Cancer Fight: Unclear Tests for New Drug

Cancer Fight: Unclear Tests for New Drug
By GINA KOLATA
Copyright by The New York Times
Published: April 19, 2010
http://www.nytimes.com/2010/04/20/health/research/20cancer.html?th&emc=th


Dr. Linda Griffith was at a conference in Singapore in early January when she felt a lump in her breast. She assumed it was nothing — a cyst. And anyway, she had no time for it. She was returning on a Sunday night and the next Tuesday morning was leaving for a conference in Florida.

Dr. Linda Griffith has faced cancer as a doctor and as a patient. It was unclear whether she should receive “targeted therapy.”

But she had a mammogram, ultrasound and biopsy within hours of getting off the plane. The news was not good: she had cancer.

Then the complications began. Dr. Griffith, director of the Center for Gynepathology Research at M.I.T., had a test to see whether her tumor had extra copies of a protein, HER2. If it did, it would respond to a drug, Herceptin, which blocks the protein and stymies the tumor’s growth.

Drugs aimed at disabling proteins that spur cancer are, many oncologists say, the future of cancer therapies. Only a few are available now but almost every new drug under study is designed to disable cancer-fueling proteins.

But these so-called targeted therapies are only as good as tests to find their protein targets. And while most patients do not yet know it, those tests can be surprisingly unreliable.

Acknowledging the problem, cancer specialists on Monday announced new testing guidelines for one protein target, but as new targets are identified, the problem continues to grow.

The test on Dr. Griffith’s tumor was negative. Or was it? One small area of her tumor stained chocolate brown, indicating lots of HER2. The rest was a cream color, indicating no extra HER2 protein.

Yet her treatment hinged on this result. A HER2 positive tumor has a bad prognosis. Herceptin can make that prognosis good, reducing the chances that the cancer will come back by 50 percent and reducing a woman’s risk of dying by 40 percent.

But Herceptin, costing $42,000 a year wholesale, causes flulike symptoms and also has a rare, serious side effect, severe heart damage that can even be fatal.

And if a tumor does not have high levels of HER2, Herceptin would be, as Dr. Antonio Wolff, a breast cancer specialist at Johns Hopkins put it, “a toxic and expensive placebo.”

Dr. Griffith had come face to face with an emerging, but rarely acknowledged, problem in today’s era of new cancer tests and therapies.

HER2 tests, for instance, can give false-positives up to 20 percent of the time, wrongly telling women they need the drug when they do not. Five percent to 10 percent of the time the tests can falsely tell a woman that she should not take the drug, when she should. And Herceptin testing for breast cancer is easy compared with what is coming next.

Genentech, Herceptin’s maker, is about to apply to the Food and Drug Administration to sell the drug to treat stomach cancer. But it is much more difficult to tell whether a stomach tumor has high levels of HER2, said Krysta Pellegrino, a company spokeswoman. Breast cancers usually are all positive or all negative. Not stomach cancers, which almost always have sections that are positive for HER2 and sections that are negative. The HER2 tests are the same, but “the interpretation and scoring are different,” Ms. Pellegrino said.

That sort of mosaic pattern is typical of cancers other than breast cancer, says Dr. Jeffrey Bloss, vice president, North America Medical Affairs at GlaxoSmithKline. And it raises questions of what a test result means.

“The science is still evolving,” Dr. Bloss said. “What was true last year may not be true this year.”

Like the HER2 tests, other molecular tests for breast cancer also have problems. Those tests, for estrogen receptors on breast cancer cells, determine whether cancer will be thwarted by drugs that deprive tumors of estrogen. They can be wrong at least 10 percent of the time. Some estrogen-depleting drugs, while generally safe, increase the risk of osteoporosis and, depending on the drug, can also cause joint pain and increase risks of stroke and cancer of the uterine lining.

Estrogen receptor tests are a muddle, noted Dr. Edith Perez, a breast cancer specialist at the Mayo Clinic in Jacksonville, Fla. Quite a few tests are being used, but Dr. Perez could not ascertain exactly how many or how good they were in predicting whether a tumor would respond to estrogen-depleting drugs.

And different labs may do tests in different ways; some even invent their own.

“How do you know they are the same?” Dr. Perez asked. “If you do the test in two different labs, you can get two different answers.”

Error rates for newer tests have not even been established.

“This is an issue that transcends breast cancer,” Dr. Wolff said. “A poorly developed test is potentially as dangerous as a poorly developed drug.”

The Food and Drug Administration says it is concerned about the quality of tests developed by clinical laboratories for their own use, said Alberto Gutierrez, who oversees diagnostic products for the agency. Some of the tests are increasingly complex, Mr. Gutierrez said, adding that there is a proliferation of laboratories offering tests without F.D.A. oversight. But, for now, the agency has no specific plan to regulate the tests, in part because of lack of money.

