Journal of the National Cancer Institute Advance Access originally published online on July 29, 2008
JNCI Journal of the National Cancer Institute 2008 100(15):1050-1051; doi:10.1093/jnci/djn286
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
© Oxford University Press 2008.
NEWS |
Gene Testing To Tailor Breast Cancer Therapy Has Arrived—Is It Ready for the Clinic?
To genotype or not? The question is not as profound as Hamlet's, but for oncologists who treat breast cancer, the answer is almost as open to interpretation.
Some studies suggest that women with variant forms of the cytochrome P450 2D6 (CYP2D6) gene may not respond well to tamoxifen—the standard therapy in postmenopausal women with estrogen receptor–positive breast cancer. Other studies, however, have found no such relationship.
In Europe, the conflicting results have prompted caution. CYP2D6 genotyping is not yet offered to patients there. In the U.S., oncologists are split on the issue. Some already offer selected patients the gene test, whereas others think that the science is too premature.
"This falls into the category of legitimate differences in opinion," said Daniel F. Hayes, M.D., clinical director of the breast oncology program at the University of Michigan Comprehensive Cancer Center in Ann Arbor. Hayes is a member of the Consortium on Breast Cancer Pharmacogenomics, which recently authored an editorial (JNCI 2008;100:610–3) that recommended against routine CYP2D6 genotyping.
Evidence
The CYP2D6 gene codes for an enzyme that breaks down tamoxifen into several metabolites with antiestrogenic effects, the most active of which is endoxifen. Between 7% and 10% of the white population, however, carries variant CYP2D6 alleles that code for an impaired enzyme, the most relevant of which is called the star-4 (*4) allele. People who have two copies of *4 metabolize certain drugs, including tamoxifen, poorly. Those with one normal allele and one *4 allele metabolize the drug at an intermediate rate, and women with two normal alleles metabolize the drug quickly.
Preclinical research had long suggested an association between CYP2D6 genotype and response to tamoxifen. But studying the question in patients was difficult because scientists thought that they could obtain an accurate genotype only from blood samples. Few cancer clinical trials, however, store blood samples from patients. Then in 2003, researchers found that stored tumor samples yielded results as accurate as those obtained from blood samples. That advance allowed retrospective studies using stored tumor samples to look for associations between CYP2D6 alleles and response to tamoxifen.
The first study to look for this association was published in 2005 in the Journal of Clinical Oncology by Matthew Goetz, M.D., of the Mayo Clinic in Rochester, Minn., and James M. Rae, Ph.D., of the University of Michigan Medical Center in Ann Arbor and their colleagues. The researchers used tumor samples from a trial conducted by the North Central Cancer Treatment Group, which included almost 200 women with estrogen receptor–positive breast cancers who took tamoxifen for 5 years. The team found that poor metabolizers (with two *4 alleles) tended to have a higher risk of breast cancer relapse than that of extensive metabolizers (those with two normal alleles). A follow-up study using the same data found that poor metabolizers had a statistically significantly higher risk of relapse or death. The study controlled for women who were taking drugs that inhibit CYP2D6 activity.
"[Goetz's] paper kicked off a firestorm," Hayes said. "People looked at that and said, ‘Jeez, it looks like 10% of our patients may be taking an inactive drug.’" Since then, researchers have published five more studies on this subject. Two have been consistent with Goetz's initial findings. One found no difference between poor and extensive metabolizers, and two others concluded the opposite—that poor metabolizers responded better to tamoxifen than did other women.
The conflicting results drew criticism on study design. There were several flaws, critics argued: The criteria for including patients differed among studies; some patients might have concurrently been taking other drugs that interfere with CYP2D6 activity; most studies were retrospective; and patients received different doses of tamoxifen for different durations.
"That doesn't mean the studies were bad," Hayes said. "It just means that because of the way the studies were done, there are inherent biases." In retrospective studies, for example, narrowly defining a patient population is difficult because crucial information is sometimes missing. For CYP2D6, not all studies used tumor samples with detailed medical records, so comparing patient groups across studies was challenging. More important, the investigators couldn't be sure that all patients complied with drug regimens. Because fast metabolizers experience more side effects, they may not have taken tamoxifen as diligently as poor metabolizers, which would have skewed results. Also, methods to test tumor samples for the presence of estrogen receptors weren't consistent—a factor that Hayes finds especially troubling.
"There haven't been clear guidelines on how tumor marker studies should be done," Hayes said. "Most tumor marker studies are studies of convenience. You pull some samples out of the freezer, you run an assay, get a [statistically significant result], and you rush it off to be published. That's exactly what's going on with 2D6."
In the Clinic
Despite conflicting data, some oncologists already offer the test to patients. For example, shortly after Goetz and his colleagues presented their data at the 2006 annual meeting of the American Society of Clinical Oncology, the Mayo Clinic began offering the test to postmenopausal women with estrogen receptor–positive breast cancer who had chosen to take tamoxifen.
