August 25, 2015
17 min read

Immunotherapy ‘revolutionary’ for lung cancer, but questions remain about cost, biomarkers

You've successfully added to your alerts. You will receive an email when new content is published.

Click Here to Manage Email Alerts

We were unable to process your request. Please try again later. If you continue to have this issue please contact

The potential of immunotherapy to treat several malignancies has generated tremendous excitement in the oncology community, and this optimism has become a reality this year for patients with lung cancer.

Checkpoint inhibitors have demonstrated significant activity in lung cancer. More than 100 clinical trials are evaluating checkpoint inhibitors as single-agent treatments or in combination with other anti-cancer therapies for lung cancer.

“Immunotherapy has ushered in a new optimism for patients and a new era of clinical investigation in oncology, specifically in lung cancer,” Edward S. Kim, MD, chair of the department of solid tumor oncology at Levine Cancer Institute at Carolinas HealthCare System and a HemOnc Today Editorial Board member, said in an interview.

Edward S. Kim

Despite the optimism, some key questions related to the use of these agents for lung cancer remain unanswered.

Their cost — nearly 10 times that of standard-of-care docetaxel — must be weighed against the survival benefit and toxicities, experts said. Further, the optimal predictors of response, including the role of PD-L1 expression, have yet to be defined.

“We are very unclear about the biomarkers,” Kim said. “If you trace back to renal cell carcinoma and melanoma ... researchers never could really find a direct association [between outcomes and different biomarker cutoffs]. You’d see one cutoff and say, ‘Oh, there’s a trend,’ and then you’d see another cutoff and there wouldn’t be a trend anymore.”

HemOnc Today spoke with lung cancer experts about the promise checkpoint inhibitors have demonstrated for squamous and nonsquamous non–small cell lung cancer, how PD-L1 status may help determine patient eligibility for treatment, and whether cost will prove prohibitive.

Nivolumab sets the pace

Nivolumab (Opdivo, Bristol-Myers Squibb) — a monoclonal antibody that binds to the programmed cell death-1 (PD-1) receptor and blocks its interaction with ligands PD-L1 and PD-L2 — received FDA approval in March for treatment of patients with metastatic squamous NSCLC who progressed on or after platinum-based chemotherapy.

A pair of phase 3 trials presented at the ASCO Annual Meeting this spring showed single-agent nivolumab conferred a significant OS advantage over docetaxel in second-line treatment of advanced or metastatic squamous and nonsquamous NSCLC.

In the CheckMate 017 trial, Spigel and colleagues enrolled patients with squamous-cell NSCLC. Patients assigned nivolumab achieved longer median OS than those assigned docetaxel (9.2 months vs. 6 months; HR = 0.59; 95% CI, 0.44-0.79). More patients assigned nivolumab also achieved 1-year OS (42% vs. 24%).

Based on these data, CheckMate 017 was stopped early in January and the FDA approved nivolumab for this indication.

The CheckMate 057 trial, conducted by Paz-Ares and colleagues, showed nivolumab also benefited patients with nonsquamous NSCLC.

Patients in that trial assigned nivolumab achieved longer median OS (12.2 months vs. 9.4 months; HR = 0.73; 95% CI, 0.59-0.89) and a higher rate of 1-year OS (50.5% vs. 39%) than those assigned docetaxel.

These trials may have established nivolumab as a new standard of care for NSCLC.

“The approach of this immunotherapy in lung cancer is going to be revolutionary,” David Carbone, MD, PhD, director of the James Thoracic Center at The Ohio State University Comprehensive Cancer Center — Arthur G. James Cancer Hospital and Richard J. Solove Research Institute, told HemOnc Today. “It’s not going to work in everyone and it’s not without toxicity, but in a subset of patients, we get good, durable objective responses with good quality of life and tolerability.”

PD-L1 is a potential biomarker in the treatment of NSCLC, although its use is a work in progress, according to Roy S. Herbst, MD, PhD. “Right now, it’s not a good enough biomarker for me to exclude someone from immune checkpoint therapy, because even the group that was negative for the biomarker still seemed to have at least equal benefit to docetaxel with a potential tail on the curve,” Herbst said.

