In a Targeted Oncology case-based peer perspectives live discussion, David Gerber, MD, addresses the role of biomarkers in guiding treatment decisions for patients with non–small cell lung cancer based on a patient scenario.
In a Targeted OncologyTM case-based peer perspectives live discussion, David Gerber, MD, a professor of internal medicine and associate director of clinical research at The University of Texas Southwestern Medical Center in Dallas, addressed the role of biomarkers in guiding treatment decisions for patients with non–small cell lung cancer based on a patient scenario.
TARGETED ONCOLOGY: Why do you think they did a biopsy on the patient’s metastases instead of the primary site?
Gerber: You can tap the pleural effusion to improve symptoms. The challenge of pleural fluid is that it might be harder to do molecular analysis because there may be fewer cells. You may have a bony metastatic site where it’s also hard to do molecular analysis because you have to decalcify the tissue and it’s not as viable.
It may be that biopsying the liver is safer than biopsying a lung lesion. You still have issues of bleeding and pain and infection, but usually you don’t have a risk of pneumothorax. They biopsied the liver lesion and it showed adenocarcinoma, so now we have pathologically proven stage IV adenocarcinoma in a former smoker.
TARGETED ONCOLOGY: What kind of molecular testing would you order for this patient?
Gerber: [I would order] next-generation sequencing [NGS] plus PD-L1 testing.
You don’t get PD-L1 status from a blood test. We get [plasma for molecular biomarkers testing]; we’re looking at DNA. We’re looking at the genes in the cancer, so you can detect a mutation. PD-L1 is a protein on the cancer cell surface.
If you have a patient with stage IV cancer, maybe 0.5% or 0.3% of the DNA in the blood is from the cancer. We [have tests that can] pull that out, so that’s another option.
These tests don’t necessarily [look for] circulating tumor cells; this is just the [circulating tumor] DNA. Circulating tumor cells have a diff erent assay. We’re not sure what to do with that information right now in a lot of cancers. There are some cancers, like breast, prostate, and maybe colon cancer, where it’s prognostic [for] response to therapy.
TARGETED ONCOLOGY: Which mutations do you usually test for in these patients?
Gerber: If we’re sending [tissue for] NGS, [the number of genes] depends on the panel. It might be 22,000 genes through [the] Caris Life Sciences [assay]. At The University of Texas Southwestern, we have a panel that’s CLIA [Clinical Laboratory Improvement Amendments] certified, and it’s 1415 genes. It pairs germline, or the patient’s DNA, with the tumor’s DNA.
In lung cancer, the National Comprehensive Cancer Network [NCCN] recommends, at a minimum, testing for EGFR, BRAF, ALK, and ROS1 mutations.¹ Those are 4 subsets of lung cancer. If you think about nonsquamous, non–small cell lung cancer [NSCLC], what is [the rate of] EGFR mutations? Somewhere around 10% to 15%. BRAF mutations are around 3% to 5%, ALK mutations are around 3% to 5%, and ROS1 mutations are around 1% of cases. [As you go down the list] they start to get rare; but if you don’t look for it, you don’t find it.
If you have a nonsquamous case, it can’t hurt to look for an NTRK mutation also because there are TRK inhibitors that are FDA approved regardless of the type of cancer.2 It doesn’t matter anatomically what it is. There are MET mutations, there are RET rearrangements, and there are KRAS mutations.
Believe it or not, there are HER2 mutations in lung cancers. I’m not talking about HER2-amplified or HER2-expressing tumors, like you would look for in breast cancer; it’s a different biomarker. [We would] look for a mutation in HER2 exon 20 that predicts response to HER2 inhibitors.
TARGETED ONCOLOGY: This patient is a former smoker; does smoking history or other factors affect the patient’s mutation status?
Gerber: If your patient is a heavy smoker, EGFR mutations will probably be in less than 5% of those cases. If your patient is a never smoker and you’re looking for EGFR, depending on the histology and the patient, [it will be about] 60%.
I do have patients who are African American men who smoked a lot and have EGFR mutations. But there are times where you may go to greater efforts to genotype. In general, the [treatable] mutations are going to be in the patients who have smoked less.
