This has been fantastic. We are at the end of our time. I want to thank all of you for this rich and informative discussion. Before we go, I would like to conclude and get some final thoughts from each of you. I’ll start with Jonathan and pose the question to the 4 panelists. What do you expect to see in the future in terms of biomarker-based oncology? Is there a biomarker out there that you’ve got your eye on that you think is going to change practice? What are your thoughts about the future, in the form of a closing remark?

 I would like to thank the other panelists; this has been a great session. I’ve learned a great deal, and it has helped me and it’s certainly going to impact the way I look at some of these more rare mutations that you might see in thyroid or lung cancer. I have also been surprised at how many targetable, biomarker-detectable mutations and translocations we’re still finding. We’ve been doing NGS [next-generation sequencing] for some time, and we’re still identifying new targets. Academic institutions and pharmaceutical companies are still developing new therapies for rarer and rarer conditions. We just recently identified a mutation in isocitrate dehydrogenase in chondrosarcoma. Seventy percent of patients with chondrosarcoma have this mutation. We have a new targeted therapy that targets the oncometabolite, and it gives me hope that between the biomarkers and the development, we’re going to continue to impact cancer.

 Mark, thank you so much for being such a great host. It’s really nice to be on this panel with everyone. And thank you to 

® for hosting us. Just to build on Jonathan’s comments and then make 1 very specific comment—biomarker-driven treatment is here to stay, and we’re going to be continuing to evolve our treatment and involve our patients in these therapies. Acquired resistance will emerge on the potent, specific inhibitors, and our next-generation inhibitors are in development already. This story does not have an end. Biomarker-driven therapy is here to stay. For advanced thyroid cancer—whatever histology, whether it is anaplastic, medullary, or a differentiated thyroid cancer—patients with advanced thyroid cancer are poster children for gene-specific therapy. There is a very high hit rate across all the various histologies, and we absolutely should be genotyping all those patients who need systemic therapy.

 Thanks so much. What a great group. Certainly, I share all the enthusiasm about how we can really transform our patients’ lives with the right drug at the right time. What will also change, and what I look forward to—aside from more targets and an understanding of resistance—is also how we can better personalize therapy for a patient in terms of escalation or de-escalation. Are there opportunities, when treating an early stage patient, to detect minimal residual disease [MRD], as Jonathan had talked about earlier? Can we treat those patients? Are there ways that we can layer on multiple drugs if we do not have the biochemical response that we would like to see? My sense is that, in the next years, we will incorporate prevention of resistance, and also really tailor therapies to patients over a much longer journey.

 It’s tough to go last because everyone stole my thunder. I just want to piggyback on what Jyoti said, because this is a “Go big or go home” comment. Two years from now, I hope we are assembled here again, talking about all the data to leverage these drugs in the early stage of disease for a cure. We’ve sat here for the past 90 minutes talking about the role of these targeted agents, advanced medullary thyroid cancer, lung cancer, and sarcoma. But for these drugs to be truly transformative, we have to leverage them for cure. We have already seen this with 

 story, and this is a call for initiatives, IITs [investigator-initiated studies], cooperative groups, and industry trials looking at these drugs in the neoadjuvant setting—or in the adjuvant setting for better chance for cure. This highlights the importance of testing, not just for patients with advanced-stage diseases but for patients with early stage disease as well. I look forward to that. Jyoti made some really good comments about utilizing CTDNA [circulating tumor DNA], MRD testing, and other technologies to decide who gets these drugs and who does not. That’s important. When we start to consider cost and toxicity, personalization is not about the genotype. It’s about understanding who really needs the drug in the first place. Those are all great points.

As the moderator, this has been a complete pleasure. I want to thank the 4 of you—you made my job easy. It was a great discussion. Thank you all, and thank you to our viewing audience. We hope you found this 

® to be useful, informative, and helpful in your practice.

Videos

Experts in oncology discuss future directions in biomarker-driven therapy. 

Future Directions in Biomarker-Based Oncology 

 Before we close, I want to talk for a couple of minutes about NRG1 fusions. This is a fusion of the neuregulin 1 gene; these are rare tumors. They do occur in less than 0.5% of many solid tumors, including lung, breast, ovarian, colorectal, pancreatic, and gallbladder tumors and sarcomas. This is primarily a fusion protein that retains an active EGF-like domain and drives proliferation through aberrant HER3 activation. Like most drivers, it is typically mutually exclusive of other drivers. We do have experience with a monoclonal antibody, selumetinib, which blocks the HER3 dimerization as well as HER3 dimerization with other HER family members. There has been a little activity, or a little data presented. What I am familiar with mostly is the NRG1 fusion–positive GI [gastrointestinal] tumors in which there have been a number of responses quite early in the drug development. There is an ongoing trial called CRESTONE, which is looking at a number of cohorts that express NRG1 fusions in this setting. With that background, is anyone seeing this in lung cancer? Jyoti, you are shaking your head yes. Do tell.

 We have picked up a couple of these patients with this NRG1 fusion, and we have been enrolling—we have CRESTONE open as well as the Tarlox trials—trials testing tarloxotinib, which is an epoxy-activated EGFR TKI [tyrosine kinase inhibitor]. There is a cohort of patients with that experience as well as a few patients that we are starting to screen for CRESTONE. These are the early days in terms of activity. Our patients with lung cancer include nonsmokers, young patients, and Asian patients. There is the enriched population of people you would treat with a fusion. Patients, historically and often, have been treated with afatinib off-label with high toxicity and low efficacy, so we will see. Certainly, there has been some early activity in our practice, but there have been really small numbers.

 Ben, have any been identified in the Hopkins practice [Sidney Kimmel Comprehensive Cancer Center at Johns Hopkins]?