Photo by Rob Lisak

Still, there are noteworthy differences between the squamous and nonsquamous populations that need to be considered, according to Roy S. Herbst, MD, PhD, Ensign professor of medicine (medical oncology) and chief of medical oncology at Yale Cancer Center and Smilow Cancer Hospital at Yale-New Haven.

“The magnitude may be about the same, but squamous patients are all smokers, whereas about 15% of the nonsquamous patients are nonsmokers with either EGFR or ALK mutations,” Herbst told HemOnc Today. “Those patients probably aren’t benefitting as much from the single-agent checkpoint inhibitors. That’s a big difference between squamous and nonsquamous patients.”

Other immunotherapies carve a similar path

Another PD-1 inhibitor under investigation for NSCLC is pembrolizumab (Keytruda, Merck), and other PD-L1 inhibitors include atezolizumab (MPDL3280A, Genentech/Roche) and MEDI4736 (MedImmune, AstraZeneca).

Alexander I. Spira, MD, PhD, FACP, director of Virginia Cancer Specialists Research Institute in Fairfax, Virginia, and colleagues presented data at ASCO that indicated patients with NSCLC assigned atezolizumab achieved longer median OS than those assigned docetaxel (11.4 months vs. 9.5 months; HR = 0.78; 95% CI, 0.59-1.03). Based on these data, researchers initiated three phase 3 trials of atezolizumab in combination with chemotherapy for first-line treatment of NSCLC.

“The survival advantages we’ve observed with these various immuno-oncology agents are impressive,” Kim said. “We don’t usually beat chemotherapy head-to-head. For years we’ve tried to beat it with doublets or other additional agents. Now, we’ve reset the boundary.

Alexander I. Spira, MD, PhD, FACP
Alexander I. Spira

“Docetaxel will not have to be a control arm in the second line any longer for lung cancer,” Kim added. “It will be nivolumab or one of the other immuno-oncology agents. That’s a big shift from where we were 5 years ago.”

Garon and colleagues presented data at the AACR Annual Meeting in April that demonstrated the promise of pembrolizumab in this setting. The results — subsequently published in The New England Journal of Medicine — showed the response rate among patients with locally advanced or metastatic NSCLC was 19.4% (95% CI, 16-23.2). Researchers reported median OS of 12 months (95% CI, 9.3-14.7) and median PFS of 3.7 months (95% CI, 2.9-4.1).

Results of a trial evaluating MEDI4736 plus the CTLA-4 inhibitor tremelimumab (AstraZeneca) — presented at ASCO by Antonia and colleagues — demonstrated response rates of 33% in patients with PD-L1–positive tumors and 27% among those with PD-L1–negative tumors. The agent has also demonstrated efficacy as monotherapy, and a phase 3 study is underway.

The efficacy is particularly encouraging given that progress in this setting typically is measured in weeks.

“We’re not talking about marginal survival benefits. We’re not talking about low PFS rates,” Kim said. “We are seeing a group of patients surviving longer.”

The biomarker debate

If any debate emerged from the lung cancer immunotherapy trials, it is whether overexpression of PD-L1 must guide patient selection for treatment.

Identifying an appropriate biomarker — as well as an appropriate cutoff for expression of that biomarker — is important to prevent patients from having to endure the costs and toxicities associated with treatments from which they may not benefit. However, current data are inconclusive with regard to the necessity of PD-L1 expression for benefit.

In the nonsquamous population, Paz-Ares and colleagues demonstrated that nivolumab benefit appeared greater in patients with PD-L1 expression at all cutoffs — which included 1%, 5% and 10% expression levels in tumor cells. Among patients with PD-L1 expression in 1% or more of tumor cells, those who received nivolumab demonstrated a median OS that exceeded 17 months, whereas patients who received docetaxel demonstrated a median OS of 9 months. Patients with 10% or greater PD-L1 expression demonstrated the greatest OS benefit with nivolumab (HR = 0.4; 95% CI, 0.27-0.59).

