Non-Small Cell Lung Cancer EGFR Testing and Treatment - Episode 10
Expert oncologists review clinical scenarios to define best treatment practices for patients with EGFR-mutated non–small cell lung cancer.
Follow Gregory J. Riley, MD, PhD, vice chair of Clinical Research in the Department of Medicine at Memorial Sloan Kettering Cancer Center (MSK) in New York, as he conducts Morning Rounds with the treating team and discusses patient care with Lauren Welch, MSN, NP-C, AOCNP, a nurse practitioner from Tennessee Oncology in Nashville, and genetic testing in lung cancer with Jason Chang, MD, a pathologist from MSK. Riely leads the discussion by taking Welch and Chang through 2 different patient cases and asking their perspective on how they would appropriately treat each patient, how to manage the adverse effects (AEs) observed, and which tests are right for them based on their condition at presentation.
“When I see a patient up front, I used to think that if they had stage I through III cancer, I didn’t need to focus on biomarker work because the pathway forward didn’t depend on that. More recently, with data suggesting the role of neoadjuvant chemotherapy plus checkpoint inhibitors…we need to know these things in earlier-stage settings. The majority of the work for biomarkers has come from the advanced-disease setting, where we focus a lot on things such as PD-L1 [status and genetic] testing,” commented Riely.
RIELY: I want to briefly go over treatment for those with EGFR exon 20 insertions. Mobocertinib is a small molecule kinase inhibitor analogous to drugs like osimertinib [Tagrisso] or erlotinib [Tarceva]. It has an AE profile similar to that of EGFR inhibitors that we know, such as erlotinib.
The data that we have for mobocertinib is from a single-arm trial [NCT02716116] that was published in JAMA Oncology last year.1 In that trial, they identified patients with an EGFR exon 20 insertion who had prior platinum-doublet chemotherapy, and they enrolled them to receive 160 mg of mobocertinib. The trial’s primary end point was overall response rate [ORR]. For patients with prior platinum treatment, the ORR was 28%. This was a significant improvement from what we saw with other EGFR tyrosine kinase inhibitors [TKIs] in this context; the response was typically around 5% for other TKIs. It established that mobocertinib had efficacy. With a median follow-up of 13 months, we were able to identify that the median progression-free survival [PFS] was about 7.3 months. Using mobocertinib, [we could give] something to these patients who didn’t have prior targeted therapy available to them.
Another recently approved drug is amivantamab [Rybrevant],2 which we’ve been evaluating for several years. It’s an antibody that targets EGFR, but it’s a little different because it also targets MET. It’s what’s called a bispecific antibody. Amivantamab has been studied in classical EGFR mutations, and we’re waiting for some more data on how that works. In the meantime, it’s been studied in patients with EGFR exon 20 insertion mutations, and the best data for are from the CHRYSALIS trial [NCT02609776], published recently in the Journal of Clinical Oncology.3 In that group of 81 patients with EGFR exon 20 insertion mutations, the response rate was 40% [95% CI, 29%-51%] and the median PFS was about 8.3 months [95% CI, 6.5- 10.9]. These were both single-arm trials without a randomized comparator arm.
Ultimately, the efficacy of these agents is relatively similar. They were both studied in the same patient population of prior platinum-based chemotherapy. They are reasonable second-line agents and they’ve both been approved as second-line therapies for patients with EGFR exon 20 insertion–positive non– small cell lung cancer [NSCLC].
RIELY: How do you approach a patient when you need to inform them that they have cancer?
WELCH: That’s one of the hardest initial conversations to have with patients, especially when they may not necessarily trust you yet. You don’t have that rapport and you’re trying to explain the importance of pausing and waiting to get back biomarker testing results. Many patients have never heard of biomarker testing and when we meet them, they want to start treatment [right away]. We’ll often spend time in the initial consultation explaining what biomarker testing is, discussing the importance of testing, and the potential impact of testing on therapy selection. Patients often have questions about how long it will take to get the test results back. Patients who have those driver mutations will frequently ask if these driver mutations are hereditary [or if] family members or children tested need to be tested. Then a subset of patients will want to know how certain biomarker results may impact their response to treatment or even their survival.
RIELY: While patients are waiting for the results, do you give them any additional resources?
