Last checked at 
By Julie Steenhuysen
CHICAGO, June 2 (Reuters) - The medical view of cancer is intransition, as cancer doctors increasingly focus on thedefective genes that are driving the disease rather than theorgan in which it takes root.
Oncologists hope that by understanding the geneticunderpinnings of cancer rather than focusing on whether itoriginated in the breast or the liver, they will be able to givepatients better, more personalized and more effectivetreatments.
But leading cancer experts say that as doctors have testedthat theory more closely, they are seeing both successes andfailures, suggesting that the ideal route to treating cancer may be far more complicated than hoped.
At the moment, there is not enough understanding of themutations that drive cancer growth, said Dr. Richard Pazdur,chief of oncology at the U.S. Food and Drug Administration
"What people want and the scientific reality are twodifferent situations," Pazdur said in an interview at theAmerican Society of Clinical Oncology (ASCO) meeting in Chicago.
He said there are currently "only a handful of therapies"that target specific cancer-causing genes. Some oncologists have begun using those drugs to treat cancers in people with thecorresponding genetic mutations, even when the drugs haven'tbeen approved for the type of cancer a patient has.
When the drugs work, the results are dramatic. Tumors seemto melt away, and advances in survival are counted in months andyears, not weeks. The transformation has led many doctors toquestion whether the current practice of approving drugs basedon the organ they target needs to change.
But there have also been disappointments. Drugs made byRoche and GlaxoSmithKline that target cancerswith a mutation known as BRAF can have a powerful, albeittransient, effect in beating back the deadly skin cancermelanoma. These drugs do not work, however, in patients withcolon cancers driven by the same BRAF mutation, according to a2012 study that began to sow doubt among some experts.
"The temptation in practice is to get the tumor sequenced,and if something comes up for which there is an approved drug,your clinician is tempted to try that, no matter what the tumoris. But it might not be that simple," said Dr. Barbara Conley ofthe National Cancer Institute.
As it turns out, colon cancers with BRAF mutations haveanother driver mutation called EGFR. In those tumors, doctorsmay need to use a combination of agents that hit both targets, ,according to new data presented at ASCO this weekend.
Dr. Bert Vogelstein, a cancer geneticist at Johns HopkinsUniversity in Baltimore, said there is scant evidence thatmatching a patient's mutation to a targeted drug improves care.
"It's definitely an unproven assumption and we ought to testit," Vogelstein said.
JUMP IN GENETIC PROFILING
Oncologists interviewed by Reuters said genetic profiling oftumors is becoming much more routine, particularly if a patientwith advanced disease hasn't recovered with more traditionalapproaches. Foundation Medicine Inc, which makes geneticprofiling tests, has seen clinical demand for its tests jump 67percent in the first quarter from the same period a year ago.
Dr. Richard Schilsky, chief medical officer of ASCO and aUniversity of Chicago Medicine oncologist, estimates that in upto 70 percent of such tests, doctors find a genetic mutationthat can be matched with a drug. Insurers, however, aren't quickto cover a treatment for an unapproved use unless there isevidence that it will work.
"They are experimenting. It's not like these therapies don'thave harms as well. They have side effects," said JenniferMalin, medical director for cancer drugs at health insurerAnthem Inc.
The issue will be studied more broadly in a large cancerclinical trial launched by the NCI. It is designed to match theunderlying genetic defect driving a person's tumor with one ormore of 20 approved or experimental drugs targeting thatgene.
ASCO is also starting a clinical trial to gather data on howpatients fare when doctors order genetic profiles and use thedata to influence treatment. The TAPUR trial will allow thesepatients to be treated with one of five FDA-approved drugsprovided for free, and their data will be collected to helpanswer this question.
Early results from a small Foundation Medicine-sponsoredtrial presented on Sunday showed a limited benefit. Researchersat University of Texas MD Anderson Cancer Center tested 339patients with very advanced cancers. For 122 patients, the teamwas able to match the genetic defects in their tumors to a drugor combination of drugs targeting those defects. They alsofollowed 66 patients whose cancers were not matched to atargeted drug.
The patients whose tumors were matched with a targeted druglived 2 months longer than the unmatched patients. Dr. JenniferWheler of MD Anderson, who led the study, thinks the outcomemight improve if the approach were tried in healthier patients.
But it will take large clinical studies to prove thehypothesis that picking a patient's therapy based on themutation in his or her tumor works. "It's like everything else.The devil is in the details," Vogelstein said. (Editing by Michele Gershberg and Sue Horton)