Higher PD-L1 expression also appeared associated with a greater response rate with pembrolizumab in the study by Garon and colleagues ( 50% PD-L1 expression, 45% vs. < 1% PD-L1 expression, 10.7%), as well as prolonged median PFS in the atezolizumab study by Spira and colleagues (9.7 months vs. 3.9 months; HR = 0.57; 95% CI, 0.28-1.11).


However, patients in the squamous population of the CheckMate 017 trial derived benefit from nivolumab regardless of PD-L1 expression.

PD-L1 overexpression potentially will be one biomarker in the treatment of NSCLC, although its use is a work in progress, Herbst said.

“Ultimately, I believe PD-L1 is going to be a biomarker,” Herbst said. “Right now, it’s not a good enough biomarker for me to exclude someone from immune checkpoint therapy, because even the group that was negative for the biomarker still seemed to have at least equal benefit to docetaxel with a potential tail on the curve. There are also major issues with tissue heterogeneity. There could even be some long-term benefit for PD-L1–negative patients.”

However, the data have showed that anyone with at least 1% positivity of PD-L1 expression — a very low threshold — may benefit, rendering the biomarker relatively ineffective, Spira told HemOnc Today.

“The only section where there is not a significant benefit is the low-expression group or the negative group and, in that scenario, the therapy is no different than docetaxel,” Spira said. “Given the side-effect profile, there is probably not a patient in the world who would not want to receive atezolizumab compared with docetaxel.”

The biggest problem is inconsistency in the biomarker assays, Spira added.

“Given what we’ve seen and what we know, PD-L1 status is going to become irrelevant,” Spira said. “Part of the issue is, it’s really difficult to get across common language because of the way these drugs are being approved. All the assays are being done for one particular drug. This is unique compared with everything else in oncology, and it is going to be hard to compare biomarkers.”

There has been some headway in the development of a biomarker for all the anti–PD-L1/PD-1 drugs, but researchers await data to determine each biomarker’s practical importance as the drugs reach FDA approval, Spira added.

The ways in which researched tested and defined “high” PD-L1 expression differs in each of these studies, Matthew D. Hellmann, MD, a medical oncologist specializing in lung cancer and member of the Immunotherapeutics Group at Memorial Sloan Kettering Cancer Center, said in an interview. Despite these questions, PD-L1 expression may still help inform treatment decisions.

“Across nearly all studies, PD-L1 positivity appears to enrich for responses. But, positivity doesn’t guarantee response, nor does lack of PD-L1 expression rule out the possibility of response,” Hellmann told HemOnc Today. “I do not think PD-L1 is a necessary biomarker for choosing whether to use PD-1 or PD-L1 blockade in the second-line setting. But, however imperfect PD-L1 testing may be today, PD-L1 testing may still be a useful tool for helping to guide treatment decisions in the first-line setting or, possibly in the future, when deciding between single-agent or combination immunotherapy.”

The more important research question — according to Hellmann — is moving beyond the focus on PD-L1 expression to understand why some patients with lung cancer respond well to PD-L1 blockade while others do not.

“The challenge to identify more sensitive biomarkers of response remains open,” Hellmann said. “We are working hard now to help answer that question so that we may have better and more personalized ways to treat patients in the future.”

The ability to select a specific population of PD-L1–positive patients who can greatly benefit from immunotherapy would be a huge leap forward, Herbst said.

“Maybe we need a better PD-L1 assay,” Herbst said. “PD-L1 is induced by interferon, so it can vary from different time points in patients and there’s also heterogeneity. It’s a very imperfect biomarker to say the least. However, there’s no doubt in my mind that the right biomarker could make a difference. Perhaps this will involve using techniques of DNA or RNA sequencing or other genomic approaches.”


Impending FDA approval

The uncertainty surrounding PD-L1 expression has rendered it difficult to predict whether the biomarker will be included on the label for additional approvals of immunotherapies for NSCLC.

“We just need to get these drugs to patients,” Carbone said. “The FDA approved nivolumab for all-comers with squamous disease, but if the FDA approves it only for those nonsquamous patients who are positive for PD-L1, then we will have to use this test for all of these patients until something better is found.”