WELCH: [I use] a couple of different strategies, depending on the patient. My resource recommendations vary depending on where the patient is in their diagnosis and on their learning preferences. I’m very supportive of and involved with lung cancer patient advocacy groups, and I’ll often refer patients to LUNGevity or the GO2 Foundation for Lung Cancer. When trying to explain the importance of waiting, a physician I work with has a great analogy about taking a road trip to California. Since we’re based in Nashville, you could get in your car and go west and eventually land in California, or you could take the time to look at a map, figure out the most direct route, avoid some construction zones, hit a good place to eat on the way, and have a much more successful and faster trip in the long run. We’re trying to get to the point that if we get this testing back, and there’s something relevant that could inform first-line decision, you’ll have a better outcome.
Regarding resources, once we have biomarker results back to review with patients, I recommend LUNGevity’s excellent biomarker testing booklet, available in both print and PDF. Also, the Biomarker Collaborative Group has compiled an excellent list of advocacy groups for lung cancer–causing biomarkers. I’ll introduce patients to these different resources for education and to help patients and their caregivers cope. [I find that they] benefit tremendously when they’re engaged and have support from other patients with cancer.
RIELY: If a patient has a high PD-L1 expression and an EGFR exon 20 insertion mutation, how do you approach initial treatment?
WELCH: We always go back to the EGFR paradigm using osimertinib, and how it does not play well with immunotherapy. We want to be thoughtful about how we sequence therapies in that context. Our first recommendation would be to hold the immunotherapy, since we know that patients with EGFR mutations don’t tend to respond as well.
RIELY: When we think about any of these drugs—mobocertinib or amivantamab—they have an AE profile to consider. How do you discuss AEs patients might experience from these 2 drugs? How can they manage them?
WELCH: Although there are differences mechanistically that you highlighted, they both have EGFR-directed effects. [Patients do experience] rash, nausea, diarrhea, paronychia, mucositis, and potentially interstitial lung disease, pneumonitis, and cardiac toxicity, although these are much less common than the gastrointestinal issues. The diarrhea can be quite significant, especially with mobocertinib. The median time to onset is within 5 days. I’ll instruct patients to have loperamide at home in advance of starting mobocertinib and to begin loperamide immediately after the first loose stool. When talking about diarrhea with patients, I make sure I inquire about their [status at] baseline. It’s important to understand what’s normal for them so you can figure out the grade of the toxicity. We’ll spend some time talking about gentle skin cleansing and frequent skin moisturization, nail care, vinegar soaks, etc. With amivantamab comes the unique issue of infusion reactions. These reactions tend to occur with the first dose, which is why we divide the first dose into 2 infusions. Normally, that reaction is better with subsequent doses, but infusion reactions can be scary for patients, their caregivers, and even the clinic staff. These EGFR exon 20 insertion mutations are still rare, and the clinic staff is not going to have a lot of experience giving amivantamab. Make sure that you prepare the patient as well as the staff for the high likelihood of the infusion reactions so they aren’t caught off-guard.
RIELY: Looking at this patient case, how would you respond to the watery diarrhea?
WELCH: This happened when we were conducting the trial with mobocertinib at my institution. I had a patient who had been on the drug for several months. The first question was, were they taking loperamide? [We wanted to] make sure that the patient was optimally using the loperamide, which is up to 8 tablets a day, and utilizing other antidiarrheals like Lomotil. If the patient continues to have grade 2 diarrhea despite those interventions, then you’re having conversations about holding the mobocertinib until diarrhea resolves, and normally it takes 3 days or so for the diarrhea to get better. Then you have to consider dose reduction vs resuming at the same dose. Sometimes we’ll use prophylactic antidiarrheals. That can be helpful with mobocertinib when you know it’s going to happen, and you can try to get ahead of it prophylactically. Also, that single-dose reduction can be a game-changer for tolerance. Don’t be shy to do that. [Finally], another thing we don’t always talk about enough is the pill fatigue for these patients. The recommended dose of mobocertinib is 160 mg a day, which is 4 pills. If you’re having them max out on loperamide, that’s 8 additional pills. Then they’re taking potassium and magnesium supplements for diarrhea. It can be overwhelming—sometimes just simply too much—and you have to embrace the dose reduction.
