial or the colon cancers. But there obviously are certain percentages of cancers that have that.

Then there are other driver alterations in specific genes that are seen much, much less frequently in Mullerian origin, but there are trials with the NCI-MATCH in which patients can match on therapy with that. And there are other rearrangements that we typically think about in lung adenocarcinoma, withRETrearrangements orFGFRrearrangements. So it kind of opens the box to potential treatment options.

David O’Malley, MD:I think that’s really important, right? We’re talking about occurrences as low as 1% to 2%, but very effective on treatment. Even in the mismatched repair protein, it’s most commonly quoted in the serous cancer. But if you look at the occurrence in mucinous and clear cell, it’s actually quite…

Gregory Riedlinger, MD, PhD:Endometrioid.

David O’Malley, MD:And endometriod. Excuse me, I almost forgot that. You’re getting up there to numbers that are not that uncommon.

Robert L. Coleman, MD:Let me ask you, if you had your choice of the primary specimen, because you had a primary debulking, or this biopsy that you took at recurrence, which 1 would you send?

David O’Malley, MD:If I took a biopsy at recurrence and there was adequate tissue to do what I call next-generation sequencing…

Robert L. Coleman, MD:Massively parallel...

David O’Malley, MD:Massively parallel generation... If it is adequate, which it often is not, I would send the recurrence, but we usually send the primary.

Robert L. Coleman, MD:Yeah. Shannon?

Shannon N. Westin, MD:Same. I think there is very little data to say whether there’s much difference. There’s data to say that within the same tumor there’s a difference in different tumor types. But the small retrospective studies that have looked at primary sequencing versus sequencing at the recurrence haven’t shown much change or much difference.

Robert L. Coleman, MD:Yeah, we’ve seen a fair amount of consistency.

Shannon N. Westin, MD:As you said, Dave, you’ve got something; you just took the biopsy. It’s the most recent tumor, and it makes sense to study that if you can.

Robert L. Coleman, MD:I know in ARIEL2, when we were looking at rucaparib, we were looking at fresh and original tissue or historical archival tissue and found out it was actually pretty close. It wasn’t very far off. I think the issue that comes up as we go forward is, are there new mutations that happen? Has it become more informative? And it’s expensive. And so how much? And ovarian cancer, frankly, doesn’t have a lot of mutations. It’s a lot of copy number alteration problems that we don’t have drugs for.

David O’Malley, MD:Some of this is going to go away a little bit with circulating DNA. We’re going to have some of that. We’re going to be able to do this multiple times, potentially. But I think what really is effective: As we’re going to talk about PARP [poly (ADP) ribose] inhibitors, have they undergone a reversion mutation? That would be 1 reason why I would use the recurrent tissue.

Transcript edited for clarity.