 No, I don’t think we’ve identified 1 across the board. Keep in mind that we just moved to a fusion panel about a year ago.

 It does. It’s an important alteration. It’s extremely rare, but there may be some tip-offs, histologically, that may clue you in. Some of the data would suggest that mucinous adenocarcinoma [MAC] is a subtype that may be more likely to harbor these, at least based on retrospective analysis. Yes, it is a rare fusion. We are weaving another story, certainly not chapter 1, 2, and 3 of the targeted therapy story—maybe it’s an appendix—but it’s something we’re going to have to routinely identify because there will be some novel therapies coming down the pike that will elicit meaningful activity. Afatinib, as Jyoti mentioned, does have some case reports. The response to some of the case reports is modest, but this is a new alteration that we’ll need to keep in mind when we start routinely testing or when we continue to routinely test our patients.

 Jonathan, when treating sarcoma, do you have any experience with the NRG1 fusions in your center [Sylvester Comprehensive Cancer Center]?

 What I can tell you is that it’s very rare. I have not seen any patients at our center with this. If we are looking in the published literature, then there has been a reported rate of 0.2%, or 1 of 620 patients who were tested. It is very rare, as far as I can tell, in sarcoma. This goes back to our biomarker testing that we discussed earlier today, and the fishing analogy: We are not going to find these if we do not look, so we just have to continue looking. We will find them.

 Yes, at least we have some agents—even though they are early in development—that look as if they clearly have activity. That’s certainly something to build on. Lori, any comments here in the NRG1 space?

 Living in the Boston area, I am a big lobster aficionado and I am trying to figure out how that big fishing net can incorporate lobsters in it—how can I do this for my analogies? In the same way that lobster really is exquisitely delicious, we do see some exquisitely delicious responses when we are using our gene-specific therapies for our individual patients, and those responses really are life-altering for our patients. It’s just phenomenal, so the more genotyping that we’re doing with broad-based panels that can identify these mutations, amplifications, and fusions, the more patients who will be benefiting in the long term.

Dr Mark Socinski describes the design and findings of the CRESTONE trial. 

NRG1 Fusions in Solid Tumors: The CRESTONE Trial

 I’m going to ask Dr Patel to give us her thoughts about the subsequent RET inhibitor, pralsetinib, and walk us through the ARROW trial. Those populations were very similar to what we saw in the LIBRETTO-001 trial, which was very nicely summarized by Lori.

It is a very similar. The pathway was perhaps the Blueprint Medicines Corp—the company. Thank you, Lori, for smiling. That took me a minute.

Thank you for that. The ARROW trial was a registrational phase 1/2 study design in which MTC [medullary thyroid cancer] and non–small cell lung cancer were both well represented. It was any 

 alteration. The lung cohort was reported first with the fusions, and similarly, patients could come on as treatment-naïve. Those patients had response rates that were 73%; the response rate of patients with prior platinum treatment was a little bit lower at 61%, but remarkably 100% of patients had some tumor shrinkage. The waterfall plot went across the board. Again, regardless of the fusion partner, we saw a long duration of response as well. Toxicity was mild, with about 20% of people experiencing hypertension or some transaminitis.

The updated data set for medullary thyroid cancer with 

 aberrations was a little more recently presented. That included 53 patients. The response rate was 60%, with some complete responses. The median duration of response, again, was not reached. For patients who were treatment-naïve, their response rate was about 82%. Again, we are seeing this broad activity. Anecdotally, for patients with lung cancer who we treated with platinum prior to this—patients who had received some immunotherapy as part of their prior platinum treatments—I think had a little more toxicity with both of these TKIs [tyrosine kinase inhibitors]. Patients who received platinum alone seemed to tolerate treatment really well. I think we are seeing how that evolves as more patients are treated. Regarding CNS [central nervous system] responses, I had a couple of patients with leptomeningeal disease who had good responses. Both drugs are quite active.

 Obviously, the CNS data come from small patient numbers, but it is pretty active in the settings that we’ve seen. Ben, what is your sense of the selpercatinib versus pralsetinib comparison?

 It may be another Coke versus Pepsi situation here. I have a lot more experience with selpercatinib than I do with pralsetinib. I think both of them—if you look at the waterfall plots of both of these drugs—look fairly similar, even though the response rates are a little higher in the selpercatinib data versus the pralsetinib data. Each of them has different toxicities with which we must wrestle. We were fortunate enough to participate in Lori’s trial with patients with thyroid cancer, and we also had patients in it with lung cancer. It was a very well-tolerated drug. Again, it is a win to have this embarrassment of riches compared to what we had before. I think both are extremely well tolerated and elicit meaningful responses below the neck and above the neck, intracranially. I believe only 2% of patients in the selpercatinib study had to come off therapy for any reason, and only 4% of those in the pralsetinib had to come off treatment. It speaks to the selective nature of these drugs and how well tolerated and how targeted they are.

 It is interesting; we talked a little about this at the very beginning, the role of molecular testing at the time of disease progression. I think there has been at least 1, if not more, small series presented about this topic…. It was patients with 

 fusions as their acquired resistance driver. There has actually been activity in that setting, which is interesting, and although it is quite rare, it would be part of the argument for retesting at the time of progression in other oncogenic driver subsets. I am not sure if it has been described in any other setting other than the 

We contributed to that experience, and it is an interesting progression when you have off-target progression. This is when you see more of the fusions with really potent inhibitors. I think, for drugs like selpercatinib, which will end up being great on target, there are hot spots where you will pick up some more off-target mutations. That’s certainly been noted with the 

 experience as well. Some of these toxicities are overlapping though, QTc [corrected QT interval] prolongation and transaminitis, and so we need to be careful when we start mixing.