RIELY: What molecular tests are given to patients with NSCLC?
CHANG: I’ll start by discussing some different platforms of molecular testing: driver gene–only panels and NGS panels. The driver gene–only testing will tell you the [status of] main driver alterations such as EGFR, ALK, and ROS1. The NGS will cover that plus additional information, such as passenger mutations and secondary alterations, or additional tumor suppressor genes. It gives us a more comprehensive picture of the entire molecular profile. Given that we must test for 8 biomarkers, it is very difficult to run sequential driver-only testing without exhausting all the limited tissue, especially for small biopsy samples. Therefore, NGS has become the recommended platform for this type of testing in most scenarios.
RIELY: How can clinicians mitigate patient fears when they don’t want to wait for the results of an NGS panel?
CHANG: Right now, our NGS panels have an average turnaround time of about 3 weeks. Obviously, that’s suboptimal from your standpoint. At MSK, we combine both approaches. We run the small panels that give you answers for EGFR, KRAS, and ALK status with a turnaround time of 2 to 3 days. At the same time, we also run the NGS panels, and then we can provide patients with a comprehensive evaluation of the molecular profile in about 3 weeks. Right now, that’s our approach, addressing both at the same time: We provide them with a quick answer covering the most important driver alterations, but trying to limit the tissue exhaustion issue.
RIELY: With a cytology test, are there any limitations for how you can approach both the diagnosis and molecular testing?
CHANG: It’s not so much of a problem for diagnosis. We often can make a diagnosis of malignancy on cytology specimens. However, we have to make sure we have sufficient tumor cells for molecular testing, which is another hurdle. We have investigated the sufficiency rate of molecular testing on cytology samples at MSK, and based on our experience, it seems that more than 90% of cytology samples would also be sufficient for our large-panel NGS testing. That suggests that cytology samples are not inferior to surgical pathology samples.
RIELY: There’s a difference between a single-pass fine needle aspirate [FNA] and samples from multiple passes of [endobronchial ultrasound]. Clearly, you need more than just a single-pass FNA to get a diagnosis and molecular testing.
CHANG: Exactly. I also wanted to emphasize that at MSK, we do a rapid on-site evaluation of cytology samples. We can immediately tell the interventional radiologist or pulmonologist who is doing the procedure whether the cytology samples are diagnostic. That helps with the sufficiency rate as well.
RIELY: Looking at this patient, what molecular alterations are at the top of your list?
CHANG: In this context, I would think about those alterations commonly associated with never smokers: EGFR, ALK, and ROS1.
RIELY: Is there more than 1 type of EGFR mutation?
CHANG: EGFR activating mutations typically occur in exons 18 through 21. In general, when we talk about EGFR mutations, we think about the typical and the atypical mutations. The typical mutations, those that occur in about 85% of cases, are either EGFR exon 21 L858R mutations or exon 19 deletions of various sizes. They respond quite well to EGFR TKIs. Then, the atypical EGFR mutations, such as G719S or L861Q, have variable responses to EGFR TKIs. Finally, we have the EGFR exon 20 insertions. Historically, those were thought to be somewhat resistant to EGFR TKIs, but now we have FDA-approved therapies for EGFR exon 20 insertion mutations.
I also wanted to make the distinction that although there are other types of point mutations that occur on EGFR exon 20, such as T790m, those are entirely different from the EGFR exon 20 insertion mutations. On the molecular report, we have to specifically look at the class of the mutations.
RIELY: What do we use to conduct molecular testing?
CHANG: The main issue here is liquid biopsy, or [circulating tumor DNA] testing, vs tissue-based testing. We have done tissue-based testing for a long time so we are quite familiar with the performance of tissue-based NGS assays. We have only started doing liquid biopsies in the last few years. With a liquid biopsy, the main issue is sensitivity. It depends on how much DNA is shed into the plasma, and on tumor size and stage as well. For early-stage tumors, the sensitivity will be a bit lower. However, many studies have shown the high concordance between liquid biopsy results and tissue-based NGS testing, so the accuracy is not so much of a concern; it’s the lack of sensitivity in the liquid biopsy setting. I would probably say that if you want to do a liquid biopsy up front, that’s totally fine, but if the result is negative on liquid biopsy, I’d recommend that you still consider sending tissue for NGS testing.
