Originally published June 6.
NEW YORK (GenomeWeb) – Two studies investigating different non-small cell lung cancer drugs – Genentech's MetMAb (onartuzumab) and Pfizer's Xalkori (crizotinib) – that target the same molecular marker, MET, came to different conclusions about whether aberrant signaling of the oncogene in patients' tumors may be a predictive marker for gauging best responders to these therapies.
NSCLC patients with tumors harboring amplified or over-expressed MET tend to have poorer prognoses than those without this marker, researchers have found. Additionally, studies show that MET can drive resistance to EGFR inhibitors, such as Tarceva.
However, these latest trials on MetMAb and Xalkori, presented at the American Society of Clinical Oncology's annual meeting this week, suggest a more complex gene-disease-drug relationship when it concerns MET. In particular, the contrasting studies illustrate that characterizing patients' responses to drugs according to molecular subtypes can be finicky business. A completely different picture can emerge based on whether researchers are measuring gene amplification or expression, if they're using flourescence in situ hybridization or immunohistochemistry, and how they're determining the cutoffs separating responders from non-responders.
Sub-optimal cutoff?
In the 500-patient Phase III METLung study, researchers led by David Spigel of Sarah Cannon Research Institute reported that the addition of Genentech's investigational MetMAb to the EGFR inhibitor Tarceva (erlotinib) doesn't extend survival in advanced NSCLC patients with MET-overexpressing tumors compared to single-agent Tarceva. Genentech's parent company Roche had previously announced the negative trial, but waited until the ASCO meeting to discuss detailed results.
In the study, patients receiving the MetMAb/Tarceva combination had a median overall survival of less than seven months, while those receiving Tarceva lived for a median of nine months; the hazard ratio was 1.27. Progression-free survival in the two arms was almost the same. "A lot was done to look for signals and subgroups [to see] if there was any evidence of some hint of efficacy," Spigel said while presenting the METLung data at the ASCO meeting. "Unfortunately, there really is no one subgroup where there is a suggestion of benefit."
MET and EGFR are often co-expressed and even though 29 patients in each arm had EGFR-mutated tumors, there didn't seem to be any differential advantage for this subset of patients. Adverse events in the study were tolerable, but there were more frequent events in the combination arm, such as peripheral edema, low levels of albumin in blood serum, back pain, dyspnea, nausea, acneiform dermatitis, and rash. "Several exploratory biomarker analyses are currently in progress," Spigel said. "A lot of tissue has been collected in this study, including serum."
Federico Cappuzzo from the Istituto Toscano Tumori Ospedale Civile Livorno, in reviewing the data from this study at the meeting, put forth that METLung likely failed due to suboptimal selection of patients. He noted that while the researchers' cutoff for MET positivity may have been sufficient to show a prognostic effect of the marker, it wasn't sufficient to demonstrate a predictive effect on treatment response.
A different finding by MET amplification
In contrast, Ross Camidge from the University of Colorado Cancer Center presented data from a much smaller Phase I study suggesting the opposite, that patients with high and intermediate MET amplification responded to Xalkori. Researchers stratified 14 NSCLC patients based on low (2 patients), intermediate (6 patients), and high MET amplification (6 patients), and reported one person in the intermediate-MET category had a partial response and four had stable disease. Meanwhile, one person in the high-MET category had a complete response, three had partial responses, and one had stable disease.
Xalkori is a MET, ROS1, and ALK inhibitor. Pfizer had initially planned to investigate the drug as a MET inhibitor, but after the association of ALK mutations in NSCLC came to light, the firm quickly brought Xalkori to market in 2011 as an ALK inhibitor. The drug is currently approved as a treatment for advanced NSCLC patients who have ALK rearrangements. Pfizer is continuing to investigate the agent in patients with ROS1- and MET-positive tumors.
Although this study was small, the objective response rates in the low, intermediate, and high cohorts were 0 percent, 17 percent, and 67 percent, respectively. Based on the findings of his group, Camidge concluded in contrast to the METLung researchers that that MET amplified disease "may represent a new subtype of non-small cell lung cancer." However, the predictive ability of this marker to pick out best responders to anti-MET agents will depend on the cutoff point for determining when a tumor is MET positive, he cautioned.
The way Camidge and colleagues established MET positivity in this trial differed from the METLung study, where Spigel et al. used IHC to determine levels of MET overexpression and included mostly patients who had IHC 2+ MET expression in 50 percent of tumor cells, while some had IHC 3+ tumors. In METLung, researchers also looked at how well patients responded to MetMAb/Tarceva according to their MET amplification status by FISH, but "there was no hint of an advantage in the group that [was] FISH MET amplified," Spigel said.
In the Phase I trial involving Xalkori, however, investigators measured MET amplification – the increase in copies of the MET wild-type gene – using a combination of different techniques. The amplification of MET can result in increased copy numbers of the entire chromosome 7 on which MET resides or increased copies of the region in the chromosome containing the gene. To account for this, Camidge explained that the researchers decided to use the ratio of MET and the CEP7 region of the chromosome to define MET positivity of tumors in the trial.
"Theoretically increases in the MET/CEP7 ratio are more likely to represent a true biological selection of MET by the cancer cell," he said. Using this method, Camidge and colleagues developed the cutoffs for determining MET-low, -intermediate, and -high groups in a study involving 800 patients. Then, in the Phase I trial, researchers prescreened patients for MET amplification by FISH, calculated their MET/CEP7 ratio, and stratified them to low, intermediate, and high cohorts.
While none of the patients in the low-MET category appeared to respond to the drug, several patients in the other categories have been on Xalkori for almost four years. A few of these long-term responders have not had disease progression, Camidge noted, "with the level of duration seeming to increase with MET amplification."
Cappuzzo noted that the data from another study presented at the ASCO meeting provides further evidence that mostly IHC 3+ MET overexpressers are also found to have very high MET amplification by FISH. However, IHC 3+ MET overexpressers are exceedingly rare, while those with IHC 2+ status are more common, which is why METLung enrolled more in the latter category than the former. Based on the failure of METLung and the positive findings in the Xalkori study by Camidge, Cappuzzo advised researchers using IHC to predict MET positivity in a drug trial, to confirm patients' amplification levels by FISH.
While some patients have discontinued the Xalkori trial due to disease progression, there haven't been any drop-offs due to drug toxicity. Patients commonly experienced gastrointestinal events and the other toxicities common to Xalkori.
A curious aspect of the trial was that a very high proportion of smokers or ex-smokers in the trial, 86 percent, had MET amplification. "The demographics in relation to smoking appear different than seen with other targetable genomic abnormalities," Camidge noted.
Data from cell lines – showing that only very high MET amplified cells are sensitive to anti-MET agents – corroborate the findings of the study reported by Camidge et al., Capuzzo said in reviewing the data. He further noted that his own research has found that smokers or former smokers tend to exhibit MET amplified tumors, some of which had squamous cell histology.
"This is a very important message," Cappuzzo told the audience at the meeting. "That means that [when] conducting a trial, if we're looking for patients with MET amplification, we should probably focus our attention on smoking patients irrespective of histology."
Although Camidge could not precisely say what proportion of NSCLC patients would exhibit sufficient MET amplification to respond to Xalkori, he admitted the marker is rare. ALK mutations are found in around 5 percent of NSCLC patients, while ROS1 rearrangements show up in around 1 percent of tumors. Cappuzzo estimated that the best case scenario would be if around 3 percent of NSCLC patients had MET amplified tumors that responded to anti-MET agents. "We actively encourage [researchers] to do MET screening now," Camidge said.