Gandara Uses Predictive Biomarkers to Determine Best Immunotherapy Options in Patients With Advanced NSCLC

David R. Gandara, MD, discusses the treatment decisions he makes and the data that support his recommendations for treating patients with non–small cell lung cancer

David R. Gandara, MD

David R. Gandara, MD, explained to a group of physicians in a recentTargeted Oncologylive case-based peer perspectives presentation the treatment decisions he makes and the data that support his recommendations for treating patients with non—small cell lung cancer (NSCLC). Gandara, a professor of medicine in the Division of Hematology/Oncology at the University of California, Davis School of Medicine, and the director of the Thoracic Oncology Program at the UC Davis Comprehensive Cancer Center, explained the factors that go into his treatment decision-making when considering the use of immunotherapy based on the case scenarios of 3 different patients with NSCLC.

Case 1

A 59-year-old Caucasian man presented to his physician with symptoms of chest pain, cough, and dyspnea. His ECOG performance status was 1. His past medical history included hypertension and osteoarthritis. He was a former smoker with a 10 pack-years history.

A chest and abdominal CT scan revealed a 9-cm spiculated mass in the left lower lobe, loculated pleural effusion in the left hemithorax, diffuse liver nodules, and right adrenal metastases. PET/CT scan showed fludeoxyglucose (18F-FDG) uptake in the left lung mass, pleura, liver, and right adrenal gland. A brain MRI was negative for metastases.

Next-generation sequencing (NGS) was negative for molecular aberrations inEGFR, ROS1, BRAF, ALK,RET, MET, HER2, andKRAS. Immunohistochemistry (IHC) testing showed PD-L1 expression of <1% and a tumor mutational burden (TMB) of 10 mutations per megabase.

What type of molecular testing would you do in a patient such as this, and what are the treatment options?

Our organization, the International Association for the Study of Lung Cancer [IASLC], has gotten together with the College of American Pathology [CAP] and the Association for Molecular Pathology [AMP] to create guidelines, which we did in 2013 and again [in 2018].1

If you are not into lung cancer, you may not know that there are now 8 treatable oncogenes that are in the guidelines. Six of the 8 oncogenes have targeted therapy agents as the treatment of choice, not immunotherapy [or] immunotherapy with chemotherapy.

The CAP-IASLC-AMP guidelines [outline] what is a mandatory “must test” as well as a “should test” oncogene. I will point out that the American Society of Clinical Oncology, the European Society of Medical Oncology [ESMO], and the National Comprehensive Cancer Network [NCCN] include all these in the “should test” category.KRASmutations were included in these guidelines, but I do not agree with that. That is left over from Memorial Sloan Kettering [recommendations that indicated testing for]KRASfirst; if the patient had aKRASmutation, then you did not have to do anything else. That is wrong. As a matter of fact, we are the very first ones to present that patients can haveKRASmutations andEGFRmutations de novo at presentation. [There was some debate over this when we first presented the findings.] I do not think you need to test forKRASin 2019.

It is recommended that you do one of the following: either perform a broad, comprehensive cancer panel, and this usually means NGS…or do the big 3,EGFR,ALK, andROS1, and if they are negative, then do NGS. The guidelines also recommend PD-L1 testing. NCCN and ESMO as of this year also recommend testing for TMB.

[This] patient is tested with a broad NGS platform, yet the patient is negative for everything. IHC shows the PD-L1 expression is <10% and the tumor mutational burden is 10 mutations per megabase.

What would be your preferred treatment recommendation for this patient with stage IV adenocarcinoma with PD-L1 expression <10% and 10 mutations per megabase?

Options for therapy include pemetrexed [Alimta] plus carboplatin; paclitaxel, carboplatin, and bevacizumab [Avastin]; checkpoint inhibitor monotherapy with pembrolizumab [Keytruda]; pembrolizumab, pemetrexed, and carboplatin; atezolizumab [Tecentriq] plus paclitaxel, carboplatin, and bevacizumab; nivolumab [Opdivo] plus ipilimumab [Yervoy]; or another regimen.