There are a couple of exceptions. I’m not sure that BRAF mutations are necessarily more common in never smokers or former smokers. KRAS mutations have long been considered [untreatable] in oncology. It is the second most common mutation in human cancer. The most common mutation in human cancer is TP53, the guardian of the genome.
TARGETED ONCOLOGY: Why do you never hear about inhibitors of TP53 mutations?
Gerber: They don’t exist. All these other mutations are proto-oncogenes, so when they are mutated, they function more. TP53 is a tumor suppressor; its job is to prevent cancer. When it’s mutated, it stops doing its job or doesn’t do it as much. You’re not trying to inhibit mutant TP53; you’re trying to increase its function. It’s a lot harder to increase the function of a molecule in medicine than it is to inhibit it or slow it down.
TARGETED ONCOLOGY: What other mutations or issues should physicians keep in mind when testing?
Gerber: We’ve been trying to target MET mutations in lung cancer for a long time, and it came down to which MET cases matter. We were looking at MET amplification, such as with FISH [fluorescence in situ hybridization]; we were counting the number of genes or chromosomes. We were looking at MET expression, so that’s the protein [detected by] immunohistochemistry, and those trials were negative. It’s the MET exon 14 mutation [that should be targeted, and this mutation] tends to respond well to MET inhibitors.
When you do a blood test looking for the DNA of cancer, there are times that the test will be negative but the tumor is positive [for a mutation]. Maybe the tumor didn’t shed enough DNA [for the test] to pick it up. There are also cases where there is negative testing in the tissue, and the patient is positive for a mutation with a blood test. Is that a false-positive blood test? It’s not; it could have to do with tumor heterogeneity. When you stick a needle in the tumor, you’re [sequencing] just 1 site, whether it’s a lymph node or a primary or metastatic tumor. That area might not have that particular mutation that you could detect.
TARGETED ONCOLOGY: Have you had cases where the first and/or second biopsy didn’t have enough tissue and you still needed to get a third biopsy?
Gerber: I have. This is going back to 2009, when we first learned about ALK-rearranged or ALK fusion-positive lung cancer. I had a patient who was a never smoker with lung cancer, and she had been on an EGFR inhibitor, hadn’t had any testing, and hadn’t responded at all [to therapy]. She received some chemotherapy; she was on hospice in a wheelchair, getting supplemental oxygen, and she came for a second opinion.
I told her I had just heard about a new subtype of lung cancer called ALK-positive lung cancer, and I thought we should look to see if she had it. Even though just 3% of lung adenocarcinoma is ALK positive, for a never smoker with adenocarcinoma, the chance of having an ALK mutation is about 50%. I did a biopsy on her liver and there was insufficient tissue. I asked if we could do 1 more, and it came back ALK positive. Our pathologist had never heard of the test, so I had to get them to order it, and it was positive. She went on to live another 5 years. These are hard decisions.
TARGETED ONCOLOGY: How does PD-L1 factor in when deciding on treatment, and is it considered a biomarker?
Gerber: If you test for PD-L1, we consider negative as 1% or less expression, ntermediate as 1% to 49%, and high or positive as 50% or greater. PD-L1 is the biomarker that helps us determine the efficacy of immune checkpoint inhibitors that target PD-L1 or PD-1. It is not a perfect biomarker.
TARGETED ONCOLOGY: Is it necessary to test for EGFR mutations if you suspect a patient has them?
Gerber: We were spoiled in thoracic oncology with biomarkers like EGFR. If I have an EGFR mutation and I’m giving an EGFR inhibitor, my response rate is over 90% with a third-generation EGFR inhibitor. If I don’t have an EGFR mutation and I give an EGFR inhibitor, my response rate is below 1%. I can only give an EGFR inhibitor if the biomarker is positive. With greater than 90% versus less than 1%, it’s a clear decision. You know what to do.
TARGETED ONCOLOGY: What makes mutation biomarkers positive in this setting?
Gerber: For the most part, when we’re looking for mutations, we tend to take any evidence of a mutation as the tumor being positive [for that aberration]. It tends not to come back measured by cells because we’re looking at the DNA and not a cell surface marker. At the cellular level, that’s difficult to interpret. With ALK, when we’re doing FISH testing, there is a cutoff saying it should be that 15% of cells or more have that rearrangement. What do you do if it’s 14%? Is that different from 16%? You can do another test looking for the protein. If that’s there, that’s not to be functionally meaningful—especially when we’re testing the blood—because the denominator of DNA is not cancer DNA, it’s all the DNA in the blood. The total cancer DNA may be 0.5%. We don’t know that, but when you get 0.2% coming back with a mutation, that’s a positive result, and you’re going to act on it.
