ImClone Systems has decided to pursue several new indications for Erbitux, its personalized medicine for colorectal cancer, but because of a miscalculation in a late-stage trial those new indications might not be personalized after all.
And, for each of the additional Erbitux indications, ImClone is hoping to remove the EGFR IHC-test requirement, and possibly replace it with another biomarker test, or perhaps nothing at all.
As ImClone readied Erbitux for clinical trials 2 years ago, the company believed that by enriching the trial-subject population for the drug's target, EGFR, it would be able to increase the proportion of patients who responded to the drug, according to Eric Rowinsky, ImClone's chief medical officer. But because of the EGFR immunohistochemistry test's limitations and the sheer prevalence of EGFR in many kinds of carcinomas, it became a hindrance.
"The problem was that there was a naïve notion that immunohistochemistry for [epidermal growth factor receptor] would be useful," Rowinsky said. "But that was done in the earliest trials of Erbitux, and got carried along with the approval of Erbitux."
ImClone now is seeking to expand the drug's indications to several different carcinomas, while leaving the EGFR IHC testing requirement behind.
"Immunohistochemistry as a test is a terrible test — every laboratory has its own procedure. As a test it just doesn't stand up — it can't be quantitated very well, nor qualified. We need to look for other predictors, and we're doing that in several trials with our partner Bristol-Myers [Squibb]."
The first of those new indications is head and neck cancer, for which the company will submit a supplemental Biologics License Application with the FDA seeking approval of the drug alone and in combination with radiation therapy, ImClone said last week in a statement. "Right now we're completing our submission … we just had a strongly positive group of trials — including a randomized trial with radiation — that very surprisingly showed [an overall] survival advantage," said Rowinsky.
Friedman Billings Ramsey analyst Jim Reddoch told the Associated Press earlier this month that the US market for head-and-neck cancer treatments totaled about $400 million annually.
Erbitux was approved in February 2004 through a partnership between Bristol-Myers Squibb and ImClone. The drug brought in total revenues of $261 million in 2004, according to a presentation given by John Elicker, vice president of BMS investor relations, earlier this month.
Analyst projections for full-year 2005 Erbitux sales average $387.6, according to an e-mail from an ImClone spokesperson. The company does not share internal projections, she said.
At the 2004 American Society of Clinical Oncology, ImClone presented its Erbitux-and-radiation-therapy findings, along with data supporting the drug's use as a second-line stand-alone therapy, in patients with squamous cell carcinoma of the head and neck. The combination-therapy data were gathered in a study of 424 patients with locally advanced SCCHN, while the stand-alone therapy data were generated by a study of 103 SCCHN patients with advanced recurrent with or without metastasis, the company said.
Other indications the company is pursuing include pancreatic cancer, for which trial data will be available in 2006, lung cancer, and a "variety of other carcinomas in combination with Genentech's Avastin," said Rowinsky. Other trials will conclude in 2007 and 2008, but Rowinsky declined to elaborate.
For each of the additional Erbitux indications, ImClone is hoping to remove the EGFR IHC-test requirement, and possibly replace it with another biomarker test, or perhaps nothing at all. "Immunohistochemistry as a test is a terrible test — every laboratory has its own procedure," said Rowinsky. "As a test it just doesn't stand up — it can't be quantitated very well, nor qualified," he said. "We need to look for other predictors, and we're doing that in several trials with our partner Bristol-Myers [Squibb]."
In one of the trials with BMS, researchers explored approximately 40 genes as candidates for biomarkers of Erbitux response using gene-expression microarrays, said Rowinsky.
Immunohistochemistry testing for EGFR does not allow for easy comparison between different testing locations, due to differences in preservation, storage, method, and other variables, agreed William Pao, a researcher at Memorial Sloan-Kettering Cancer Center, who has worked extensively on the EGFR pathway. Problems with the quantitation of IHC signals often spring from its interpretation — the scale used to evaluate signals can differ among facilities, as can a pathologist's perception, he said.
In the case of Iressa, "there hasn't been any correlation between protein expression and response or survival, so far," said Pao.
There are at least two problems ImClone has with the use of EGFR IHC diagnostics — inter-laboratory comparability and the actual usefulness of EGFR. Regarding comparability, the company is talking informally with the FDA about finding a way to resolve an Erbitux post-approval commitment to test EGFR-negative cancer patients, Rowinsky said. Using a standardized central lab to confirm results from regional labs, about "40 to 60 percent" of samples turn out to be false negatives, he said.
The company is in discussions with the FDA to remove EGFR testing from Erbitux's label because it doesn't live up to high standards either, said Rowinsky. About 25 percent of patients with EGFR-undetectable disease in lung cancer and in colon cancer respond to Erbitux, he said. "The bottom line is that we don't have a good test to predict the functionality of EGFR in cancer — that is, 'What is it that makes EGFR important in some tumors as a target, and not in others?'"
A possible outcome is that ImClone might continue to use EGFR as an indicator of Erbitux efficacy, albeit perhaps not expression level, said Pao. "It doesn't make sense if your tumor doesn't express it, for you to use [the drug]," he said. "I don't know what the data is in head-and-neck [tumors] — for lung cancer, at least for Iressa, there does not appear to be any correlation between protein expression [level] and response," said Pao.
For "most of the epithelial tumors, like lung, and ovarian, et cetera," EGFR expression is probably occurring, even though it may not be detectable by IHC, Pao said. "But I'd be hesitant to use [Erbitux] unless there were studies showing it was beneficial in a certain amount of patients," he said.
At present, there is not a better biomarker for Erbitux response, Pao said. However, a May 2005 study printed in The Lancet by an Italian group searching for response biomarkers in colon cancer identified gene amplification of EGFR as the best correlate, he said. The test "could be a readily available test, because we do that testing in breast cancer — for Her2, you look for gene amplification," he said.
Also, data show that biomarkers for lack of response may be on the horizon. The data is still incomplete, but recent data suggest that lung-cancer patients with certain EGFR mutations do not respond to Erbitux, said Pao.
Dako Cytomation, the Netherlands-based manufacturer of the most widely used IHC EGFR test, also acknowledges a change might be necessary. "We are aware that EGFR as a marker in IHC may not be optimal in cancer forms like colon, head & neck, and lung," said a Dako spokesperson in an e-mail to Pharmacogenomics Reporter. The company believes the EGFR IHC test might remain useful in other cancers or in combination with drugs, said the spokesperson. The company is also "working with an assay for detection of the amplified gene," and is "collaborating with the pharmaceutical industry regarding various markers for drugs" to use in "the development of new pharmacodiagnostic products," she said.
— Chris Womack ([email protected])