Like many oncologists, Eric P. Lester, M.D., was faced with a dilemma: seven patients with advanced, incurable cancer, an arsenal of drugs that may or may not help them, and not enough solid proof about treatment efficacy to guide him.
So Dr. Lester devised what he called a "simple-minded experiment" that illustrates the promise of personalized medicine. Using DNA microarray "chips," Dr. Lester analyzed his patients' tumors for expression of genes associated with good response to various anti-cancer drugs, and based his drug treatment plans on the results. Four out of seven patients with advanced cancer enrolled in the extremely limited study had a better outcome than expected.
The finding shows that "a personalized molecular oncology approach, basing chemotherapy on relative gene expression in tumors, holds promise even at the relatively crude level employed here," said study investigator, Dr. Lester, president of Oncology Care Associates in St. Joseph, Mich.
To obtain and analyze chip data, Dr. Lester worked with Craig Webb, Ph.D., Director of Translational Medicine at the Van Andel Research Institute in Grand Rapids, Mich.
The study is unusual because oncologists don't yet base most of their treatment decisions on gene profiling, especially when it might involve pairing drugs together in a novel combination or using varied doses, Dr. Lester said. "Much of clinical medicine is an educated guess, and this was an attempt to come up with a better approach by using the technology of a gene chip to make multiple, highly educated guesses simultaneously," Dr. Lester said.
Dr. Lester added that one of the seven participating patients died before the gene chip was used to direct therapy.
Many current clinical trials involving gene expression examine effectiveness markers for individual drugs rather than combinations of drugs or different doses of agents used together for the first time. To truly help the most patients, Dr. Lester said, all potentially effective drugs and combinations must be matched up against the unique genetic profile of a patient's tumor, he said.
"Effective cancer treatment depends on understanding the biology driving the cancer, but because each tumor is different, it is very hard to personalize care and do a rigorous scientific experiment at the same time."
In this study, Dr. Lester said he "stayed within the envelope of a reasonable standard of care" in treating his patients. That standard is often based on what insurance companies will typically reimburse for treatment given published studies about the effectiveness of a drug on a certain tumor type, and whether or not the drug is federally approved for that indication. Dr. Lester and Webb surveyed the scientific literature and compiled a list of genes whose expression levels may predict response to a drug given the tumor type.
In some cases, treatment strategies suggested by the chips varied significantly even for the same type of cancer. For example, one patient whose lung cancer had spread to his brain and bones achieved a "near complete response" when treated with two chemotherapy drugs, in addition to Tarceva and Avastin, while another lung cancer patient responded to third-line drugs such as etoposide.
Acknowledging the risk involved with using novel combinations of drugs where no set safety profile exists, Dr. Lester said that "this is constantly done in medicine. People are taking antibiotics at the same time as using heart and cholesterol pills, and blood pressure medication."
"This kind of polypharmacy will become more common in cancer, but at the moment, it is hard to figure out the difference between doses that are effective or that could be toxic," he said.
The best way to get around such issues is to build a database of gene expression data and match them with patient outcomes, he said. "Now when I see new patients I am itching to look at what the genes can tell me," Dr. Lester said. "It is a smarter way to treat cancer."
This finding was presented in Atlanta, Ga. at the American Association for Cancer Research's second International Conference on Molecular Diagnostics in Cancer Therapeutic Development.
Cite This Page: