Epidermal growth factor receptor expression levels could predict which non-small-cell lung cancer patients may benefit from a combination of Erbitux and chemotherapy, new biomarker data from the First-Line Erbitux in lung cancer trial has suggested.
"High EGFR expression is a tumor biomarker that can predict survival benefit from the addition of [Erbitux] to first-line chemotherapy in patients with advanced NSCLC," wrote researchers led by Robert Pirker of the Medical University of Vienna in a paper published earlier this month in Lancet Oncology. "Assessment of EGFR expression could offer a personalized treatment approach in this setting."
Merck KGaA, which markets Erbitux (cetuximab) in Europe, funded the reanalysis of the FLEX trial data and earlier this year submitted the results to the European Medicines Agency hoping to gain approval for Erbitux as a combination regimen with first-line platinum-based chemotherapy for advanced or metastatic NSCLC patients whose tumors have high EGFR expression.
EMA has already approved Erbitux for the treatment of patients with EGFR-expressing, KRAS wild-type metastatic colorectal cancer, and patients with squamous cell cancer of the head and neck.
The FLEX trial originally tested the benefit of combining Erbitux with chemotherapy in a NSCLC patient population loosely selected based on whether subjects had even one tumor cell showing expression of epidermal growth factor receptor by immunohistochemistry. However, according to Pirker, this broad selection criterion was not sufficient for EMA approval.
After receiving a negative decision for the Erbitux/chemo regimen from the EMA in 2009, FLEX researchers decided to give a second go at European approval and retrospectively reanalyzed the prospectively collected trial data by stratifying patients according to their level of tumor EGFR expression.
Pirker expects the EMA will respond "within weeks" to Merck's latest submission data from the FLEX trial, described in full in the Nov. 4 issue of Lancet Oncology.
According to Pirker, the researchers planned from the beginning to evaluate a variety of biomarkers in the FLEX trial to identify a best responder NSCLC subpopulation for Erbitux. Initially, the group considered several biomarkers, including KRAS and EGFR mutations in addition to measuring EGRF expression by immunohistochemistry, Pirker said.
At the end of the main FLEX trial, published in 2009 in the Lancet, the researchers first evaluated the frequency of positive immunohistochemical staining cells as a predictive marker, but frequency alone did not predict a difference in patient outcomes. KRAS and EGFR mutation also weren't predictive for treatment response in the study.
"So we came back to the original plan to look into detailed expression of EGFR, and that's what we have now published," Pirker said.
For the study published in Lancet Oncology, the researchers measured staining intensity as well as staining frequency and combined the two measures into an expression level score. "At each expression level we took the percentage of staining cells and then put it together," he said. "So 100 percent [of] cells with staining intensity level two would mean a score of 200, and a staining intensity of three in all cells would mean a score of 300."
The researchers used response data to select an outcome-based threshold score of 200 on a scale of zero to 300, the authors reported. Over 200 would be considered high EGFR expression and under 200 would be considered low.
In this retrospective analysis of the FLEX trial, EGFR immunohistochemistry data were available for more than 99 percent of the patients from the study cohort; 345 evaluable patients scored high and 776 patients scored low, they wrote.
In the main FLEX trial, patients given chemotherapy plus Erbitux survived longer than those in the chemotherapy-only group — a median of 11.3 months vs. only 10 months, respectively — but the 1.3 month improvement was not significant enough to warrant EMA approval.
By contrast, the new biomarker analysis showed that patients in the high EGFR expression group who received the Erbitux/chemo regimen survived an average of 12 months, while those in the chemotherapy-alone group lived only 9.6 months — a 3.6 month improvement. Meanwhile, patients with low EGFR expression did not show improved survival with the addition of cetuximab at all, surviving approximately the same amount of time on average in both treatment arms.
"We were able to show a significant difference in overall survival in the paper [for the high-expression group] with a hazard ratio of 0.73 instead of 0.87, which we had in the global population, so nearly 30 percent risk reduction," said Pirker. "We now believe this is clinically relevant difference in survival," he said.
According to Pirker, the new analysis is also strong because the response in high EGFR-expressing subjects was consistent across a variety of tumor histologies or types.
He explained that the because about 30 percent of the FLEX population had high EGFR expression, and the whole FLEX cohort represented 85 percent of all non-small-cell lung cancers, approximately 25 percent to 30 percent of all patients with advanced NSCLC would fall into the high-expression category and might benefit from the addition of cetuximab to cisplatin-based chemotherapy.
"For these patients, we believe we have a clinically relevant improvement by adding cetuximab to chemo," Pirker said. The Erbitux/chemo combination treatment "was not approved initially based … on the initial FLEX data, so I at least hope that it will be now approved for this selected group of patients."
Pirker held off from speculating on whether the EMA would explicitly require that advanced NSCLC patients be tested for their EGFR expression levels if the agency approves Erbitux's use based on the latest FLEX data. In the study, the researchers used a Dako immunohistochemistry kit to measure EGFR expression.
Pirker said immunohistochemistry had been used in earlier phase II trials of cetuximab with positive results, so the researchers decided to maintain that strategy in FLEX.
"The data [from the biomarker analysis] in my opinion are quite robust, and are consistent with the mode of action of cetuximab and also we have a clinically well established technique with immunohistochemistry," he said.
Retrospective data may have to suffice in this case, according to Pirker, noting that in light of the new findings, a prospective, randomized Phase III validation trial where some patients with high EGFR expression levels wouldn't get the Erbitux/chemo combination would be hard to ethically justify in his opinion.
He said he is not aware if the companies that market cetuximab in the US — Lilly and Bristol-Myers Squibb — may be seeking a similar extension with the US Food and Drug Administration.
Lilly subsidiary ImClone Systems and Bristol-Myers Squibb submitted an application to the FDA for the approval of Erbitux in NSCLC at the end of 2008, but withdrew their application in January 2009 when the FDA asked for more information about the difference between the chemical composition of Erbitux used in the FLEX trial and the version of the drug marketed in the US. At the time, the companies said that they were planning to resubmit another application to the FDA for Erbitux in this indication at a later time.
In 2010, researchers reported in the Journal of Clinical Oncology the results of an additional Phase III trial, BMS-099, of Erbitux plus a taxane and carboplatin as a first-line treatment for metastatic non-small cell lung cancer. The study did not meet its primary endpoint of progression-free survival as assessed by an independent radiology review committee.
Last week, Dako, which provided the immunohistochemistry testing kit used in the FLEX trial, announced a partnership with BMS to develop pharmacodiagnostic tests intended to identify patients more likely to benefit from treatment with investigational drug candidates under development.
Dako and BMS did not respond to inquiries about whether this partnership includes the development of a companion test for Erbitux in the US.
Have topics you'd like to see covered in Pharmacogenomics Reporter? Contact the editor at mashford [at] genomeweb [.] com.