TARGETED ONCOLOGY: How would you handle this patient, considering the results by IHC?
Gerber: The relationship between mutations of a cancer and the likelihood of immunotherapy working is antigenicity. The more mutations the cancer has, the more foreign it appears to
the patient’s immune system and the more likely immunotherapy will work.
[This patient has] 14 mutations per megabase; is that a high level or a low level? It depends on the company and the clinical trial. Bristol Myers Squibb uses 10 mutations per megabase [as a threshold for high expression], and AstraZeneca uses 16 mutations per megabase in their durvalumab [Imfinzi] with or without tremelimumab studies, so it depends. It’s something that’s hard to agree upon.
TARGETED ONCOLOGY: What are the treatment choices, and which factors influence your recommendations for this patient?
Gerber: You don’t want to give checkpoint inhibitor monotherapy. You want to give chemotherapy alone [or in a combination]. We have platinum [therapy] plus pemetrexed [Alimta] and pembrolizumab [Keytruda]. We also have platinum- and taxane-based therapy plus atezolizumab [Tecentriq] with or without bevacizumab [Avastin].
There was a time when I could look at the NCCN guidelines and they told me what to do for patients with lung cancer. That is no longer the case, and I’m sad to say they’ve lost some of their usefulness.
In this case, the patient has NSCLC and PD-L1 expression status of 2%. About 70% of cases fall into that category. About 30% are going to be PD-L1 negative, so I’m talking about the 70% [that are positive].
[The NCCN recommends] platinum chemotherapy, pemetrexed [Alimta], and the immune checkpoint inhibitor pembrolizumab as category 1. [You may also use] the platinum/taxane combination with the antiangiogenic [agent] bevacizumab plus anti–PD-L1 atezolizumab. We also have platinum nanoparticle albumin-bound paclitaxel [Abraxane] as an option. There’s no FDA indication for this, but combination immunotherapy [is also suggested].1
TARGETED ONCOLOGY: Do you feel that there’s a difference between the approved drugs for this setting (pembrolizumab, nivolumab [Opdivo], atezolizumab, and durvalumab) in toxicity or efficacy?
Gerber: Some are PD-1 versus PD-L1, some are humanized antibody versus fully human antibody, and some are immunoglobin G1 [IgG1] versus IgG4. I’m not sure there is a meaningful difference in efficacy or toxicity between PD-1 and PD-L1. IgG1 versus IgG4 would tell you how [the] complement interacts with the antibody; I don’t think that’s relevant to the mechanism. And a fully humanized versus a human antibody would tell you the likelihood of an infusion reaction.
With a humanized antibody—that’s atezolizumab or pembrolizumab—the chance of an infusion reaction is probably 1% to 2%. With a fully human antibody like durvalumab or nivolumab, the risk of an infusion reaction is less than 1%. The single time in my practice I have changed from one checkpoint inhibitor to another was when a patient had pembrolizumab and they had an infusion reaction; not autoimmune toxicity, not pneumonitis, not nephritis, not colitis—they had an allergic infusion reaction. I said, “I’m going to switch to a fully human antibody.”
TARGETED ONCOLOGY: What are the design and efficacy of relevant trials in this setting?
Gerber: The KEYNOTE-042 [NCT02220894] study [involves] patients that had PD-L1 status of less than 1%, and they’re receiving immunotherapy [of pembrolizumab] or chemotherapy.3
Looking at the progression-free survival [PFS] of the group whose status is at least 50%, there was the greatest benefit in PFS, with a hazard ratio of 0.81 [95% CI, 0.67-0.99; P = .017]. For patients with less than 20% PD-L1 status, there was a hazard ratio of 0.94 [95% CI, 0.80-1.11]; it’s not significant. When you look at patients with PD-L1 expression greater than 1%, that’s also not significant [HR, 1.07; 95% CI, 0.94-1.21]. But it’s never worse than chemotherapy.