 I completely agree that we first consider safety. We have to be sure, particularly when treating some of our patients who may have other comorbidities. I think it is great to have 2 options. I mentioned before, regarding “The Pathways of Progress” that Ross Camidge MD, PhD, and I co-chair; we had this debate where we considered selpercatinib versus pralsetinib. You can probably guess that it was a Coke versus Pepsi thing. We could not distinguish based on toxicity. We could not distinguish based on efficacy. There is not enough experience with these drugs for us to confidently know any cost issues, but that may be the leading factor moving forward. We know that the pricing and cost of these expensive oncolytics in the oral setting can vary based on contracts and so forth. There are a lot of things to learn in that particular area.

Dr Jyoti Patel presents data from the ARROW clinical trial. 

 

Tumor Agnostic RET Inhibition: The ARROW Trial 

Lori, talk to us about the LIBRETTO-001 trial. It obviously changed our world, at least in the realm of lung cancer treatment, from the 

 fusion point of view. Talk us through the design and so forth.

One thing I want to comment on is that, interestingly, the dirty kinases like vandetanib used in MTC [medullary thyroid cancer] have much better activity than in 

fusion non–small cell lung cancer, which is such a curious thing. I do not know why that is, but we can see really good responses with both vandetanib and cabozantinib in patients with MTC. However, we do now have specific inhibitors. Selpercatinib was designed along the same lines as larotrectinib, to potently and specifically inhibit just the RET kinase as much as possible and have as little off-RET inhibition as possible. It was also particularly designed to have as little KDR, VEGFR2 activity as possible, with the idea that with the MKIs [multikinase inhibitors] that do have RET activity, that off-target toxicity not only causes a lot of adverse effects, but it also limits the dose escalation you can get. So you cannot get really potent RET inhibition because of the dose limitation factors of the other kinase inhibitors.

Selpercatinib was designed just to hit RET as hard and as specifically as possible. Pralsetinib was designed similarly. Selpercatinib was also designed to inhibit all of the 

 fusions that we see in non–small cell lung cancer and other cancers. It was designed to inhibit all of the known 

 mutations seen in medullary thyroid cancer. Then it was also designed to head off acquired gatekeeper resistance mutations at the codon V804, which was hypothesized to account for acquired resistance. Subsequently, that was shown in several patients who were treated with cabozantinib or vandetanib and then developed resistance. The LIBRETTO-001 trial was a phase 1/2 trial that enrolled patients with 

-altered cancers. There were several different cohorts, including patients with non–small cell lung cancer that had been previously treated, and non–small cell lung cancer with 

 fusions that had not been previously treated. There were 3 thyroid cancer cohorts. The first thyroid cancer cohort included patients with 

-mutated MTC who had previously received either cabozantinib, vandetanib, or both. Then there was a cohort of patients with 

-mutated MTC who had received neither cabozantinib nor vandetanib. There was also a smaller cohort of patients with 

 fusions and advanced thyroid cancer who had been previously treated.

There were 2 separate publications in 

. The lung publication reported on the primary analysis set of all of the consecutively enrolled patients with lung cancer who had been previously treated. Then we did the same with the thyroid cancer publication. The primary analysis set concerned patients with 

mutated MTC who had previously been treated with cabozantinib, vandetanib, or both. In that primary analysis set of patients with medullary thyroid cancer who had been previously treated, the overall response rate was 69%. When we looked at patients who hadn’t received prior vandetanib or cabozantinib, the overall response rate for that patient population was 73%. In that smaller group of patients with advanced 

 fusion thyroid cancer, their overall response rate was 79%. These responses are all very durable. In the primary analysis set, the median duration of response and median PFS [progression-free survival] were not yet reached by the time of the data maturity for the publications. In terms of non–small cell lung cancer, we see similar degrees of activity; the previously treated 

 fusion cohort had an overall response rate of 64%, with a median duration of response of 17.5 months.

There was an 85% overall response rate in patients who had not been previously treated. In terms of safety, we see primarily grade 1 and 2 adverse events, and these are reversible. There are some treatment-related grade 3 or 4 adverse events that we observed, the most common of which included hypertension, transaminitis, and diarrhea. A small number of patients had QTc [corrected QT interval] prolongation; a couple of patients had grade 3 QTc prolongation. That is certainly something of which to be aware. Overall, however, selpercatinib is very well tolerated. One of the ways we can specifically take a look at tolerability in our patient population with thyroid cancer is by analyzing the frequency of dose discontinuation. With the other multikinase inhibitors, we often see rates of dose discontinuation of 15% to 20% of patients in the trials who have thyroid cancer. With selpercatinib, only 2% of patients had to have dose discontinuations due to treatment-related adverse events.

One of the questions I have had about selpercatinib concerns how much VEGF effect it may have. I cannot remember, what was the grade 3 hypertension rate? It was not insignificant, if I remember correctly.

Yes, it was 11% or 17%, or something like that.

I think it is about 11%, if I am remembering correctly. S. Mike Rothenberg, MD, PhD, is a friend of mine who was involved in the development of selpercatinib early on; he would say, “There’s definitely a little tickle of KDR there.” So we are seeing some off-target VEGFR2 activity, but certainly, it is nothing compared to our other friends we have discussed.

Lori Wirth, MD, reviews the phase 1/2 LIBRETTO-001 trial. 