Benjamin C. Creelan, MD, medical oncologist at Moffitt Cancer Center, told HemOnc Today — when commenting on the nivolumab studies at ASCO — that he believes nivolumab will be approved for the overall population.

“Would I not treat a patient who is PD-L1 negative?” Creelan said. “I would still have a discussion with them and I would still give the drug if the patient wanted it. If a patient knows there is a 10% to 12% response rate with this drug even if they are PD-L1 negative, I’m guessing a lot of patients are going to go for it. Most people would rather go for a smaller chance of a long-term durable remission with fewer side effects and better quality of life than going with our traditional options.”

These therapies likely will also receive approval for the first-line setting. Clinical trials in this area have finished accrual, and researchers await the results.

“If that shows that the immunotherapy is positive in first line, it’ll change everything, and some patients may never have to have chemotherapy,” Carbone said. “One feature of all of the first-line trials is that they are limited to PD-L1 positive patients. We won’t know about the PD-L1–negative patients in first line, at least at first.”

The cost hurdle

The excitement surrounding these advances may be tempered by cost.

Nivolumab costs approximately $60,000 a month, compared with the $6,000 monthly price of docetaxel. Payers and the federal government must determine if the survival benefit is worth the price.

“You have to think about this from a patient’s point of view,” Herbst said. “If, in fact, a combination of drugs clearly improves survival and the toxicity is acceptable, I’d want to use it. There would be some cost constraints, but we need to be really careful and sensitive when we measure those costs.”

Bristol-Myers Squibb is providing nivolumab free of charge in Europe for patients with advanced lung cancer who are not eligible for surgery; however, this will only be an option while the drug is fast-tracked through the licensing process.

“The costs are becoming staggering,” Kim said. “But, I don’t believe in picking on the ‘new kid on the block’ just because it’s expensive. ... There just has to be a fundamental shift on how the payers and the reimbursements are integrated into our process. As a clinician, I want any type of drug that can benefit a patient available to me so I can treat them.”

On the one hand, physicians want to improve their patients’ chances of survival; on the other, they don’t want to see their patients go bankrupt.

“A lot of it is driven by the insurance coverage,” Carbone said. “But, with all things being equal, if patients are told they could pay $50,000 and live another year, they would probably say, ‘Yes.’ That year could be a good, high-quality year with these drugs.”

The cost problem may resolve with time, Carbone added.

“It is really unfortunate that they do cost so much, but I’m hopeful that with so many competitors, competition will drive these prices down,” Carbone said. “As time goes on, the drugs will come off patents and biosimilars will come out, so 20 years from now, we’ll be in a better situation. But, with one drug approved on the market, it is going to be difficult.”


If the drugs are limited to a biomarker-positive population, providers likely would pay for a greater proportion of the overall cost of therapy.

However, the debate about the predictive nature of the biomarkers leaves that approach on shaky ground.

In a Kantar Health Oncology Market Access survey conducted in April, 86 commercial payers expressed lukewarm support for the value of biomarkers for cancer immunotherapies.

The poll showed 58% of payers believe biomarkers are useful in some cancers but not others. This represented an increase from 31% of payers who felt that way in 2014 and 27% who expressed that opinion in 2013.

However, the poll showed that only 37% of payers believe biomarkers improve overall outcomes in the target population, and 29% of payers believe biomarkers help save money on drug costs. Fewer than one-quarter (23%) of the payers surveyed have coverage policies in place for most biomarker tests.

Although payers are unlikely to deny coverage if the drugs are considered clinically appropriate, defining the appropriateness could have many variables, including PD-L1 expression.

One approach that may help reduce costs is defining an appropriate length of treatment.

In both CheckMate trials, patients received nivolumab every 2 weeks indefinitely for as long as it worked.

“The day Paz-Ares presented his data, the BMS stock price dropped 4%,” Spira said. “The thinking online was that patients were treated, on average, less than expected. Patients don’t need to stay on the drug until progression. Most patients only got benefit in the 6- to 8-month time frame, which is significantly less than people thought.”

A current research questions focuses on whether the immune system will continue to fight the cancer after treatment is stopped.