This KEYNOTE-189 trial is pivotal.2This is a trial of either pemetrexed plus platinum chemotherapy, or the combination with pembrolizumab—I call this “pem-pem”—in first-line nonsquamous NSCLC with the possibility for a crossover. As you can see, it is a positive trial. The Kaplan-Meier curves for overall survival [OS] and progression-free survival [PFS] separate early and stay separated, and there was a pretty good proportion of patients who crossed over.

This was a positive study, not only for PFS but for OS. If you look at OS by PD-L1 levels, you can see that the greatest impact, even with the addition of platinum chemotherapy, is still in the group with a PD-L1 level that is ≥50%. However, you can see that it is still quite positive in the patients who are negative for PD-L1 expression. My own impression is that with all the biomarkers that we have for immunotherapy, if you throw in chemotherapy, you dilute out some of the potential for the biomarkers. Chemotherapy is relatively agnostic for the biomarker.

The other thing about this study and the combination of pembrolizumab with chemotherapy is that there was a previous trial—KEYNOTE-024—that tested pembrolizumab alone in patients with high PD-L1 expression, ≥50%.3Realizing the vagaries of comparing trials, if you look at the high expressers in both of these trials, the results are pretty equivalent. This leads me to the opinion that pembrolizumab as a single agent is appropriate therapy for patients regardless of histology, both squamous and nonsquamous, with high PD-L1 levels. That forms the basis for my opinion, and most of my colleagues in lung cancer agree with that.

Going back to pemetrexed and platinum chemotherapy with pembrolizumab versus just chemotherapy, response rates favored those patients with the highest level of PD-L1 expression.

Atezolizumab is another anti—PD-L1 agent. The pivotal study [IMpower150] led to its FDA approval in nonsquamous NSCLC, regardless of PD-L1 level.4It tested the combination of atezolizumab plus bevacizumab, carboplatin, and paclitaxel [ABCP] compared with BCP or atezolizumab plus carboplatin and paclitaxel, and it had multiple treatment arms.

The one that we are focusing on is the 4-drug regimen. This again is a positive study. Everyone always says, “Compared with KEYNOTE-189, this does not look as positive.” That is true, but remember the comparator here is a regimen with bevacizumab, where we already know that there is a survival advantage over chemotherapy alone. To me that is not surprising, and it does not diminish the result.

You may ask, “Why would you use this 4-drug regimen, and who would you use it in when compared with the prior regimen we just looked at?” Well, one of the interesting things is that all the first-line studies excluded patients withEGFRmutations orALKtranslocations because the second-line studies showed that they did not do well. As a matter of fact, the hazard rate showed that there was harm when giving single-agent immunotherapy versus giving docetaxel, so these patients were excluded. However, this study included them, and it was beneficial. This 4-drug regimen was beneficial for patients withEGFRandALKmutations. My impression is that this was due to bevacizumab. We know that bevacizumab has some magic withEGFRnonsquamous NSCLC. So this is now an FDA-approved regimen.

What are the different assays for TMB, and how do they relate to one another?

CheckMate 227 is the only trial to have a co-primary endpoint of TMB.5There are different assays for TMB. As I have said, it is not an FDA-approved predictive biomarker, but it is incorporated in both the NCCN and ESMO guidelines. We can measure total mutational burden and what are called neoantigen loads by tracking the neoantigens by a variety of techniques: a research tool called whole-exome sequencing, or comprehensive genomic profiling; commercial tests, such as FoundationOne CDx; or in blood. There are 2 assays for use in the blood; the one that I developed using the Foundation blood assay and the new one by Guardant Health.

How did we first know about TMB? What are the rationale and cumulative data for using TMB in therapeutic decision making for immunotherapy?

This is one of the negative trials, CheckMate 026.6It tested single-agent nivolumab versus platinum-based chemotherapy in first-line NSCLC. This had the same design as KEYNOTE-024, which was positive. When these results came out back-to-back with KEYNOTE-024, everyone asked whether it was because nivolumab is a different agent. What is going on here? There are some differences in the demographics of the patients, but this was published in theNew England Journal of Medicine,and that journal does not usually publish negative trials.