Meanwhile, Dr. Griffith’s doctor, Eric Winer at the Dana-Farber Cancer Institute, had a gradual awakening.

“In my naïve view, which wasn’t so many years ago,” Dr. Winer said, “I thought HER2 was a switch that turns on or off and it was pretty easy to tell when it’s on or off. It turns out that it is not nearly as straightforward for a large number of tumors.”

Now, recognizing the problem, Dr. Winer had Dr. Griffith’s tumor retested with a different method, hoping the result would help him and Dr. Griffith figure out whether she could benefit from Herceptin.

And Dr. Griffith was left facing the uncertainties of cancer medicine.

“Me as a scientist says it’s very interesting,” she said.

But, she said, as a patient she sees it differently.

“It’s really hard to know what to do,” Dr. Griffith said.

The two large national studies of Herceptin for women with HER2 positive early-stage breast cancer were just starting in 2001 when Dr. Perez, of the Mayo Clinic, a principal investigator, had a moment of truth. Women were having HER2 tests at a variety of places — community hospitals, major medical centers, national labs. Dr. Perez decided to retest tumors in a central lab to confirm the results.

The outcome stunned her and her colleagues. Twenty percent of the first 119 women whose initial tests indicated their tumors had excess HER2 turned out not to have it on retesting.

“We all felt, ‘Oh boy, we have a problem,’ ” said Dr. Wolff, a study investigator. “This was huge.”

So the studies were modified to require central labs to retest all the tumors.

Yet the discordance remained — one-sixth of women told by local labs that they were HER2 positive were not on retesting.

“We were all horrified,” said Dr. Elizabeth Hammond, a pathologist at Intermountain HealthCare in Utah.

The result of that moment of horror was HER2 testing guidelines by the College of American Pathology and the American Society of Clinical Oncology, dictating criteria for declaring a test positive or negative and requiring proficiency testing, among other things.

In a way, the effort was a huge success. About 900 of the nation’s estimated 1,500 labs agreed to follow the guidelines.

But even so, said Dr. Bloss of GlaxoSmithKline, there seemed to be approximately a 20 percent discordance between labs. GlaxoSmithKline makes Tykerb, which also focuses on HER2.

There are all sorts of reasons why different labs can get different results, said Dr. Mitch Dowsett of the Royal Marsden Hospital in London and a member of the United States committee that formulated HER2 testing guidelines.

In borderline cases, pathologists can disagree. Or stain can pool in areas where a tumor was crushed or damaged, making it look, to inexperienced eyes, like a positive stain.

Twelve years after Herceptin was approved for women with advanced breast cancer, “we’re still trying to refine the testing,” said Ms. Pellegrino of Genentech.

Then there is Dr. Griffith’s problem: what to do when part of a tumor is positive and the rest is negative.

The College of American Pathologists wants to develop testing guidelines for every molecular target for cancer drugs. On Monday, for example, it released new guidelines for estrogen receptor testing.

And Dr. Hammond has become driven to make sure pathologists know about and follow the HER2 guidelines.

At pathology meetings, she asks her audience how many know about the guidelines and are following them.

“Almost everyone raises their hand,” Dr. Hammond said. “I am preaching to the choir. They chose to come to the meeting. It’s the ones who did not choose to come that I am worried about.”

But even the best labs can differ, as some women learned.

When Sheila Maloney had breast cancer surgery in October, her doctor wanted to test her tumor for HER2.

“I had never heard of it,” said Mrs. Maloney, a 64-year-old hostess at an Olive Garden restaurant in Lady Lake, Fla.

She is now seeing Dr. Perez, and ended up having her tumor tested four times with four different commonly used HER2 tests. The first test was positive, the second negative, the third positive, the fourth negative.

Dr. Perez recommended that Mrs. Maloney take Herceptin.

As for Dr. Griffith, the two tests for HER2 turned out to agree, but with that mixed result, it was hard to know what to do. Her tumor was on the fence — part negative, part weakly positive.

Medical experts say there are no easy answers. For now, their best advice is for women to ask that their breast cancer tissue be sent to experienced labs that follow accreditation procedures like those recommended by the College of American Pathologists.

But Dr. Griffith did all that. And Dr. Griffith, a scientist whose own research involves the HER2 protein, also read and examined the literature on HER2 to prepare for a discussion with Dr. Winer.

“Here I sit as a patient. My situation is ambiguous,” Dr. Griffith said.

In the end, the studies, along with Dr. Winer’s clinical perspective, did not convince her that the drug would help. The risk of serious heart damage and other side effects was scary. And, she said, she cannot ignore the drug’s price, even though her insurer would pay.

Dr. Griffith decided not to take Herceptin, but she is having standard chemotherapy.

“I am very comfortable with my decision,” she said.