McClatchy-Tribune Information Services -- Unrestricted
June 3, 2008
Dr. Matthew Ellis stands at the dawn of personalized medicine, his work heralding a day when breast cancer patients will receive just the right treatment for their specific type of disease.
Ellis, a researcher at Washington University School of Medicine and co-founder of St. Louis-based University Genomics Inc., is developing genetic tests for analyzing individual tumors. The tests could predict patients' responses to treatment options.
Some would be spared the harsh side effects and cost of chemotherapy because it is not warranted, while others with aggressive tumors would quickly be given experimental treatments.
"That comes from really understanding what the patient has, what their tumor's biology is," Ellis said. "You're individually tailoring your approach to their exact problem."
It may be years until that is the routine course of cancer treatment. But work done by Ellis and his colleagues is shedding new light in a direction that experts say oncology is destined to go.
"This will only go forward. ... The entire field of cancer diagnosis and treatment is moving" toward it, said Dr. Mark Clanton, chief medical officer of the American Cancer Society's High Plains Division, which includes Missouri.
Ellis and his team presented their research Monday at the American Society of Clinical Oncology's annual meeting in Chicago. ARUP Laboratories, a Salt Lake City-based company that is partnering with University Genomics to commercialize its work, passed out information there on a test it hopes to bring to market this fall.
The test, called the Bioclassifier, measures 55 genetic indicators in a tumor tissue sample. It uses a mathematical algorithm to parse the results and divide patients into different buckets of disease -- such as the stage of tumor development, or whether female sex hormone estrogen will make it grow -- that indicate the best treatment options. It works for all types and stages of breast cancer, Ellis said.
ARUP and University Genomics say the Bioclassifier is more versatile and specific than two competing technologies on the market: Oncotype DX, made by Genomic Health Inc. of Redwood City, Calif.; and MammaPrint made by Dutch firm Agendia.
Experts say that would be a good thing. Because, while Oncotype and MammaPrint provide interesting information, they are not accurate enough to influence treatment decisions.
The tests predict whether a woman is at low risk of having her cancer return after surgical removal, radiation treatment, or both. If so, she can be spared postoperative chemotherapy.
But "imagine that if the test was wrong one out of six times, and the cancer came back, how that woman, her family and her doctor would feel about her decision not to take chemotherapy. That is essentially the situation that is facing us with the tests we have available today," said a blog item written in February by Dr. J. Leonard Lichtenfeld, deputy chief medical officer for the national office of the American Cancer Society.
The Food and Drug Administration is developing new guidelines for approval of genetic tests, recognizing that if they influence the use of drugs and course of cancer treatment, they require rigorous review. Tests that pass are more likely to be covered by public and private insurers -- and to properly win physicians' confidence, Clanton said.
ARUP will bring the Bioclassifier to market using a "home brew" regulation that allows a single laboratory to offer a test that it conducts, Ellis said. But it requires full FDA approval before University Genomics can begin its work: licensing the test to corporations that operate large numbers of diagnostic labs.
University Genomics plans to sell test kits to the labs. But the samples will be analyzed by Ellis and his partners, Charles Perou at the University of North Carolina and Phillip Bernard at the University of Utah. They will control the algorithm software and return test results through a secure website.
"The market itself is huge. ... Anyone who has breast cancer should take this test to determine how to be treated," said Rosana Ellis, Matthew's wife and the company's business consultant. "But we don't know what percentage we can capture yet. And technology moves so quickly, you never know who's going to come around the corner with something new."
But Ellis and his partners are not going to rest with a single test. They envision developing a series of increasingly specific prognostics, capable of further narrowing expected outcomes to certain types of tumors.
Already, Ellis has made a finding that he will present at the Clinical Oncology conference and that could underpin University Genomics' next test.
He found a unique genetic signature that unmasks a type of high-risk tumor that on the surface appears to be a low-risk variety -- estrogen-receptor positive cancer, which depends on estrogen to grow and is easily treated in postmenopausal women with drugs that inhibit the body's production of the hormone.
Doctors use chemical tests to spot the estrogen receptors and follow with hormone-inhibiting therapy in addition to operating to remove the tumor. But in certain cases, the cancer comes back and is aggressive -- and when then tested proves to be estrogen-receptor negative.
In a study of 56 women who appeared to have estrogen-receptor positive tumors, Ellis' test was able to identify 15 percent of patients whose cancers were estrogen-receptor negative -- wolves in sheep's clothing.
He believes the test will, for the first time, enable doctors to correctly identify this type of virulent cancer and treat it more effectively. Today, most estrogen-receptor positive patients are given the hormone-inhibiting drug as well as postoperative chemotherapy, just in case they fall into the high-risk group. If Ellis' test works, they would be spared chemotherapy.
But a much bigger clinical study is needed before it can be put into practice.
"Like all advances in medicine, it answers some questions but raises others that you need to answer," Ellis said.
Dr. W. Fraser Symmans, professor of pathology at the M.D. Anderson Cancer Center at the University of Texas, said Ellis' approach "is a work in progress that has a lot going for it."
"There is clearly a need for higher-quality diagnostic information for treatment planning, no question about it. And his sort of approach, and others, will get us there," Symmans said.
Eventually, experts say, doctors will have an array of genetic cancer tests they can apply to every tumor -- much as a single blood sample now can be run through a variety of tests to provide an accurate diagnosis of well-understood diseases. University Genomics hopes to provide several pieces to that puzzle.
"There is going to be a flood of technologies. ... At some point, it is expected that virtually any woman with breast cancer, no matter what her profile, will benefit from these kinds of tests," Clanton said. "We are at the dawn of targeted identification and treatment of cancer."
Copyright (C) 2008 St. Louis Post-Dispatch