Carbone and colleagues found that if they analyzed the trial by patients with high TMB expression, it was positive. And in fact, with low to medium TMB expression, [chemotherapy has a better response]. In this trial, TMB was very telling, but it was a research tool.

Then they looked at the FoundationOne CDx and compared it with whole-exome sequencing and showed that there was a good correlation with a commercial test except with the lower levels of TMB.

Furthermore, in the nivolumab arm, patients who had high PD-L1 and TMB expression simultaneously did extraordinarily well. In a totally negative trial, we can figure out who benefited with predictive biomarkers.

This brings us to CheckMate 227, which had a co-primary endpoint of TMB ≥10 mutations per megabase by the FoundationOne CDx assay, which compared nivolumab alone or in combination with ipilimumab with platinum chemotherapy alone.5The problem here is that the investigators and Bristol-Myers Squibb, [the manufacturer of both nivolumab and ipilimumab], used a retrospective analysis of 100 patients to come up with a TMB threshold of 10 mutations per megabase. Foundation Medicine has studied 35,000 human genomes to come up with their cutoff of 16 mutations per megabase, which was not used in this trial. This will become important.

In this trial, those patients who had high TMB had a positive outcome for PFS with nivolumab plus ipilimumab compared with chemotherapy. This was true in both squamous and nonsquamous disease. The interesting thing about the study is that if patients had low PD-L1 expressions, 0%, but high TMB, they benefited from nivolumab plus ipilimumab. There was also an arm of nivolumab plus chemotherapy over chemotherapy alone. If TMB was low and PD-L1 expression was low, there was no benefit. This study has not yet led to an approval by the FDA, although it is in the guidelines, and it is reimbursed for nivolumab and ipilimumab. Part of the reason is that the investigators picked a TMB of 10 mutations per megabase. We know that according to the press release, OS was not quite statistically positive.7This would be positive, I think, with a TMB of 16 mutations per megabase, but that is not how they designed the study.

Case 2

A 62-year-old man presented with persistent right-sided neck pain, along with decreased appetite, lethargy, and a dry cough. His ECOG performance status was 1. His past medical history was remarkable for hypercholesterolemia and arthritis. He was a former smoker but had not allergies or family history of lung cancer.

A chest CT revealed a 4.3-cm right upper lung (RUL) mass with enlarged right hilar and right paratracheal lymph nodes. PET/CT scan showed18F-FDG uptake in the RUL mass, the hilar and paratracheal nodes, and multiple cervical and thoracic vertebrae. A brain MRI was negative for metastases.

CT-guided biopsy of the RUL mass identified adenocarcinoma, which was positive for thyroid transcription factor-1. He was diagnosed with metastatic lung cancer, adenocarcinoma subtype, T2N2M1b.

Molecular testing included NGS, which wasnegative for EGFR, ROS1, RET, BRAF, HER2,andNTRK; andIHC, which was negative forALKgene rearrangements.PD-L1 expression was found in 0% of cells, and TMB was not available.

The patient is started on pemetrexed, carboplatin, and pembrolizumab. After 2 cycles of treatment with the triplet, imaging revealed progression in the RUL (5.2 cm) and several bone lesions. Laboratory results showed increased carcinoembryonic antigen (34 ng/mL), decreased albumin (3.2 g/dL), and decreased hemoglobin (10.2 g/dL).

In this case, with a PD-L1&le;1%, what would be your choice of treatment?

I would not give a patient like this pembrolizumab alone. With a TMB that is 0 mutations per megabase, you would not give nivolumab plus ipilimumab. If we know that TMB is low and PD-L1 expression is absent, you could make a case for pemetrexed plus carboplatin or paclitaxel, carboplatin, and bevacizumab.

This patient was started on pembrolizumab plus pemetrexed and carboplatin, or the “pem-pem” regimen, and after 2 cycles, unfortunately, there was progressive disease. So it does not work in everyone.

What are the options for second-line therapy for this patient?