Tumor Agnostic RET Inhibition: The LIBRETTO-001 Trial 

We are going to transition to our next topic; we will be talking about 

 in this setting. I am going to ask her to set the stage for the role of 

 in oncology, why don’t I start out by discussing the identification of 

rearranged in transfection? This was first discovered, I believe, in 1986 by Masahide Takahashi, MD, PhD, in an NIH3T3 cell line. Then 1 year later, Alfredo Fusco, MD, in Italy identified 

 fusions in papillary thyroid cancer. Then shortly after that, the 

 gene was cloned in patients with familial MEN2A and MEN2B [multiple endocrine neoplasia type 2A and 2B] syndromes, causing as we know medullary thyroid cancer, as well as the other endocrine neoplasias associated with those syndromes. It was not until 

 was identified as a fusion driving non–small cell lung cancer that the pharmaceutical industry really got interested in potently inhibiting RET. We did have the multikinase inhibitors like vandetanib, cabozantinib, and others that do have anti-RET activity along with VEGFR2, KDR, EGFR, and so forth.

I think it took the identification of KIF5B 

 fusions in non–small cell lung cancer to really turn the spotlight specifically on RET. RET is a juicy target because it is a receptor tyrosine kinase. Interestingly, it does not homodimerize exactly with itself with ligand binding, but there is a GDNF [glial cell line-derived neurotrophic factor] coreceptor that the ligand binds to, and then that coreceptor leads to homodimerization, forming a complex of activated 

. There is signaling, as a result of that, through multiple signaling pathways. 

 fusions are, of course, seen in papillary thyroid cancer and other thyroid cancers in the subset of patients with non–small cell lung cancer. They can be rarely identified in other cancers, including pancreatic adenocarcinoma. Then we see 

 point mutations, deletions, and a couple of other alterations in the majority of patients with medullary thyroid cancer, whether it’s a germline mutation in MEN2A or MEN2B, or a somatic mutation arising in a sporadic cancer.

In that situation I’m under the impression, Lori, that you might see 

 mutations in other tumors, but it is really felt to be a driver only in the setting of medullary thyroid cancer. Is that a fair statement?

I do think it is a fair statement. Again, I think it is because 

 is not necessarily expressed in other tissues. A point mutation is not going to have any sort of oncogenic effect if it is not expressed to begin with.

Let’s discuss the impact of fusions and mutations on clinical outcomes. Before we had the real potent, modern-day RET inhibitors, how effective were those drugs, like sunitinib, vandetanib, and cabozantinib, and those other “dirty kinase” treatments, as we call them? What was your experience prior to the modern era?

There are MKIs [multikinase inhibitors] that are approved for iodine-refractory differentiated thyroid cancer that have anti-RET activity. There are also MKIs approved for medullary thyroid cancer. There is activity in cases of DTC [differentiated thyroid cancer] whether or not 

 fusions are present. There is also good activity in cases of MTC [medullary thyroid cancer] whether or not 

 mutations are present. However, at least regarding patients taking cabozantinib who are being treated for medullary thyroid cancer, a subset analysis correlating activity with a specific 

 mutation did show that it looked like cabozantinib is more effective in patients who harbor a 

 M918T mutation, which is in the tyrosine kinase domain. There is even a survival benefit that was seen in that particular patient population.

It is not too surprising, because in MTC, there is a really interesting genotype-phenotype correlation. The various 

 mutations do cause various degrees of aggressiveness of the cancer. In the familial syndromes, we see not just aggressiveness in terms of the age of onset of medullary thyroid cancer, but also in the risk of developing the other endocrine neoplasias like hyperparathyroidism and pheochromocytomas.

I’m under the impression that, at least in thyroid cancer, the germline issues are more commonly seen than the somatic mutations. Is that fair?

Not quite; about 25% of medullary thyroid cancers will be seen in patients who have MEN2A or MEN2B. They all have germline 

 mutations; MEN2A germline mutations are more common because MEN2B is just very aggressive. Of the 75% of patients who do not have the hereditary MEN2A or 2B syndromes, approximately 60% will have somatic 

 mutations driving their MTC. Overall, the somatic 

 mutations are more common in MTC because sporadic MTC is more common. Getting back to tumor testing and so forth, this germline versus somatic mutation adds a whole other level of confusion regarding how to test. One thing to point out is that we are talking mostly about NGS [next-generation sequencing] testing in today’s program, but for patients with medullary thyroid cancer, we absolutely have to make sure that those patients do not have a germline 

 mutation. It has implications not only for their individual health and care, but it also has implications for all of their first-degree family members.

Thank you. Ben, I want to give you the microphone for a second here, but I am going to limit your comments. When I first went to UPMC [University of Pittsburgh Medical Center], in 2011, or shortly thereafter, we were testing for 

. Occasionally we did find it in patients with lung cancer. In my experience with the drugs we mentioned before, like cabozantinib and sunitinib, the multikinase inhibitors have a lot of off-target effects and were not particularly active in the 

 fusion-positive lung cancer population. I am wondering if you had a similar experience. I do not want you to talk about the newer agents selpercatinib or pralsetinib because I want to give Lori the stage for that in a second. But there has been such a change in our ability to target this alteration that I want to give people a sense of what it was like prior to the modern era.

These drugs have already been bounced around during the discussion, and we have tried them in a lot of these tumors where we have identified some sort of 

 alteration. Vandetanib, cabozantinib, sorafenib, sunitinib—these were all drugs that may have had some RET activity. Now that we have these drugs, they do not elicit the meaningful activity that we are now seeing with these newer drugs, and they are not as well tolerated. I think this has been a win to have these new drugs, but certainly, dirty TKIs [tyrosine kinase inhibitors] are just that. They are not that selective; the depth of response and meaningful response are not as common, and there is more toxicity to deal with. I can recall several experiences with some of these drugs where you are just not getting a lot of activity out of what you think may work. It is an oral therapy. A lot of patients may think, “I am not getting chemotherapy, so maybe it will work,” but my experience with these treatments has not been as favorable as with the newer drugs.