What’s interesting to me is that the OS is better with immunotherapy than it is with chemotherapy regardless of the PD-L1 cutoff.
There was an exploratory analysis, and [it showed that] no difference in OS whether you give immunotherapy first or chemotherapy first. We don’t know about chemotherapy plus immunotherapy; that has not been formally asked.
TARGETED ONCOLOGY: Please discuss the important aspects of the IMpower150 study.
Gerber: IMpower150 [NCT02366143] was a 3-arm study; patients received chemotherapy plus bevacizumab, an antiangiogenic drug; chemotherapy plus the immune checkpoint inhibitor atezolizumab; or both immunotherapy and antiangiogenic therapy.4
I want to point out that this trial suggests that immunotherapy works better if the patient has liver metastases. I don’t know why that was looked at [as a stratification factor]. There is another trial that looks at liver metastases, and those patients did worse. I don’t know why this question is being asked. However, if you ask enough questions, you’ll eventually find something.
The other trials that I [mentioned], which added chemotherapy or immunotherapy to chemotherapy, excluded patients with EGFR mutations or who were ALK positive. Those patients were [about 13% of the entire cohort] were included in this trial, and they benefited. We didn’t know whether they would benefit from the addition of immunotherapy to platinum [therapy and] pemetrexed; that question hadn’t been asked in that study.
TARGETED ONCOLOGY: How does the CheckMate 227 [NCT02477826] trial compare to these other therapies?
Gerber: The investigators enrolled about 1700 patients, so it was a big trial. If a patient was negative for PD-L1 expression, there was a different 3-way randomization than if they were PD-L1 positive. For example, in the PD-L1–positive cases, 1 of the arms is single-agent immunotherapy of nivolumab. That was not offered to the patients with negative PD-L1 expression. The second arm was chemotherapy alone, and the third was combination immunotherapy, with nivolumab and ipilimumab [Yervoy]. In the PD-L1–negative comparison, there was combination immunotherapy, chemotherapy alone, and then chemotherapy plus immunotherapy. The main comparisons were combination immunotherapy versus chemotherapy.5
TARGETED ONCOLOGY: What was the efficacy of this trial?
Gerber: The OS in all patients showed an improvement with combination nivolumab plus ipilimumab versus chemotherapy alone [median OS with nivolumab and ipilimumab, 17.1 months versus 14.9 months with chemotherapy; P = .007].6 Looking at the subset of patients who had at least 1% PD-L1 expression, there was a signifi cant improvement in OS. I don’t [know] the hazard ratio, which is always more meaningful to me than the P value. You can drive your P value and get it lower just by having more patients on your trial, regardless of the diff erence in efficacy.
In the PD-L1–negative cases, combination immune checkpoint inhibitor is better than chemotherapy. The only biomarker we’re looking at is PD-L1; we’re not looking at tumor mutation burden.
TARGETED ONCOLOGY: What else did the CheckMate 227 trial show about biomarkers?
Gerber: They looked at both PD-L1 and tumor mutation burden [TMB]. In general, we think of [tumors with] high PD-L1 and high mutation burden as more likely to respond to immunotherapy. How much do these groups overlap? If you look at these groups of patients, are they synonymous with each other? Not at all. The overlap is only about 60%, believe it or not.5
For PD-L1–negative cases that are TMB low versus high, it’s almost the same number of patients treated—111 versus 86, respectively—and that’s amazing. For PD-L1–positive tumors, more are TMB low than are high. I think that when it comes to single-agent immunotherapy, PD-L1 is a more meaningful biomarker. But when you add in the CTLA-4 [antibody] and you think about what biomarkers perform with those combinations, I would say TMB performs better than PD-L1.
If you add them together, there was benefit across all these [subgroups]. You’re seeing slightly greater benefit with the higher level of PD-L1. But is this the reason, when you look across all [recommended therapies], that the NCCN [recommends] this as a reasonable option for our patients.
TARGETED ONCOLOGY: Would you change treatment in this patient? If not, why not?
Gerber: I’m going to assume that it’s a clinically meaningful change, because [if] it’s a 2-mm diff erence, I may say that’s not [clinically meaningful and doesn’t warrant a change].