“I don’t think people need to be treated indefinitely, because once the immune system is reset you can easily take a break,” Spira said. “We never do that with well-tolerated medicine, but considering the cost, I think it’s a reasonable thing to do. Biologically it makes sense — the half-lives of these antibodies are reasonable. The effects should be long-lasting, and one could probably take a break for some time to save on cost.”

Until costs are addressed by all involved parties — pharmaceutical companies, providers, physicians and the federal government — it is hoped that patients are not uniquely burdened about cost when making a decision, Hellmann said.

“Cost is a critically important question for society as a whole to work on, to find ways to ensure that our health care system remains innovative, yet solvent,” Hellmann said. “But the individual treatment decisions I make together with my patients are based on what the best medical option is for them and their family rather than cost.

“I believe optimal care should be available to all,” Hellmann said. “For patients who respond to immunotherapy, those who are able to live better and longer than they ever could have with other medicines, the benefits are priceless.” – by Anthony SanFilippo


Garon EB, et al. N Engl J Med. 2015;doi:10.1056/NEJMoa1501824.

Kantar Health. What are payers pondering about the new wonder drugs? Available at: Accessed July 28, 2015.

The following were presented at the ASCO Annual Meeting; May 29-June 2, 2015; Chicago:

Antonio SJ, et al. Abstract 3014.

Paz-Ares LG, et al. Abstract LBA109.

Spigel DL, et al. Abstract 8009.

Spira AI, et al. Abstract 8010.

For more information:

David Carbone, MD, PhD, can be reached at The Ohio State University Wexner Medical Center, James Thoracic Oncology Center, 300 W. 10th Ave., Columbus, OH 43210; email:

Benjamin C. Creelan, MD, can be reached at Moffitt Cancer Center, 12902 USF Magnolia Drive, Tampa, FL 33612; email:

Matthew D. Hellmann, MD, can be reached at Memorial Sloan Kettering Cancer Center, 1275 York Ave., New York, NY 10065; email:


Roy S. Herbst, MD, PhD, can be reached at Yale Cancer Center, 33 Cedar St., New Haven, CT 06510; email:

Edward S. Kim, MD, can be reached at Levine Cancer Institute, 1021 Morehead Medical Drive, Charlotte, NC 28204; email:

Alexander I. Spira, MD, PhD, FACP, can be reached at Virginia Cancer Specialists, 8503 Arlington Blvd., Suite 400, Fairfax, VA 22031; email:

Disclosure: Carbone reports consultant roles with AstraZeneca, Bristol-Myers Squibb and Genentech. Creelan reports speakers fees from Bristol-Myers Squibb. Spira reports consultant roles with Novartis and Roche. Hellman, Herbst and Kim report no relevant financial disclosures.



Should checkpoint inhibitors be limited in the front-line NSCLC setting based on PD-L1 expression?


PD-L1 expression should guide selection of patients for clinical trials in the front-line setting.

Immunotherapy, specifically nivolumab (Opdivo, Bristol-Myers Squibb), has been approved in the second line for squamous non–small cell lung cancer based on the results from the CheckMate 017 trial. In that trial, patients assigned nivolumab achieved a median OS of 9.2 months, whereas the median OS in the docetaxel arm was 6 months (HR = 0.59; 95% CI, 0.44-0.79). The benefit associated with nivolumab persisted regardless of PD-L1 expression levels.

Tarek Mekhail

However, the CheckMate 057 trial in patients with adenocarcinoma demonstrated similar results with some intriguing findings regarding the correlation of response and PFS based on PD-L1 expression. PD-L1 positivity at 1%, 5% and 10% expression levels appeared associated with benefit from nivolumab.

Further, the study published on pembrolizumab (Keytruda, Merck) in The New England Journal of Medicine showed significant improvement in response rates for a selected population with PD-L1 expression of more than 50% compared with lower response rates for patients with less PD-L1 expression.

In selection for patients with more treatment options — such as patients in the first-line setting — the bar is higher. I suspect this may partially explain why the FDA has been slower in approving checkpoint inhibitors for adenocarcinoma — where we have more treatment options — despite fairly comparable results to squamous cell carcinoma.

