Roundtable Discussion: Experts Debate Types of Testing and Therapies in NSCLC

Case-Based Roundtable Meetings Spotlight, Case-Based Roundtable Meetings Spotlight: December 2, 2021,
Pages: 3

Joel Neal, MD, PhD and Karen Kelly, MD recently teamed up to moderate a roundtable discussion around molecular testing and choice of treatment for a 59-year-old patient with non–small cell lung cancer.

NEAL: The patient is in your clinic—physically, not on video—[and] you have the opportunity to order blood biopsy if you choose. You also have access to tissue that the pathologist gave you [in] the report yesterday, so [nothing] has been skipped in terms of sending anything out for molecular testing.

What are our opinions in terms of what’s the fastest yet most comprehensive molecular testing here?

SHIHABI: The fastest way would be liquid biopsy, but that may not be the most accurate way, perhaps, if it doesn’t show any molecular aberrations—I would still need to test the tissue. Given the fact that there is no immediate need for treatment, perhaps he’s not very symptomatic, I would send the tissue for NGS [next-generation sequencing].

NEAL: What do you expect for a turnaround time in your experience with full DNA NGS vs liquid biopsies?

SHIHABI: One week for the liquid and 2 weeks for the tissue.

NEAL: [Does anyone] find that their experience is faster or slower than that in their practice? I think many of the liquid biopsies have 7- or 10-day options, depending on how comprehensive you want to be. The 10-day option can be a little bit faster, too. I find that one of the limiting factors of a tissue NGS is getting the tissue out of the pathology department, physically out to the company for testing. That 1-week turnaround for NGS is possible, but it’s the sample, getting it out of pathology, that can take a while.

KELLY: I would say that typically we do the liquid [biopsy]. We sometimes send both. We do the liquid first and send the tissue as well. I think that ultimately it depends [on whether] the patient is symptomatic. If the patient truly isn’t symptomatic, then I will tend to just send the tissue. But if I have a patient [who’s] really symptomatic, I’m going to do both because I want the answer as quickly as possible. I don’t know if anybody else sends both.

ZHOU: In this case, why do you need a PET scan? Also, if I have enough tissue, I always go with tissue. I don’t know about a liquid biopsy, but maybe I am behind. I [think] the liquid biopsy is not that reliable.

NEAL: I do believe [in] tissue more. If there is an adequate biopsy, which it sounds as though this patient has, and if I have control over getting the biopsy, I can often get EGFR results within a week. I’ll do a rapid EGFR, ALK, or ROS1 [test] and get those results within a week, especially for never-smokers, where the EGFR pretest probability could be high.

I prefer still doing PET/CT [scans]. I’ve been burned a couple of times just settling for a CT of the chest, abdomen, and pelvis, and I’ve had patients with bone metastases that are occult on CT but then cause pain or fractures later. So at least 1 survey of knowing where all the disease is, I prefer [that] in the beginning. But for follow-up, I often do chest CTs, if it’s relatively easy, to follow a disease.

KELLY: You’re in a place where you have these rapid tests going on for the tissue that maybe the rest of us, including myself at UC Davis, don’t have the ability to do— those rapid tests, the EGFR, ALK, and ROS—so we have to send that tissue out. I always tell patients they are going to be waiting a minimum of 3 weeks if a patient truly is asymptomatic with not a lot of disease. The patients are willing to wait that amount of time because, I agree, tissue is ideal.

But I do have patients who are just too anxious and/ or symptomatic [so] that I just want to do both. If I get lucky on the liquid [biopsy], let’s say I find an EGFR, that’s great, but it doesn’t preclude me from wanting to know family mutations, or comutations. Because if I find a TP53 [mutation] or if I find something else that wasn’t picked up in that blood test, I use the entire NGS to [determine] how I think that patient is going to do. I still think it’s valuable when I send both to get that additional information from the tissue.

In terms of the PET scan, I’m with you, typically, but I can’t say that Dr. Zhou is wrong. I can’t say that he’s wrong because it’s not going to change anything that you do. You used that word occult, which I think is what you would be picking up from a PET scan in the bone. So I can’t say that he’s wrong in not doing the PET scan.

I think that some of the problems that we face also are that you [order] staging PET and then you can’t get any more PET scans [even though] the patient wants [one]. The insurance company isn’t paying for any more PET scans, and you don’t truly have an apples-to-apples comparison. I don’t know whether you all have been running into this, but we do all the time. It’s frustrating, sometimes, comparing PET scan with CT scan, because of the quality.