"We present these women with an option," said James N. Ingle, M.D., professor of oncology at Mayo Clinic in Minnesota. Most oncologists treat postmenopausal breast cancer patients with tamoxifen or aromatase inhibitors, another class of antiestrogenic drugs. At Mayo, oncologists explain the pros and cons of each drug and allow women to choose. If patients choose tamoxifen, the oncologists also recommend that they take the gene test. If a woman turns out to be a poor metabolizer, doctors recommend aromatase inhibitors instead of tamoxifen. "We have confidence in our data," Ingle said.
Oncologists at the Mayo Clinic aren't the only ones who are confident that CYP2D6 genotype influences drug response or cancer recurrence. In 2006, an advisory committee to the U.S. Food and Drug Administration recommended that tamoxifen's label be changed to include information about genetic factors and drug interactions that may affect the efficacy of the drug. They also recommended that the label state that CYP2D6 genotype testing is available.
Others agree that the data are convincing. "This isn't just a simple association," said Hiltrud B. Brauch, Ph.D., coordinator of the oncology program at Margarete Fischer-Bosch Institute of Clinical Pharmacology in Stuttgart, Germany. "This association is in strong agreement with the biological phenotype." CYP2D6-impaired genotypes affect tamoxifen metabolite levels, and "this is where we take our confidence from," Brauch said.
Brauch's group published a study that supported Goetz's findings in 2007 in the Journal of Clinical Oncology. But although she is confident that ongoing research will eventually support routine CYP2D6 testing, she is more circumspect about implementing it now. After her report appeared, German clinicians flooded her with requests for the gene test. But she refused, she said, because evidence is still lacking. Although studies have associated genotype with clinical outcome and genotype with tamoxifen metabolite levels, no study has yet linked genotype to metabolite levels to clinical outcome all within the same patient population.
"Until we have unanimous research results, we're reluctant to spell out any recommendations for clinicians," she said.
Conflicting data aren't the only reason that some scientists hesitate to introduce gene testing. Research suggests that there's more to tamoxifen metabolism than just CYP2D6. Endoxifen levels in intermediate metabolizers, for example, vary widely, suggesting that other factors also influence response to tamoxifen.
Tumor heterogeneity is one explanation. Estrogen receptor–positive tumors are not all the same and thus may respond to the drug differently. Other genes may also be involved. Also, CYP2D6 isn't the only enzyme that metabolizes tamoxifen. Other cytochrome enzymes act on the drug to create additional metabolites, some with antiestrogenic activity (although endoxifen is the most potent). From a biological perspective, "it's not clear that complete loss of 2D6 activity should necessarily result in a worse outcome," Hayes said.
Such blips in the data make European oncologists, such as Vincent O. Dezentjé, M.D., of Leiden University in The Netherlands, even more hesitant to use gene testing. "I think it's too early to say that there's even an association between CYP2D6, tamoxifen metabolism, and clinical outcome," he said. "The studies that support testing were done retrospectively and on small groups of patients. The studies that conflicted with these were criticized. But these studies surely do have some value, and they can't be discarded."
Moreover, he added, there's no evidence to suggest that a poor responder would necessarily do better on an aromatase inhibitor. Aromatase inhibitors haven't been expressly studied in poor responders. But studies in postmenopausal breast cancer patients do suggest that the drugs are certainly no worse than tamoxifen. In fact, they may even be more effective. Studies comparing the two drugs found that women taking aromatase inhibitors have between a 2 and 5 percentage point improvement in disease-free survival over women on tamoxifen. The findings prompted the American Society of Clinical Oncology to recommend recently that all postmenopausal breast cancer patients take an aromatase inhibitor at some point during their treatment. Because the CYP2D6 enzyme does not act on aromatase inhibitors, there's no reason to think that a poor responder would do worse on this class of drugs, Ingle said.
But these data aren't enough to convince skeptics that postmenopausal poor responders should be shifted to aromatase inhibitors. "If for whatever reasons tamoxifen is better for a given patient, and if the 2D6 story turns out not to be true, then that patient will be placed on a less optimal therapy," Hayes said. "We are just too early in this story to use this test routinely to make clinical decisions."
More Data on the Way
Meanwhile, several groups have come together to form the International Tamoxifen Pharmacogenomics Consortium to clear up the discord. The group's main purpose is to gather data from around the world on the clinical efficacy and toxicity of tamoxifen in women with various cytochrome genotypes. "We are trying to gather as much retrospective data as possible," said Brauch, who is a member. In the process, they hope to standardize study design so that retrospective studies on stored tumor samples can be compared. The members of the Consortium on Breast Cancer Pharmacogenomics also plan to conduct more retrospective studies.
Although experts agree that prospective studies would yield better data, many express apprehension about the cost and the follow-up time required. Using retrospective studies, Brauch estimated that the consortium's collaborative efforts might be able to gather enough data to publish within a year or two.
Nevertheless, researchers at Leiden University have just launched a prospective study that will include 650 women and genotype all cytochrome alleles, not just *4. Dezentjé said the study will test endoxifen levels and statistically correlate them to clinical outcome after 4 years. The study will try to clarify whether poor and intermediate metabolizers might benefit from receiving higher doses of tamoxifen.
Until the results of these and other studies are available, Hayes suggests that oncologists wait for stronger data. "This field is in its infancy, and we have to be careful about it."
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||