Expert oncologists discuss RET-driven therapy for select solid tumors.

RET as a Therapeutic Target for Solid Tumors

 I want to turn to Jonathan. Ben gave us a nice overview of larotrectinib, but do you have any other comments with regard to the clinical development of larotrectinib and other end points that might be of interest?

 Larotrectinib certainly is the first NRTK inhibitor that was studied. It has been studied in a number of studies, and researchers have looked at many subsets in this entity. One of the subsets that has been looked at is the sarcoma subset. As I mentioned, in the subset, the most common histology was spindle cell sarcoma or undifferentiated pleomorphic sarcoma. That number makes up a third of the sarcoma patients that have 

Other patients with different sarcoma subtypes can also harbor these mutations. These subtypes include GISTs [gastrointestinal stromal tumors], MPNSTs [malignant peripheral nerve sheath tumors], chondrosarcomas, and angiosarcomas. We generally have to be on the lookout for the NRTK trans-locations in all our patients with sarcoma. We have treated patients who participated in the larotrectinib study, but entrectinib is equal as active as far as we can tell in patients with sarcoma. I cannot really say if 1 is more appropriate than the other in a given situation, although they seem to be equivalent or reasonably equivalent from the available sarcoma data.

 What has your experience been like, assuming you have used larotrectinib and entrectinib, in terms of their adverse-effect profile? Do you see any major differences or any issues with this class of drugs?

 I have the most experience with larotrectinib. In my patients with sarcoma, I have not really seen any marked toxicity problems that would preclude me from prescribing either of these medicines. They are both very well tolerated, and we use a lot of TKIs [tyrosine kinase inhibitors] because of my patients with GISTs, who make up about a third of my practice. They are on imatinib, sunitinib, regorafenib, and ripretinib and participate in clinical trials.

We are really used to managing the side effects, so our patients seem to tolerate it pretty well. We aggressively manage the adverse effects. You have to realize that most of our patients with sarcoma had high-doses of doxorubicin or ifosfamide. Some of them have had high doses of methotrexate, so they are moving from this high-dose chemotherapy in certain situations to this treatment, and they tolerate it really reasonably well.

 Lori, what has your experience been like when treating thyroid, head, and neck cancer to date with these 2 agents?

 fusions in our patients with head and neck cancer with squamous cell carcinoma. In the secretory carcinomas that arise in the salivary glands, we identify them because we are testing. Our surgical pathology group will often do our fusion panel as part of diagnostic testing for salivary gland tumors, and they also look for thyroid cancers. We will often identify these tumors, but they tend to be fairly well behaved. Often, they are treated with surgery. They may have adjuvant radiation but do not need systemic therapy. We have a number of patients with thyroid cancer who have been enrolled in the larotrectinib and entrectinib trials. There certainly is activity. However, my go-to drug to treat my patients with thyroid cancer is larotrectinib at this point. It is just because there are more patients who have been enrolled in the larotrectinib trials with thyroid cancer.

Overall, there were 28 patients with thyroid cancer who have been reported who have received larotrectinib. Maria Cabanillas presented that data at ESMO [European Society for Medical Oncology Congress]. In differentiated thyroid cancer, there is a 90% overall response rate with larotrectinib. How can you not offer it? We do not see quite as high a response rate with anaplastic thyroid cancer, unfortunately. I cannot remember the exact number of patients with anaplastic thyroid cancer treated, but the response rate in those patients was 29%.

One of the other really nifty things with larotrectinib and 

-fusion cancers—there was a case report recently in the 

on a patient treated in France with larotrectinib harboring an 

 fusion who had iodine-refractory, widely metastatic, papillary thyroid cancer. They did a whole body radioactive iodine scan in this patient, who was on larotrectinib. They found that, on larotrectinib, the patient was converted from being iodine refractory—they did not take up any radioactive iodine—to taking up radioactive iodine. This suggests that if we block the regulated pathways with an NRTK inhibitor in these patients with 

 fusion, we may be able to redifferentiate those patients and treat them with a combination of drugs, like larotrectinib and radioactive iodine. You are going to be seeing more case reports emerge along those lines in the future.

 Has that been a one-off experience to date in this setting? Have there been others?

 There is a submitted paper, but it is not in print. I have a couple of off-label patients whom I am hoping to treat. I have success, and then I write up their case reports. But unfortunately this is a setting where, right now, it is very difficult to do a clinical trial because of the rarity of the patients and the expense. We would need to do a trial with numerous centers even to get a small number of patients. It is a real challenge in that regard because of feasibility issues.

 In your sense, you have heard the others weigh in. They have compared and contrasted larotrectinib and entrectinib. Again, I use the Coke and Pepsi analogy. Is that your position?

 In thyroid cancer treatment, there were so many more patients treated in the larotrectinib trial. We have more robust data. We have experience with entrectinib, including a patient of mine who is enrolled in the trial, but it is such a small handful of patients. It is a big glass of Coke and a smaller glass of Pepsi, perhaps. I do not know if that matters.

Jonathan Trent, MD, PhD, describes the safety and efficacy data from clinical trials of larotrectinib in patients with NRTK fusion–positive solid tumors.