I can’t say that he’s wrong, but I do agree that I typically get that 1 staging PET scan to look at the bones. But I don’t think it’s 100% necessary.

NEAL: During COVID-19, we’ve been doing most of our new patient visits on video, and it’s sometimes harder and slower to invite a patient to get the liquid biopsy.

You can do mobile phlebotomy and all that, but…my NGS at Stanford in-house takes about 3 weeks. It’s not quite as fast as some of [those we send out]....I’ll totally underscore reassurance, especially for those patients who have a high probability of having never-smoking targetable mutations. We’re now up to 7, 8, 9 [targeted therapies], all of which you would use in the first-line setting but [there are] many things that you don’t want to miss so that you open those treatment opportunities later. So for the asymptomatic patient, [we offer] reassurance.

KELLY: Many [clinicians] do reflex testing, [but] we don’t do that. You come in and they’re anxious and they want to start treatment today, and then you say, “Oh, but it’s precision medicine, we want you to wait, and you’re asymptomatic, your burden is not very high, it’s not going to typically grow in a month or so.”

Then you say, “But I also want to warn you that you may have insufficient tissue to send and that you might need a repeat.” I am very upfront about all this, that it can be a very frustrating process because of these things. I had a patient, she had 2 biopsies with [more than] 10 passes, insufficient tissue for molecular analysis, and she didn’t have anything in the blood.

I say, “I just want to be transparent about this, [it] can be a little challenging.” But I always also say, “If something changes, you get symptomatic, plan B is we can give you chemotherapy. No immunotherapy [IO], but we can always start chemotherapy.” It gives them a little reassurance.

NEAL: The broad NGS panel, if you get it in 2 or 3 weeks and the patient is asymptomatic and can wait, it is the most comprehensive. We’re debating now whether you also need an RNA plus a DNA NGS panel. Some of these [are] rare fusions, so RET, ROS1, NTRK, even the MET exon 14 mutations are probably picked up a little bit more reliably on RNA assays than they are DNA assays, but RNA is finicky and has even more stringent tissue requirements than regular NGS.

Any commercial NGS is close to the gold standard in terms of ruling these in or ruling these out. Once you find a driver, you’re unlikely to find more than 1. Sometimes there are overlaps, but unlikely. So until you have the NGS, ignore the PD-L1.

NEAL: So [the diagnosis is] adenocarcinoma, light smoking history, and the PD-L1 was 1% to 10%. I think this is a good candidate to consider IO, especially after all the negative molecular drivers. Pemetrexed [Alimta] works well on adenocarcinoma, pembrolizumab [Keytruda] is a fairly well-tolerated drug, and [there also is] a carboplatin backbone.1 We used to debate carboplatin vs cisplatin, but that debate is gone in the first-line metastatic lung cancer setting now. I think carboplatin plus pemetrexed plus pembrolizumab [is] well tolerated and well established.

Carboplatin plus nab-paclitaxel [Abraxane] plus atezolizumab [Tecentriq] is an FDA-approved regimen; one that I think of more as the squamous regimen and with either atezolizumab or pembrolizumab.1 For that regimen, I’m thinking more squamous or a patient who has poor creatinine clearance. So I’ll go down to a creatinine clearance of 40 mL/min. How low do you go for pemetrexed? The package insert says 45 mL/min, but I get nervous when it’s less than 40 mL/min, and I just say no, we’re not going to do this, let’s think about something else for the backbone.

KELLY: I go down to 35 mL/min, and our pharmacists have shown some data on this; 35 mL/min is my limit.

BEHL: Is there any reason you did not include the CheckMate 9LA regimen [NCT03215706] as a choice; nivolumab [Opdivo] plus ipilimumab [Yervoy] and 2 cycles of chemotherapy?2

NEAL: I think that’s perfectly worth including as a possibility.

[As for the rest of the options in the poll], carboplatin plus paclitaxel plus bevacizumab [Avastin] plus atezolizumab, the IMpower150 regimen [NCT02366143], has the liability of paclitaxel—the hair loss, the neuropathy.2 Nivolumab plus ipilimumab is the new chemotherapy-free regimen. And then CheckMate 9LA3 is the nivolumab/ipilimumab plus 2 cycles of chemotherapy at the beginning, the induction chemotherapy plus the maintenance nivolumab/ipilimumab regimen.