Keep in mind that the patient has already had a platinum-based combination, and they have already had a checkpoint inhibitor. Docetaxel is the longtime standard of care. In a randomized phase III trial, ramucirumab plus docetaxel were positive [for this indication]. Nivolumab plus ipilimumab are undergoing study in the second line in this setting.

I think one of the points of this case—and it is kind of likeEGFR-mutated lung cancer where if you use osimertinib [Tagrisso] in the first line, what do you do after that—is that when you give a combination, like platinum-based chemotherapy with a checkpoint inhibitor first, then what you have for third-line therapy moves up to the second line. What I typically do in this situation is use the docetaxel plus ramucirumab regimen, and that was done in this case.

Case 2 (continued):

The patient received docetaxel plus ramucirumab (Cyramza) and achieved a good partial response.

What is the rationale for this regimen for this patient?

This rationale was based on the REVEL study, which included participants with all histologies who had [progressed during or after a platinum-based chemotherapy regimen, with or without bevacizumab or maintenance therapy].8This is a different antiangiogenic, but it works by a different mechanism, and it was compared with docetaxel alone. And it is a positive study. You can see the hazard ratios were modest, with median OS at 0.86 and median PFS at 0.76. It has a doubling of the response rate.

The inclusion of the antiangiogenic agent seemed to have helped in this phase III trial, mostly in patients with aggressive disease. As you can see, depending on the duration of first-line therapy, you can see the OS and the other parameters were primarily in patients with rapid progression.

Case 3

A 71-year-old man presented with cough, dyspnea, and fatigue. His past medical history was remarkable for hypercholesterolemia, hypertension, and mild psoriatic arthritis without treatment for 3-plus years (only nonsteroidal anti-inflammatory drugs in the past). He was a former smoker with a 30 pack-years history, and he remained physically active. His ECOG performance status was 1.

A chest CT revealed a 2.5-cm mass in the left upper lobe and lymphadenopathy in the left hilar and bilateral mediastinal nodes. PET/CT scan showed18F-FDG uptake in the lung mass, left hilar and both mediastinal lymph nodes, and multiple bones metastases. A bronchoscopy and transbronchial lung biopsy were performed. Pathology identified the tumor as squamous cell carcinoma. IHC testing showed PD-L1 expression level of 65% and TMB of 16 mutations per megabase.

Ina patient with squamous cell lung cancer, high TMB, and high PD-L1 expression, what would be your choice of therapy?

I would use pembrolizumab alone in this case. Some might try the pembrolizumab plus chemotherapy option because the chemotherapy will offset the psoriatic arthritis.

KEYNOTE-024 is a first-line trial of pembrolizumab versus platinum chemotherapy.3This was very positive despite crossover for pembrolizumab in patients with high PD-L1 expression. The PFS and OS showed positive benefits. What I will point out in this is that the hazard ratio is among the factors that predicted benefit for pembrolizumab in squamous histology. You can see that these are not statistically significant because of the number of patients—it is only 56. But the hazard ratio is 0.35, so there is a reason for that. Patients with squamous cell lung cancer tend to have higher TMB and tend to be more homogenous. We say that the mutational burden is homogenous versus adenocarcinoma.

Squamous lung cancer for immunotherapy becomes a poster child because the mutational burden density is homogenous compared with adenocarcinoma. It is not just the number of mutations but also the kinds of mutations. We occasionally see never-smokers who have squamous cell disease, and in my experience, even though they have squamous cell disease, they end up having anEGFRmutation or anALKtranslocation. They are different beasts.

One other study to consider is KEYNOTE-042.9This again compared single-agent pembrolizumab versus doublet platinum chemotherapy, but instead of PD-L1 expression ≥50%, this was in patients with PD-L1 expression ≥1%. It is a positive study, but in OS, it is most positive in patients with the highest level of PD-L1 expression. Although still, a positive hazard ratio of 0.81.

KEYNOTE-047 is the trial that recently led to the FDA approval for pembrolizumab for all squamous cell lung cancer,10similar to KEYNOTE-189. The chemotherapy in this trial is different; it is either paclitaxel or nab-paclitaxel plus carboplatin. They did not study gemcitabine in this trial. It is a positive trial, although the data are relatively immature compared with the other studies, but it has already resulted in an FDA approval.