TRK Fusion Inhibitors in Tumor-Agnostic Clinical Trials

 Ben, talk to us about available treatment options if you find 1 of these 

 It is important to delve into a little of the history of 

 gene fusion was discovered as an oncogene in 1982 in colon cancer. It only took 35 years to then develop a therapy, or therapies, that are 

directed. Thankfully, we have those now. The first drug that came onto the scene for 

fusions of cancer was larotrectinib. This was a study that enrolled, at the time, 55 patients—the study reported on 55 patients—with 17 unique 

fusion–positive tumor types. It was pretty remarkable because there were different cancers. The age range for the trial was also pretty remarkable. Patients were 6 months all the way to 70 years of age. The study featured a wide range of patients, with different ages and tumor types. These patients were highly pretreated. About a third of these patients had received 3 or more lines of therapy. A majority of these patients had received prior lines of therapy.

What we saw with this single-agent, oral TKI [tyrosine kinase inhibitor] was a response rate of north of 70%. We have learned, in later iterations of this study, that there are meaningful responses intracranially, in the brain. These responses are durable. Going back to the question you asked before, about the fusion partner, it didn’t seem to matter in this study. The drug was active regardless of fusion type, tumor type, or age. This is your perfect basket study that is emblematic of targeted therapy. Based on this study, larotrectinib was approved.

More recently we have experimented with entrectinib. This is a drug that is a potent TRK inhibitor, but it also hits ROS1 and ALK a little. It was designed to cross the blood-brain barrier, and we have data on this as well that are published. A small number of patients—50 to 60, I believe—showed lower response rates of around 60%. But these responses are durable, and there are intercranial responses as well. It is a perfect story for targeted therapy: higher, durable, and intercranial response rates, given the proclivity of some of these lung cancers to go to the brain.

The last thing I will say about both these drugs is that they are reasonably well tolerated. They do not offer any new adverse effects that we have not seen before, minus a few on-target adverse effects that we see, with proprioception, hyperphagia, as well as increased weight. Most of the adverse effects from both these drugs are reasonable and manageable. The clinical implementation, development, and implementation of these drugs highlights what we have said already. You must test. You need to cast that wide net in the lake. Then you can identify these fusions, and these patients can go on to genotype-directed therapy with NRTK inhibition and NRTK inhibitor therapy.

 It sounds as if you are describing Coke and Pepsi. How does one decide?

 It is hard. Obviously, we do not have any cross-trial comparisons, and I have already shown you my cards. I have treated only 1 patient for this. I used larotrectinib in that setting, and the patient did reasonably well and continues to do well. I do not know if I would necessarily think there is a specific advantage of 1 over the other. 

We certainly have intercranial response rates for these drugs, so they are both reasonable options. I am not trying to play it safe, because I do like to put my nickel down on a choice, but both are reasonable. In my case, I chose larotrectinib because it was around at the time of identification, and entrectinib had yet to be approved.

 I am involved with Elsevier, formerly Via Oncology’s Via Pathways. Ross Camidge and I are the cochairs of the lung committee. We obviously approved the pathway of larotrectinib because it was the only 1 in town at the time. When entrectinib was approved, we were asked by Via Oncology’s Via Pathways to make a decision based first on efficacy and then on toxicity. The committee could not differentiate based on either efficacy or toxicity.

The third variable is cost, and there was a significant cost difference between the 2 agents. That was a decision from the Via Oncology’s Via Pathways’ point of view, based simply on the cost of the 2 drugs. I will leave you hanging in terms of what the outcome of that was.

Benjamin Levy, MD, reviews the available treatment options for patients whose tumors harbor NRTK fusions. 

Treatment Options for NRTK Fusion–Positive Solid Tumors

 Let’s transition to a different topic. We are going to focus on NRTK-directed therapy and talk a little about it. I am going to ask Lori to give us just a brief overview of what we know about NRTK therapy and cancer. We alluded to this issue of fusions and point mutations and so on. Then I’ll ask the group how often we see this and so on. Lori, your thoughts?

 We have talked a little about the difference between fusions and point mutations. We do see oncogenic fusions involving 

 with multiple prime fusion partners across different solid tumors, even in hematologic malignancies. It is that fusion with the 5 prime fusion partners that is often a homodimerization domain, like a coil domain, that leads to the expression and activation of the kinase. 

 would not be expressed in normal lung tissue, so if there is a point mutation, it is not going to be expressed or active outside the setting of those oncogenic fusions. It is the fusion that is really the mechanism that leads to the carcinogenesis with 

The other thing to think about—the basic biology of it, which is what we’ve mentioned a bunch of times—is that there are multiple 5 prime fusion partners. The more we look, the more we find. In addition to RNA-based assays, if an assay is used that is agnostic to the 5 prime fusion partners, then you are going to happen upon fusion partners that probably are driver-fusion partners that have not been described in rare cases. Across solid tumors, 

 fusions are pretty uncommon. There are a few diagnoses. We’ve talked about a few of them already that are enriched for 

fusions, like infantile sarcomas. Then there is head and neck cancer.

Things that occur above the clavicle are poster children for 

-fusion cancer in adult solid tumors. Secret carcinomas that are salivary gland cancers almost invariably harbor 

 fusions. Then we also mentioned the papillary thyroid cancer, poorly differentiated thyroid cancer, and even anaplastic thyroid cancer can harbor 

fusions. In larotrectinib testing, for example, if you lump together all the adult patients who were enrolled, the most frequent adult patient population had tumors above the clavicles.

 Fusion partners are certainly important in testing, as you mentioned, but is there any evidence that the fusion partner makes a difference from a clinical outcome point of view?

 I am glad you raised that. As far as I know, no. We’ve seen that with 

 experience in lung cancer so much, but one would think that the 5 prime fusion partners would matter more than it seems to matter. The kinase that gets activated by that fusion, and that probably explains why. Generally, that kinase will be wild type, at least early on before acquired resistance develops.

 Jonathan, you live in a world with more frequent 

fusions. What are your comments about how common they are in your discipline?

