Push is on to tailor cancer care to tumor's genes

Everyone with advanced colon cancer now is supposed to get a genetic test before taking two of the leading treatments. It's a major change adopted by oncologists last month after studies found that those pricey drugs, Erbitux and Vectibix, won't work in 40 percent of patients.

Scientists are furiously testing similar genetically tailored care in breast and lung cancer. It's a flurry of work that reflects a huge problem: Most medications today benefit at best about half of patients but it usually takes trial-and-error to tell.

That means a lot of people suffer side effects for nothing, and it's incredibly costly. When the American Society of Clinical Oncology recommended giving colon cancer patients that $300 test for a gene called KRAS, it estimated the move could save a stunning $600 million a year — by keeping drugs that cost up to $10,000 a month away from patients who won't benefit.

Here's the critical consumer issue: As tantalizing as this personalized medicine is, gene testing is like the Wild West. Laboratories often introduce new tests at the first clues they might work, not waiting for final proof. Few tests so far have won the backing of major medical groups like ASCO, the cancer specialists, making research studies a best bet for many patients.

"A bad test is as dangerous to a patient as a bad drug," notes Dr. Richard Schilsky, ASCO president and a University of Chicago oncologist. "The tricky part is to figure out which of those (genetic differences) are clinically important and which are just variations that exist."

This is not about testing if people carry so-called cancer genes that make them prone to illness. Instead it's about finding a tumor's genetic signature — a pattern of gene and protein activity that signals if the cancer will grow fast or slowly, be more or less likely to recur, and whether it would be susceptible to treatment.

"We're getting into science fiction sort of, if now medicine is being able to analyze things at the genome level," breast cancer patient Claire Weinberg of Oxford, N.C., says in wonder.

A community hospital initially dismissed Weinberg's breast lump but she fortunately sought a second opinion at Duke University Medical Center — where, cancer confirmed, she enrolled in a study of gene-directed chemotherapy.

"I felt it could only benefit me for them to know even more about me," she says.

The ultimate goal: "What's the right recipe for those patients?" explains Dr. Matthew Ellis of Washington University in St. Louis, co-inventor of a different breast cancer genetic approach.

Under study:

A less precise test already can tell certain breast cancer patients if they're at high or low risk of relapsing, helping the chemo-or-not decision. But which chemo? Duke's Dr. Kelly Marcom is genetically profiling breast biopsy tissue from nearly 300 newly diagnosed patients headed for pre-surgery chemo. Some are randomly assigned to one of two standard chemotherapy cockails; the rest get the cocktail that matches their tumor profile.

It's too early to tell if the gene-directed approach helps more tumors shrink.

But, "I can have no regrets," says Weinberg, who learned after surgery that she'd been in the gene-tailored group and her tumor shrank enough to save her breast. She's also getting post-surgery chemo in case any rogue cells remain.

Instead of custom profiling, an experimental test unveiled last week examines 50 breast cancer genes to determine which of four disease subtypes the woman has.

If it pans out — and much larger studies are planned — the Breast Bioclassifier could change breast cancer's very names. When studied on stored samples of old tumors, researchers found some women safely skipped chemo — their subtype responded better to post-surgery tamoxifen, or hormone therapy. A more aggressive type was sensitive to most chemo choices but not hormone treatment, the team reported in the Journal of Clinical Oncology.

And still another group didn't respond well to either care, a group that desperately needs new options, said Ellis, who co-developed the test with doctors at the University of Utah and University of North Carolina, Chapel Hill.

Next up, lung cancer. Hospitals nationwide are recruiting 1,200 lung cancer patients to study who carries extra copies of the tumor-spurring gene EGFR. They'll get either of two top treatments, Tarceva or Alimta, to see which is best for which genetic condition.