This also applies to front-line treatment, where the treatment options and combination of treatments are much more abundant than the second- or third-line settings. Consequently, it is harder to prove efficacy in the front-line setting compared with standard treatment.

When the bar is higher and there are more options, selection has to be more restricted to patients who have a higher chance of responding to treatment based on endpoints that are harder to reach. PD-L1 expression may help define the patient population appropriate for treatment.

None of the checkpoint inhibitors are currently approved in the front-line setting, awaiting supportive evidence. With these parameters, we should select the population based on PD-L1 expression provided the evidence we have so far. Although PD-L1 expression testing has not yet been standardized, I believe more evidence is yet to come.



Garon EB, et al. N Engl J Med. 2015;doi:10.1056/NEJMoa1501824.

The following were presented at ASCO Annual Meeting; May 29-June 2, 2015; Chicago:

Paz-Ares LG, et al. Abstract LBA109.

Spigel DL, et al. Abstract 8009.

Tarek Mekhail, MD, MSc, FRCSI, FRCSEd, is medical director of the Thoracic Oncology Program at Florida Hospital Cancer Institute, associate professor at University of Central Florida and a HemOnc Today Editorial Board member. He can be reached at Disclosure: Mekhail reports no relevant financial disclosures.


Certain patients with non–small cell lung cancer may derive benefit from immunotherapy in the first-line setting regardless of biomarker expression.

The recent data on the anti–PD-1 antibody, nivolumab (Opdivo, Bristol-Myers Squibb) — in the squamous population (CheckMate 017) and non-squamous population (CheckMate 057) — highlight the high efficacy, low toxicity and substantial clinical benefits of the PD-1 checkpoint inhibitors in the second-line setting in advanced/metastatic NSCLC when compared with docetaxel chemotherapy, regardless of PD-L1 expression. However, impressive data from Garon and colleagues published in The New England Journal of Medicine also demonstrate that the anti–PD-1 antibody pembrolizumab (Keytruda, Merck) had impressive sustained responses and high clinical benefit in selected highly expressing PD-L1–positive patients (> 50% via immunohistochemistry). The use of PD-L1 as a selective biomarker, varying methods for detection and cut-offs levels have been vigorously debated and have been an area of substantial controversy in the second-line setting.

Laura Q. M. Chow

In the first-line setting where platinum-based cytotoxic chemotherapy has had a firm established role with known response rates, PFS and OS, biomarkers for molecular drivers (eg, EGFR and ALK gene mutations) have been successful for selecting patients who may derive even greater benefit in this setting than from chemotherapy. The impressive tolerability and substantial benefit of the PD-1 checkpoint inhibitors in the second-line setting has led to high enthusiasm for use of these agents in the first-line setting. However, the role of PD-L1 as a biomarker to select responders to immunotherapy in the first-line setting and the benefits of immunotherapy when compared with chemotherapy are still pending data from current ongoing clinical trials. In the clinical management of NSCLC, there are many patients for whom tolerance to the toxicities of cytotoxic chemotherapy would be challenging — the extreme elderly, the frail, less fit patients, those with a high risk for infection, those with conflicting comorbidities, and some patients who may have a strong philosophical opposition to chemotherapy. These patients may potentially derive some benefit with PD-1 checkpoint immunotherapy in the first-line setting regardless of biomarker status due to the differing and potentially lower toxicity profile of the immune checkpoint inhibitors.

It is clear that immunotherapy with a checkpoint inhibitor is an exciting new option that hopefully many patients with lung cancer may see in their treatment algorithm, whether they are biomarker selected or not, regardless of the line of therapy.



Garon EB, et al. N Engl J Med. 2015;doi:10.1056/NEJMoa1501824.

The following were presented at ASCO Annual Meeting; May 29-June 2, 2015; Chicago:

Paz-Ares LG, et al. Abstract LBA109.

Spigel DL, et al. Abstract 8009.

Laura Q. M. Chow, MD, treats patients with lung, head and neck, and thyroid cancers at Seattle Cancer Care Alliance. She is an associate member of Fred Hutchinson Cancer Research Center and associate professor of medical oncology at University of Washington School of Medicine. She can be reached at Disclosure: Chow reports no relevant financial disclosures.