Another reason to not change treatment is [the possibility of] pseudoprogression. There are data out there that say that if the patient feels better but [the progression is] worse, then it’s pseudoprogression; but if they don’t feel better and [the progression is] worse, then it’s not pseudoprogression. I don’t make my decisions based on that. At The University of Texas Southwestern, we have published cases of patients who clinically deteriorated and went on hospice, and then 18 months later came back for a visit and their tumor had shrunk, so you can’t rely on clinical status [alone].
The other interesting thing about immunotherapy is this concept of hyperprogression. That’s where not only do you have disease progression but also…the cancer growth rate has at least doubled after starting immunotherapy. I’ve had tiny pleural base disease after starting immunotherapy; the entire hemithorax was gone, and the patient couldn’t move his arm.
People do studies trying to quantitate this and determine how often it happens; it’s thought to happen in about 5% of patients who get chemotherapy—they meet that definition of a doubling of growth—and up to 15% of patients with immunotherapy. I had cases in my practice, before the combination era with chemotherapy, where I was giving them immunotherapy and they progressed; I switched them to chemotherapy, and then they had [a great] chemotherapy response. Is that response to chemotherapy, is that resolution of pseudoprogression, or is it both?
TARGETED ONCOLOGY: What are the options for second-line therapy, and what data support these approaches?
Gerber: The ramucirumab [Cyramza] clinical trial data came in an era before immunotherapy [was available] for patients with lung cancer. The idea of a positive clinical trial was different. You can enroll 1300 patients to a clinical trial and [have barely any separation of the] Kaplan-Meier survival curves, but they had so many patients on the trial where the P value was significant.
The REVEL [NCT01168973] study investigated ramucirumab and docetaxel [versus placebo and docetaxel]. Ramucirumab is an antiangiogenic, antivascular EGFR antibody. This was in patients with NSCLC [of all histologies].7
The investigators looked at whether the patients progressed within the first month of therapy. Within the first month, the OS hazard ratio was 0.40 [95% CI, 0.22-0.73]. That tells me that [if this patient] progressed within the first 4 weeks of getting firstline therapy, [he would respond to ramucirumab].7
1. National Comprehensive Cancer Network Clinical Practice Guidelines in Oncology. Non-Small Cell lung Cancer (version 3.2020). Updated February 11, 2020. Accessed April 15, 2020. https://www.nccn.org/professionals/physician_gls/pdf/nscl.pdf
2. FDA approves entrectinib for NTRK solid tumors and ROS-1 NSCLC. FDA website. FDA. Updated August 16, 2019. Accessed April 15, 2020. https://www.fda.gov/drugs/resources-information-approved-drugs/fda-approves-entrectinib-ntrk-solid-tumors-and-ros-1-nsclc
3. Mok TSK, Wu YL, Kudaba I, et al. Pembrolizumab versus chemotherapy for previously untreated, PD-L1-expressing, locally advanced or metastatic non-small-cell lung cancer (KEYNOTE-042): a randomised, open-label, controlled, phase 3 trial. Lancet. 2019;393(10183):1819-1830. doi: 0.1016/S0140-6736(18)32409-7
4. Socinski MA, Jotte RM, Cappuzzo F, et al. Atezolizumab for fi rst-line treatment of metastatic nonsquamous NSCLC. N Engl J Med. 2018;378(24):2288-2301. doi:10.1056/NEJMoa1716948
5. Hellmann MD, Ciuleanu TE, Pluzanski A, et al. Nivolumab plus ipilimumab in lung cancer with a high tumor mutational burden. N Engl J Med. 2018;378(22):2093-2104. doi:10.1056/NEJMoa1801946
6. Hellmann MD, Paz-Ares L, Bernabe Caro R, et al. Nivolumab plus ipilimumab in advanced non-small-cell lung cancer. N Engl J Med. 2019;381(21):2020-2031. doi:10.1056/NEJMoa1910231
7. Garon EB, Ciuleanu TE, Arrieta O, et al. Ramucirumab plus docetaxel versus placebo plus docetaxel for second-line treatment of stage IV non-small-cell lung cancer after disease progression on platinum-based therapy (REVEL): a multicentre, doubleblind, randomised phase 3 trial. Lancet. 2014;384(9944):665-673. doi: 10.1016/S0140-6736(14)60845-X