A Decade of Lessons Learned From EGFR-Targeted Therapy

Get Permission

To summarize the lessons learned from the development of epidermal growth factor receptor (EGFR)-targeted therapy, one only has to go back about 10 years, according to Frances A. Shepherd, MD, Scott Taylor Chair in Lung Cancer Research at Princess Margaret Cancer Centre and Professor of Medicine at the University of Toronto.

“In 10 years, we’ve gone from first- to fourth-generation EGFR tyrosine kinase inhibitors,” she said during the closing Plenary Session at the 2016 World Conference on Lung Cancer in Vienna, Austria.1 “The speed is dizzying, but we are in this race together, and we must keep up.”

Where We Were

At the end of the past century, Sanger sequencing was the standard test used to detect somatic mutations, and mutations were tested individually. KRAS was the “oncogene du jour,” and it and most other mutations were mainly of interest for their prognostic effects. No driver mutations had been identified.

At that time, EGFR was assessed by immunohistochemistry staining for protein expression, and expression was present in > 90% of non–small cell lung cancer (NSCLC). “And the misleading term ‘overexpression’ was bandied about indiscriminately, even by the most learned of oncologists,” admitted Dr. Shepherd.

To me, pathologists are now the rock stars of oncology. They are part of our team, and they are directing our treatment.
— Frances A. Shepherd, MD

Tweet this quote

“Fifteen or so years ago, we had some interesting new drugs,” she said. Exciting phase II data had been presented for the EGFR tyrosine kinase inhibitors gefitinib (Iressa) and erlotinib (Tarceva), and EGFR monoclonal antibodies cetuximab (Erbitux) and necitumumab (Portrazza) had been studied in several phase II and III trials.

“We had these exciting results from phase II trials with EGFR tyrosine kinase inhibitors, so we all knew that if we added them to our chemotherapy doublets, things would get better,” she said. “There was no selection based on EGFR, but why bother when 95% of samples would be positive?” Tissue collection was requested, but not mandatory, for any study.

However, in studies of gefitinib and platinum-based chemotherapy, no benefit was seen in approximately 2,000 patients. Another 1,000 patients were studied in trials of erlotinib and platinum-based chemotherapy, and again no benefit was observed. “At that point, there was not only no benefit, but a hint at some early harm by the combination,” she acknowledged.

What Was Learned?

Before randomly performing combination studies, preclinical studies should be performed to understand the interaction of the agents in the pathway, she said. EGFR tyrosine kinase inhibitors cause G1 arrest, thereby reducing the effectiveness of chemotherapy.

At exactly the same time as first-line studies were being done, the BR.21 placebo-controlled trial of single-agent erlotinib in the third-line setting was being performed on patients who had no other options. This trial was positive, demonstrating a significant survival advantage for erlotinib. “Barely a second had passed before we were told that EGFR tyrosine kinase inhibitors didn’t work in wild-type, and that all of our results were driven by a subset of mutated cancers that we hadn’t identified,” she continued. This was not in fact true, although the benefit of erlotinib in the wild-type subgroup was modest at best.

After the identification of these mutations, the next step was to determine whether tyrosine kinase inhibitors were superior to chemotherapy in the mutated subgroup. Numerous trials compared EGFR tyrosine kinase inhibitors with platinum-based chemotherapy for first-line treatment, and EGFR tyrosine kinase inhibitors become the standard of care for EGFR-mutated NSCLC. “Does it matter what tyrosine kinase inhibitor we use? No,” stated Dr. Shepherd. “It also doesn’t matter if you are Asian or Caucasian. You benefit if you have the driver mutation.”

Progression-free survival hovered around the 1-year mark for all of these studies, but no survival benefit was seen in any of the early trials; then afatinib (Gilotrif), a second-generation irreversible inhibitor, was developed and similar trials were performed. When two randomized trials (afatinib vs ­cisplatin/pemetrexed [Alimta] or cisplatin/gemcitabine) were pooled, a statistically significant overall survival benefit was observed. This benefit seemed to be limited to the subgroup of patients with EGFR exon 19 deletions.

