Molecular Testing in Treatment Selection

Lori J. Wirth, MD: Let’s move on to gene-directed therapy, which has been another avenue of progress in the world of thyroid cancer in the last few years. I’m curious to know what you do at the University of Michigan in terms of genotyping your patients with radioactive iodine-refractory differentiated thyroid cancer [RAI-refractory DTC]. Let’s not talk about biopsying nodules and the role of genotyping in that setting, which is interesting but a separate conversation altogether. What do you do when patients are referred to you as a medical oncologist?

Frank Worden, MD: When patients come to us being RAI-refractory, the first thing we do if it hasn’t been done is perform panel sequencing with the internal platform we have that tests for a select number of genes. I have worked with our pathologists to make sure that all of the targeted therapies available for thyroid cancers are on that panel. It’s an RNA-sequencing panel. It takes about a week-and-a-half to 2 weeks to get back the results. We’re looking for those common mutations that we’ll talk about shortly, that we could potentially target up front. We also have our own in-house genetic sequencing panel that looks for germline mutations, as well as somatic mutations, that covers a broader array. Sometimes we’ll pick up other mutations that may provide some insight, or benefit perhaps to prognosis or some other therapies, in addition to those targeted mutations that we just spoke about. Generally, that takes about 2 or 3 months to get back, so if we want something quickly because we’re trying to decide what we need to do, we use the in-house sequencing panel. I personally rarely use FoundationOne, Tempus, or the other national panels, just because I have an in-house panel that is reliable and quick to get back results.

Lori J. Wirth, MD: Do you have difficulty with insurance coverage? 

Frank Worden, MD: That’s a good question, Lori. The pathologists sometimes look for the most common mutations. If they see, for example, a tall cell variant that more likely than not is BRAF, they’ll test for BRAF. If the test is positive, they don’t have to worry about the other genes, so the insurance is more accepting to that vs the whole panel. But occasionally, I do have trouble with insurance coverage. In fact, at a tumor board today, I had a columnar variant of a papillary thyroid cancer that we were discussing, and the NGS [next-generation sequencing] panel was refused by the insurance company. Our internal MI-ONCOSEQ [Michigan Oncology Sequencing Center] panel was covered under the confines of a clinical trial, so I was able to receive data that way. 

Lori J. Wirth, MD: Do you ever turn to the NCCN [National Comprehensive Cancer Network] guidelines when you’re arguing with insurers to get coverage because the NCCN guidelines do point out that there is a role for genotyping in RAI-refractory DTC? 

Frank Worden, MD: Yes, we do write letters. A lot of times the panel sequencing is done through the pathology group in terms of the insurance provider and not necessarily medical oncology. The drug therapies come through us. Our internal pathology group is aware of the panel guideline recommendation and uses that to argue. Since it’s not directed therapy in terms of medical therapeutics, sometimes the companies are more reticent. 

Lori J. Wirth, MD: Do you go to the primary tumor and have that tested? Do you feel strongly that you need to test a progressing metastatic lesion? What sort of approach do you take? 

Frank Worden, MD: The University of Michigan’s sequencing program prefers the freshest tissue if it’s available, easy to attain, and the patients are amenable to biopsy. A lot of times in thyroid cancer, we have peripheral lung nodules, as you know, or hepatic lesions that are easy to get. So, if patients are amenable to getting a biopsy, we will do that. However, the presence of the driver mutation is usually there from the beginning of the metastatic disease. Oftentimes, we’ll use archival tissue. Now with the COVID-19 [coronavirus disease 2019] pandemic, because of the risk of biopsies, our group will not biopsy patients, so we have to use fresh frozen tissue for both the internal panel as well as the sequencing throughout the clinical trial. But if for some reason, patients don’t have tissue readily available—if the patient comes from a different part of the state or a different part of the country and we can’t attain the tissue—then we will probably put them through a biopsy, but that rarely happens. 

Lori J. Wirth, MD: To your point about the driver alterations for all of the alterations that we’re talking about, for which drugs and potential clinical trial options are available, they’re all there from the start. For the initial genomic testing, in general, the primary specimen is perfectly fine. Also, it’s important to note that these driver alterations are mutually exclusive. If you’re taking a sequenced approach where you rule out BRAF mutation first because that’s the most common, you’re going to see it in 60% of the PTCs [papillary thyroid carcinomas], if you see it, you don’t have to go looking for an NTRK fusion because you’re not going to find it. 

Frank Worden, MD: Correct. 

Lori J. Wirth, MD: That is a feasible and rational approach to take. I’m sorry, you were going to say something. 

Frank Worden, MD: It’s interesting you bring that up, too, Lori. One of our pathologists, Tom Giordano, MD, PhD, will oftentimes say, “Well, I can stain for BRAF,” because if 60% of these are going to be positive, they would save the insurance issue and sometimes cost. For a lot of our tall cells if they can be easily stained, we’ll go ahead and stain for that first, and if it’s negative, then we’ll proceed with the paneling.

Lori J. Wirth, MD: I agree. The BRAF immunohistochemistry does work very well, so we rely on that a lot. 

Then on the other end of the spectrum, it is trickier to test for the fusions. As we know, RET fusions are seen in slightly less than 10% of all PTCs. They’re seen much more frequently in pediatric and young adults, and then even more frequently in radiation-induced cancers. We see NTRK fusions as well in a small subset of thyroid cancers. ALK fusions are uncommon, and we’ve even seen a ROS1 fusion. We know that they’re out there. RET has now more than 20 five-terminal prime fusion partners that have been identified. In NTRK we seen NTRK1 and NTRK3 fusions, and both of those have multiple 5-terminal prime fusion partners as well. It’s a pretty complicated set of molecular diagnostics that we’re talking about, if we’re going to be identifying all of the patients who harbor actionable fusions, which is a challenge. What do you guys do?

Frank Worden, MD: We’re fortunate at the University of Michigan. Arul M. Chinnaiyan, MD, PhD, has a multimillion-dollar laboratory set up there devoted to sequencing and NGS. That was his project that he started as a junior faculty in pathology and has continued to this day. In terms of that, he has been in tune to the presence of these various fusions in these cancers. Then working with the various oncology groups at the University of Michigan, we are able to make sure that all of those fusions are present. 

The pathology sequencing, which is a shorter panel, will highlight the major fusions, and that’s part of the reason why I offer patients both. Once in a while I pick up on some germline mutation, but if there is a fusion perhaps that isn’t seen in one panel, it can be picked up on another. Occasionally I see that, but for the most part, the most frequent fusions at the University of Michigan are also found in these panels. This is probably true for the more national or standard panels as well. That’s how we have dealt with it. Other centers handle it differently. Perhaps you do things differently there at Harvard Cancer Center, but that’s how we approach it here.

Lori J. Wirth, MD: I’m similarly privileged to have good molecular diagnostics in our pathology group. John Iafrate, MD, PhD, created the Archer assay, which is agnostic to the 5-terminal prime fusion partner, so we can detect rare fusions that aren’t necessarily on panels. We have identified patients with previously nonreported fusions, and when we say less than 10% of patients with PTC harbor RET fusions, we’re probably undercounting them. We probably have been historically missing some, and it may not have mattered so much in the past, but it matters now that we’ve got drugs that target these alterations and work pretty well.

Frank Worden, MD: Absolutely. 

Transcript edited for clarity.