In some settings, liquid biopsy is superior to tissue-based testing because it reflects the global disease burden in the patient. You don’t have to sample multiple tissue sources to demonstrate the required resistance mechanisms. Often, you’ll be able to get a broad picture based on the liquid biopsy profile.
RIELY: What should oncologists know about PD-L1 testing?
CHANG: There are different [versions] of the antibody assay out there. It is up to each individual laboratory to validate their own assay to make sure it shows good concordance with the published recommended assay.
We have a quantitative score for PD-L1, and your treatment decision also depends on the specific cutoffs for the PD-L1 scores. It can be somewhat subjective as well because of the variability in the intensity of staining. The interpretation isn’t cut and dry, but we try to do our best.
RIELY: Does anything specific about this patient suggest waiting to begin treatment?
CHANG: This is a perfect scenario in which it makes sense to do both NGS testing and some rapid driver gene testing upfront. This patient, because of the demographic, has a very high likelihood of carrying an oncogenic driver mutation like EGFR or ALK; often, we can just do a rapid PCR [polymerase chain reaction]–based assay that can give you an answer in 2 to 3 days. The likelihood of finding an alteration would be very high, and then you can start treatment right away. PCR-based or single gene–based liquid biopsy testing are also [options], so we can do EGFR testing on liquid biopsy as well. Given that the patient has metastatic disease, the liquid biopsy for EGFR would likely be positive if the patient has an EGFR mutation.
RIELY: What is the top priority for an oncologist, interventional radiologist, or pulmonologist, especially if the patient has bone metastases?
CHANG: The top priority from my perspective, the one thing that we are most concerned about, is to obtain sufficient tissue to complete all the testing. Some additional challenges are associated with bone metastases, because regular acid decalcification degrades the DNA; you can’t run traditional molecular testing on acid-decalcified specimens. However, at our institution, we have validated EDTA [ethylenediaminetetraacetic acid] decalcified bone specimens, and they seem to be working quite well with our molecular platforms. In general, with other institutions that haven’t validated their EDTA decalcification procedures, they would probably have to obtain some other sort of specimen that is [not] decalcified.
RIELY: Would anything specific about this patient make you wait to determine a treatment path?
WELCH: The [patient’s demographics] as never smoker and of Asian descent are both clichés for having an actionable mutation. They’re clichés because they’re often true. With this patient in particular, we’d be talking about the likelihood of her having some kind of oncogenic driver mutation. It’s so important to select the right first-line therapy, because we know outcomes are better when we match people to the right one when there’s something targetable. Trying to emphasize overall outcomes will be better if we wait and get all the appropriate tests back.
RIELY: Let’s say this patient has PD-L1 expression of 55% and no driver oncogenes during NGS. What do you do?
WELCH: If that’s the case, then we’ll probably suggest chemotherapy plus IO [immunotherapy]. You can go about that in a couple of different ways. We still most commonly are using platinum chemotherapy, pemetrexed, and pembrolizumab [Keytruda], but there are other choices, too. You could have a conversation about PD-1 inhibitors alone. I just wouldn’t feel great given all the things we’ve talked about that make this patient unique; you just wonder if you’re missing something. Technically, you could give PD-1 inhibitor monotherapy. I think, more and more, we’re doing chemotherapy plus IO in these cases.
RIELY: If a patient has an EGFR mutation, which treatments do you pick for second- or third-line therapies?
WELCH: We always consider the logistical and travel burdens to the patient: what kind of support they have, if they are still trying to work, and who can bring them to clinic. At the beginning of any therapy, we want to keep a close eye on them. If they start an oral targeted therapy, we want to see them back in the clinic within a couple of weeks to see them in person, check their labs, and talk to them about how they’re doing.
For patients who are good historians and good advocates, and you can tell that they understand what you’re recommending, we’ll do telemedicine with them and perhaps do local labs to try to ease the travel burden. When we’re selecting chemotherapy regimens, some are obviously more demanding and involved than others. Once you get through your first 4 cycles of platinum IO, you can transition to the every-6-week pembrolizumab [regimen]. Things like that that help the patient live as normal a life as possible, where they are not always stuck in your infusion room.