I’d argue that the single-agent checkpoint inhibitor option is probably the only one that I wouldn’t favor in this setting, unless this patient had poor performance status because of age, not tumor, or had some other contraindication to chemotherapy that was significant. I think pembrolizumab is approved in the 1% [PD-L1] and higher expressors, so technically it’s approved in this indication.4 But we have seen a benefit from chemotherapy in addition to the checkpoint inhibitor. I wouldn’t favor that unless the patient said, “I’m not getting chemotherapy.” But then nivolumab/ipilimumab is compelling, too, because that’s [PD-L1] 1% and higher.

KELLY: I agree; I do not think that a single-agent immune checkpoint inhibitor would be a good treatment option, particularly for this patient....If they’re a little more frail, I don’t know that I would use it then either. It just would depend on if somebody truly didn’t want chemotherapy. I think that’s my only indication for using something single agent.

NEAL: I think all of these are reasonable otherwise, but the bottom line is chemotherapy plus IO clearly leads to longer progression-free survival. We can’t always eyeball somebody and say, “We’re going to guarantee that you can get second-line treatment, so we should give A followed by B.” These are A plus B regimens. I think we sometimes have patients who have lost the opportunity for second-line therapy, and that’s the reason to give it all up front. On the other hand, you can back off from one or the other if there’s a lack of tolerability once you get going.

NEAL: I think there are certain circumstances [where you could use a single-agent immune checkpoint inhibitor]. I just published an article [making the argument that] patients who are older, want to avoid chemotherapy, and maybe have contraindications to chemotherapy are patients [with whom] you could try the IO and see [whether] it works. But I would stress repeat imaging soon, and don’t do it for the patients who if their tumor doubles [in size] they’re going to be really symptomatic.

KELLY: Have any of you used it in this setting?

BEHL: Yes, I have. It went better than I expected. It was an 89-year-old patient, he was 89 physiologically, and he did pretty well—he lived for approximately a year after treatment.

KELLY: Was his PD-L1 closer to 49%, or was it closer to 1%?

BEHL: No, it was [approximately] 20%.

RASILA: I had a patient whose PD-L1 was 30%, he was in his early 90s, but his performance status was not too bad. He wanted to try treatment but didn’t want chemotherapy. He lived for approximately a year and a half after treatment.

NEAL: I have, for what it’s worth, another patient in his 90s who received single-agent IO. His PD-L1 was [approximately] 20%. I wasn’t going to give him chemotherapy. And he celebrated many more birthdays afterward. We stopped the drug completely [and the] tumor is not growing. It’s amazing when that happens.

KANKIPATI: How about using a doublet when you’re in the middle; would you use [a pembrolizumab combination]vs omitting the carboplatin? Where does that stand in this patient population where you don’t want to give a triplet?

KELLY: This is a question that comes up; I think you’re talking more about the performance status 2 patient population. We have tried to do a study in SWOG—I know ECOG also tried to do a study in this patient population—asking these questions. There is preclinical evidence to say that pemetrexed plus IO, at least in preclinical models, does show some additivity. We also know that platinum [therapies] don’t show this immunogenic cell depth. So taking out the platinum isn’t a bad idea, but we could not come to a consensus on a study design in this patient population because, if you wanted to do the right study and compare it with chemotherapy, nobody wanted [to do the] chemotherapy-alone arm.

The only way to provide what I would say is evidence-based medicine is [by] comparing it with the standard of care. We spent months and never came up with a design that was going to be palatable for SWOG, which I think is really unfortunate. I know ECOG tried as well and they also were not able to come up with something that was approved by the National Cancer Institute.

I think this is a good question, but we can’t seem to prove it. I think it’s unfortunate because everybody has already adopted some IO, whether it’s a modification, it’s going to be pembrolizumab plus IO, or it’s going to be pembrolizumab plus carboplatin with modified doses of chemotherapy plus IO. And then, rarely, [it is] single-agent IO. That’s a good question that we’re never going to have an answer to unless we do some real-world data registry to find out.

NEAL: I think the real-world answer is that these regimens are idealized for a clinical trial population that was carefully selected with many eligibility criteria, and every patient we meet in the clinic who is not eligible for a clinical trial, which we’d now still prefer, has some issues. So there’s always some reason, maybe we can’t give the full dose of carboplatin, maybe we can’t give 4 cycles total, maybe the creatinine clearance isn’t great, maybe we use this and modify the idealized regimen. That’s what we do as medical oncologists. In reality, how do we tailor it for our patients in front of us?