Monoclonal Antibodies

In studies of monoclonal antibodies, those with higher numbers of patients had statistically significant outcomes compared with studies of smaller populations. In fact, though, most studies showed similar results with modest benefits from the addition of the EGFR monoclonal antibody. Dr. Shepherd underlined the importance of evaluating the clinical relevance of the response and survival benefits in clinical trials and not just the statistically significant P value, comparison to, rather than focusing on one or the other. “We cannot just look at a hazard ratio, because two times nothing is still nothing, and we cannot just look at a P value,” she said. “We have to put the package together and have clinical and statistical relevance.”

In regard to mutation status with monoclonal antibodies, Dr. Shepherd noted that the data were unexpected. “When we did the KRAS studies in lung cancer, we were very confident that the modest, clinically questionable benefit in the studies overall would become greater and clinically relevant in KRAS wild-type lung cancer, and sadly that was not the case,” she admitted. The interaction of chemotherapy with monoclonal antibodies or tyrosine kinase inhibitors can undoubtedly be different, and molecular driver mutations may not have the same effects or interactions in different tumor types or even within tumor subtypes.

Back to Tyrosine Kinase Inhibitors

Resistance always develops, and the most frequent resistance mechanism is the EGFR T790M mutation, she said. Osimertinib (Tagrisso) received U.S. Food and Drug Administration approval in 2016, based on phase II AURA 3 study results showing an enormous benefit in favor of osimertinib vs chemotherapy in patients with EGFR T790M mutation. “There also was activity in the brain with osimertinib, and since 50% of our patients with EGFR drivers have brain metastases at some time, this is very good news for us,” Dr. Shepherd added.

“Will we be able to prevent or delay T790M resistance by taking our second- or third-generation EGFR T790M inhibitors into the first line?” The FLAURA study, which includes patients with EGFR-sensitizing mutations, has completed accrual, and ASP8273 currently is accruing. In the absence of trial results, these inhibitors are not being used in the first-line setting yet, continued Dr. Shepherd. 

“But we are already up to fourth-generation agents now,” she said. AUY922 has shown activity in exon 20, and AZD3759 currently is in development, specifically for patients with brain metastases.

“How naive we were a decade ago!” she emphasized The EGFR mutation landscape is much more complex than simply exons 19 and 21. Different EGFR mutations have differing implications; rare mutations are being identified much more frequently; and EGFR mutations may coexist with other mutations. Now we understand that the mutation spectrum may evolve over time and that we have to analyze tumors repeatedly throughout a patient’s cancer journey to understand the mechanisms of resistance. This means that repeated molecular testing is mandatory for us now in lung cancer.

There are hurdles ahead, and the agents are costly. “To identify these mutations, we must make available timely and broad molecular diagnostic testing and circulating tumor DNA testing and have the ability to perform multiple invasive biopsies,” added Dr. Shepherd. However, circulating tumor DNA testing is becoming more widely available and may alleviate the need for invasive biopsies. Clearly, she said, “one-by-one mutation testing is a thing of the past.”

“To me, pathologists are now the rock stars of oncology,” she added. “They are part of our team, and they are directing our treatment.” Molecular testing is now standard practice for lung cancer management and should be done at diagnosis for all targetable mutations; next-generation sequencing with broad molecular panels should replace individual gene testing; and longitudinal liquid biopsies should become the new standard of care, according to Dr. Shepherd. ■

Disclosure: Dr. Shepherd is a consultant for Lilly, AstraZeneca, GlaxoSmithKline, Boehringer Ingelheim, Recombio, Synta, Bristol-Myers Squibb, and Roche; has received honoraria from Lilly, AstraZeneca, Bristol-Myers Squibb, Merck Serono, Roche/Genentech, Merck/Schering Plough, and Boehringer Ingelheim; and is a stock shareholder in Lilly and AstraZeneca.


1. Shepherd F: A life in thoracic oncology: Reflections from giants on milestones in the treatment advances in lung cancer. 2016 World Conference on Lung Cancer. Abstract PL05. Presented December 7, 2016.