And if you look at this trial in squamous disease, you can see quite a bit of benefit even in patients with no PD-L1 expression. My guess, although the data are not out yet, is that this will be used in low PD-L1 expression but high TMB in a fair number of these patients.

Case 3 (continued):

The patient was started on pembrolizumab, and a PR was achieved. There was no exacerbation of underlying mild psoriatic arthritis.

Does this patient&rsquo;s autoimmune disease indicate that he is contraindicated for checkpoint inhibitor therapy?

The patient got single-agent pembrolizumab and had a PR. There was no exacerbation of the underlying psoriatic arthritis.

A study by Leonardi and colleagues published in theJournal of Clinical Oncologyin 2018 looked at the use of PD-L1 inhibitors in patients with NSCLC and an autoimmune disorder.11This retrospective analysis showed that only 23% of patients had exacerbations of autoimmune disease [4 required corticosteroids]. Autoimmune disease is a wide spectrum; there is severe, and there is mild. This person had mild psoriatic arthritis. They did not have lupus or a more severe form of autoimmune disease.


  1. Lindeman NI, Cagle PT, Aisner DL, et al. Updated molecular testing guideline for the selection of lung cancer patients for treatment with targeted tyrosine kinase inhibitors: guideline from the College of American Pathologists, the International Association for the Study of Lung Cancer, and the Association for Molecular Pathology.J Thorac Oncol. 2018;13(3):323-358. doi: 10.1016/j.jtho.2017.12.00
  2. Gandhi L, Rodr&iacute;guez-Abreu D, Gadgeel S, et al; KEYNOTE-189 Investigators. Pembrolizumab plus chemotherapy in metastatic non—small-cell lung cancer.N Engl J Med. 2018;378(22):2078-209 doi: 10.1056/NEJMoa1801005.
  3. Reck M, Rodr&iacute;guez-Abreu D, Robinson, AG, et al; KEYNOTE-024 Investigators. Pembrolizumab versus chemotherapy for PD-L1—positive non&ndash;small-cell lung cancer. N Engl J Med. 2016;375(19):1823-183 doi: 10.1056/NEJMoa1606774.
  4. 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.
  5. Hellmann MD, Ciuleanu T-E, 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. Carbone DP, Reck M, Paz-Ares L, et al; CheckMate 026 Investigators. First-line nivolumab in stage IV or recurrent non—small-cell lung cancer.N Engl J Med.2017;376(25):2415-242 doi: 10.1056/NEJMoa1613493.
  7. Bristol-Myers Squibb provides updates on the ongoing regulatory review of Opdivo plus low-dose Yervoy in first-line lung cancer patients with tumor mutational burden &ge;10 mut/Mb [news release]. Princeton, NJ: Bristol-Myers Squibb Company; October 19, 2019. Accessed February 25, 2019.
  8. 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, double-blind, randomised phase 3 trial.Lancet. 2014;384(9944):665-673. doi: 10.1016/S0140-6736(14)60845-X.
  9. Lopes G, Wu Y-L, Kudaba I, et al. Pembrolizumab (pembro) versus platinum-based chemotherapy (chemo) as first-line therapy for advanced/metastatic NSCLC with a PD-L1 tumor proportion score (TPS) &ge;1%; open-label, phase 3 KEYNOTE-042 study.J Clin Oncol. 2018;36(suppl 18; abstr LBA4).
  10. Paz-Ares L, Luft A, Vicente D, et al; KEYNOTE-407 Investigators. Pembrolizumab plus chemotherapy for squamous non—small-cell lung cancer.N Engl J Med. 2018;379(21):2040-2051. doi: 1056/NEJMoa1810865.
  11. Leonardi GC, Gainor JF, Altan M, et al. Safety of programmed death-1 pathway inhibitors among patients with non—small-cell lung cancer and preexisting autoimmune disorders.J Clin Oncol. 2018;36(19):1905-1912. doi: 10.1200/JCO.2017.77.0305.