I think it’s reasonable to dial in therapies. I sometimes do that, especially when I’m using bevacizumab with chemotherapy. I’ll say, “Let’s give a couple of cycles, see how it goes, maybe dial in the bevacizumab.” You could think about dialing in the carboplatin or just doing a test cycle of pembrolizumab, [or] pembrolizumab alone. I think all of these things are reasonable.

At least from what I can tell with Medicare, for the Medicare population, they’ll ask, “Is the drug on the list or is the drug not on the list?” They’re not looking at details; did you give it all together in the right line of therapy and what was the PD-L1? So some insurers are a little bit pickier, but we’ve managed to escape rigid pathways for now for most of our patients.

I think the real-world answer is whatever combination your patients can tolerate....[From the] studies before we had IO, the answer was if you can give them platinum, give them platinum. Trying even less platinum is better than giving them single agents if you can get away with it. So try the platinum if you can.

NEAL: Does anyone support the idea of treatment beyond progression? Does anyone ever add radiation or something like that, as we do with targeted therapies, and continue the IO?

RASILA: Yes, definitely if they’re on IO and there is progression. But if it’s just 1 or 2 spots, then yes, I have done some stereotactic body radiation therapy and continue with the IO.

NEAL: We have a clinical trial using that strategy, of patients who progressed a little or had stable disease overall, radiating it. By definition, it improves progression-free survival because you radiated the target lesion. But the question is: Does it improve overall survival? There are some limited data about that. The oligometastatic trials’ data [showed] radiation may improve overall survival even after IO. So there is an ongoing cooperative group study [investigating] that in a prospective way with radiotherapy and immune therapy.

Changing treatment, I think, always comes down to a discussion with the patient, seeing what the risk benefits are, tolerability, what are your additional future options. Assuming no targets were uncovered, and our other option is second-line chemotherapy-based regimens, I think we may sometimes treat beyond progression and say, “Let’s give it another few months,” especially with those [patients] who responded well.

KELLY: I think one thing about this patient is [that] he is symptomatic....I would not say let’s just go with the IO. I would do the radiation [on] the 1 lesion and the adrenal. [If] you could radiate both of those, I agree [on] radiation. In [a patient] who is symptomatic, [I wouldn’t] say I’m going to treat beyond progression. I would say let’s add radiation, or I would switch if that wasn’t going to be the case.

If a patient is asymptomatic, and I do agree if it’s a tiny bit, maybe then I would say let’s wait and watch, but this is a symptomatic patient. I don’t know what others would do here.

CHEN: If you change chemotherapy, would you keep the checkpoint inhibitor?

NEAL: That’s always something I’ve wondered, but there are no approvals for the second-line drugs that we use most commonly, docetaxel or gemcitabine, with the IO.... Once the IO stopped working, I’m unlikely to induce a response by just changing the chemotherapy backbone. So I’d usually say it’s time to just cleanly move on.

NEAL: We have a number of second-line clinical trials open at our institution. Dr Kelly probably does, too, for this patient population. Really, what do we do after progression on pemetrexed, carboplatin, and IO? What do we do to try to induce that? Some of them are VEGF or VEGFR based, which I think is one of the areas that’s exciting in development now. I wouldn’t continue the IO, unless maybe you consider that nivolumab/ipilimumab switch.

KELLY: I agree. We are anxiously awaiting the results from the LUNG-MAP study [NCT03971474] where we randomized patients to standard of care, whether that was docetaxel plus or minus ramucirumab [Cyramza] or gemcitabine to pembrolizumab plus ramucirumab in patients who had previously failed PD-L1. I think there will be exciting data; [it will be exciting] to see what happens.


1. NCCN. Clinical Practice Guidelines in Oncology. Non-small cell lung cancer, version 7.2021. Accessed November 15, 2021.

2. Socinski MA, Jotte RM, Cappuzzo F, et al; IMpower150 Study Group. Atezolizumab for first-line treatment of metastatic nonsquamous NSCLC. N Engl J Med. 2018;378(24):2288-2301. doi:10.1056/NEJMoa1716948

3. Paz-Ares L, Ciuleanu TE, Cobo M, et al. First-line nivolumab plus ipilimumab combined with two cycles of chemotherapy in patients with non-small-cell lung cancer (CheckMate 9LA): an international, randomised, open-label, phase 3 trial Lancet Oncol. 2021;22(2):198-211. doi:10.1016/S1470-2045(20)30641-0

4. Keytruda. Prescribing information. Merck & Co Inc; 2021. Accessed November 15, 2021.