 We do see them, but even in sarcoma these are rare entities. This is estimated to occur in less than 1% of all our sarcoma patients. However, there was a poster presentation at CTOS [Connective Tissue Oncology Society Annual Meeting] a couple of years ago that showed that, in subsets of spindle cell or unclassified pleomorphic sarcoma, the incidents could be as high as 6%. We do see it, but it is still rare. I agree with Lori completely on the fusion partners. We have a lot to learn in that space.

 My lung cancer colleagues, can you offer commentary on how common they are? Jyoti, have you seen 1 in lung cancer?

Yes, actually. I have seen several, but it is uncommon. They are found in 1% of patients. It is an enriched patient population. They tend to be people who have never smoked. They tend to be a little younger. Lung cancer is common. If it is less than 1%, you are still seeing almost 2000 patients a year who could potentially have an 

fusion. Histologically, there are also variants. I just took care of a patient recently who had 1 with more neuroendocrine differentiation. Others I’ve seen tend to be a little more adeno or adeno with a fair bit of mucin. It stands to bear that you do not know the results based on histology unless you test.

 We have seen 1, but the identification of this is a reflection of how bad our testing was internally at Hopkins [Sidney Kimmel Comprehensive Cancer Center at Johns Hopkins University]. We were not routinely testing for 

fusions until about a year ago with an RNA-based platform. We probably missed a few. For a lot of those patients who have never smoked, if they have tissue left over, we’ll send it to a commercially available RNA-based platform to protect these. They are uncommon, but that does not mean they should not be tested, so my experience has been limited to 1 patient.

 Can I crystalize the feeling of the panel and say that everyone agrees that testing for 

fusions should be part of your comprehensive panel across tumors? These do happen across a wide range of tumors, even though it is very uncommon or rare in some circumstances. Would we all agree with that statement? Any dissenters?

 I certainly agree, and that is the beauty of these panels: We are doing it to find not just the 1% or 2% of patients with lung cancer who harbor 

fusions, but we are looking for the other uncommon fusions as well. Even if this testing was just for identifying that 1% or 2% of patients with lung cancer, for those patients it makes a really big difference. It is 1 of the challenges we have: wrapping our heads around testing all patients for these rare entities and the expense involved. If we had cost-benefit analysis, we would see that the treatment options that are unveiled for that smaller percentage of patients makes such a large difference in their therapy that, in the end, the cost-benefit analysis would end up being favorable.

To add to what Lori is saying, it is not a zero sum game. It is not to get 

. You get both with NGS [next-generation sequencing], so it truly is offering a broad number of options to our patients with full testing.

 It is like going fishing with a big net, right? You want to catch everything you possibly can.

Dr Lori Wirth describes the importance and implications of testing for NRTK fusions in patients with solid tumors. 

NRTK Fusions in Oncology 

 Let’s close this session by talking about some of the unmet needs regarding biomarker-driven therapies in oncology. I’ll start with Jonathan to get his view of the challenges in this area.

 I agree with everyone. There are still plenty of unmet needs in relation to biomarkers. In cases where patients have sarcoma, we have looked at circulating tumor DNA [ctDNA] platforms and their ability to detect minimal disease in patients who are progressing, or even before they on CT scan. In patients with sarcoma, the circulating tumor DNA to detect progression prior to a CT scan or MRI is just not there. I would love to see a blood-based biomarker assay that was more sensitive than a CT scan, an MRI, or a PET [positron emission tomography] scan.

 Do you think we have that in front of us? I have often made the statement. We used to have a saying when I was in North Carolina: Don’t go looking for trouble until trouble comes looking for you. If you had a highly effective therapy that you were willing to transition to at a very early time point, that might be the case. The important aspect is to really prove the clinical utility of knowing that before you see resist progression.

Remember that, at least currently, all our criteria that we use in evaluating cancer therapy is focused on resist criteria. Maybe someday ctDNA will play a role in that, and maybe a PET scan will play a role in that. I do not know how you feel about PET scans, Jonathan, but I think they are 1 of the most overutilized modalities in cancer therapy. In treating lung cancer, where details of the chest CT are so important to decide on treatments, it is a really terrible-quality chest CT, even if we have a dedicated chest CT. Ben, what unmet needs are in this area?

 I would say a couple of things are unmet needs. One is we need to remember that, despite all this alteration in the diagnostic algorithm we have for patients, patients continue to be woefully undergenotyped in the community. Paper after paper has shown missed alterations, and not everyone is getting complete molecular testing. That is 1 unmet need. We still need to get the word out. 

We need to understand this has a lot to do with—there are many factors. Tissue may be insufficient. There is tissue heterogeneity. It is sending it to the right platform. All are important. Then there is the importance of liquid. We all may have a hard time finding a good interventional radiologist or pulmonologist, but we all know a good phlebotomist. It’s not that difficult to get 10 cc of blood and the turnaround time is pretty rapid. That is 1.

Second, we need to remember the pie is filling in for those with lung cancer with 

 G12C mutations and those who are positive for 

 [human epidermal growth factor receptor 2]. Now we have all these alterations. Remember, we still have a big slice of that pie—perhaps a slice the size of a New York City slice of pie—that is still unknown. We do not know what the genome type is, and that is a call to better understand tumor biology outside the other alterations that we already know.

 Your point about a clear unmet need relates to undertesting. We see this time and again, at least when treating lung cancer. The patients do not get a full coverage of all the things for which there are FDA-approved targeted therapies.

 Mark, let me just point out that is the case for treating lung cancer, where ASCO [American Society of Clinical Oncology] Clinical Practice Guidelines recommend testing quite strongly. Think about the other tumor types and how many more patients are not being tested.

 I guess I should ask Jonathan too. In the sarcoma world, is undertesting an issue?

 It is a huge issue. For instance, our patients with GI [gastrointestinal] stromal tumors—70% of them have a 

 mutation, for which we have a drug. The other 30% of patients are treated with a different drug, so 30% of our patients are predicted to be resistant to imatinib. Yet regarding patients with metastatic GISTs [gastrointestinal stromal tumors], according to the National Cancer Database, only 30% are undergoing any type of NGS [next-generation sequencing] or mutation analysis. Seventy percent of these patients are just being treated blindly. Some of these patients are treated with adjuvant therapy for 3 years. It is a really undermet need.

 Jyoti, do you have any thoughts on unmet needs?

Mine is a really practical 1. It revolves around understanding how we can better communicate results to providers. Often, we do not test, but hopefully education will help with some of that. As we send testing out, how do we incorporate results in an EMR [electronic medical record], so people know it’s actionable?

That other piece is just about a health system buying into the communication and infrastructure to make it as important as the initial histologic diagnosis. This is really the definer of the population. Team dynamics and education are very important. It is not just the physician. It is really that everyone on the team understands the importance of this send-out lab to what you are doing for the patient in front of you.

 I have always taught people I train that there are 2 things you absolutely have to get right for every patient with cancer. That is the stage and the diagnosis. In 2021, the diagnosis includes a comprehensive molecular panel because we have some highly effective therapies that we do not have access to unless you make the diagnosis. We can then transition to that therapy. We need to stress the education of these sorts of things. There really should not be a patient who is undertested, even though we all know that that really does happen on a basis that is far too frequent than what we would like.

I would like to believe in our practices, and that the people in this webinar do not undertest. We go to great lengths, even if we have to do another biopsy or plasma-based things. Remember the comment about plasma? In my opinion, it is helpful only if you find something, so there could be issues of sensitivity with the plasma. You have to go the extra mile if you do not think your test is reliable enough, particularly in our population of patients who never smoked. Sometimes a second biopsy is indicated to make sure you are not missing a driver.

Dr Mark Socinski and panel discuss the challenges and unmet meets of biomarker-driven therapy across solid tumor types.

Solid Tumors: Unmet Needs Related to Biomarker-Driven Therapy

 I want to go back to Dr Patel. So many of our targeted agents historically have been developed with a companion diagnostic that is looking for whatever the target may be. Personally, I have not necessarily found them to be that helpful in day-to-day practice, but what are your thoughts? What might their impact be, and how should we look at them?

 Sure, I think that is a great question. When clinical trials are being done that examine targeted therapies, it’s advantageous for pharmaceutical companies to develop a companion diagnostic so they can quickly screen and pay for their own test…and have a homogenized set of data that they can submit to the FDA. My experience is that, after that approval, a lot about the companion diagnostic goes away. We know that we have great testing in-house as well as through multiple vendors, so I think the need for that is diminished. We use it very infrequently. I guess a good example of that, which is ubiquitous, is PD-L1 testing. That has made the interpretation of clinical trials more difficult than it needs to be, based on test characteristics and the performance of immunohistochemistry. Outside of that, I think there is great fidelity between PCR [polymerase chain reaction testing], Sanger sequencing, and NGS [next-generation sequencing]. It probably is less impactful to people who are seeing patients.

 Ben, what are your thoughts on companion diagnostics, specifically in lung cancer? Do you find them helpful?

 Not particularly, to keep the answer short. I think there is a lot of confusion, Jyoti really highlights this. We have this PD-L1 testing; we had a companion diagnostic and a complementary diagnostic, if you recall. There was a lot of confusion around this when pembrolizumab and nivolumab were being developed. I think that, given the advent of NGS and fairly high concordance across platforms, I do not find them all that useful.

 Lori, I will come back to you. You had mentioned that you have 1 marker, the 

. I do not even know if there is a companion diagnostic for that particular regimen. I am not sure about that. What are your thoughts on companion diagnostics?

 I agree with the rest of the people on the panel; we live in the Wild West with so many different NGS assays available to us now. I think it makes much more sense for physicians to be able to use an assay that they know well, they know the operating characteristics, they know how to interpret the data, rather than to have to use a companion diagnostic for 1 particular diagnosis for 1 particular agent. Hopefully, we are getting away from the need to develop companion diagnostics along with targeted therapy drugs.

 I would think if we had to use a companion diagnostic for all the biomarkers we have in non–small cell lung cancer, we would use up tissue so quickly that….

Experts in oncology provide their insights on the use of companion diagnostics in the treatment of solid tumors. 

Companion Diagnostics in Oncology for Solid Tumors

 I am going to ask Jonathan here to give us his thoughts on, we’ve had a couple of indications recently over the past year or so for a pan-tumor biomarker; 

 comes to mind. How has that played out in molecular testing and oncology? What is the impact of these pan-tumor biomarkers in solid tumors?

 This is a great question and a great topic. I am a sarcoma medical oncologist. For patients with GI [gastrointestinal] stromal tumors, we have been doing and requiring molecular testing for close to 12 years. It has been a very active part of our practice. We were early embracers of this because we noticed that although there are 175 different types of sarcoma, GI stromal tumors are probably the most common. There are 6 different molecular subtypes of GI stromal tumors: 

 mutations make up about 70%, PDGF receptor mutations make up 10%. Another 10% is attributed to loss of succinate dehydrogenase.

Then there is the more rare subset, which is seen in about 10% of patients, and is divided among 

A panel of oncology experts in lung, thyroid, and sarcomatous tumors discuss the role and impact of pan-tumor biomarker testing. 

Jonathan Trent, MD, PhD, Sylvester Comprehensive Cancer Center

Articles