heading qaThe goal of personalized medicine is to match the genetics of an individual’s cancer with genetically-targeted treatments. In this model, between a tumor and its treatment is a test. CU’s Marileila Garcia designs the tests that show a tumor’s genetic problems, and for two decades has been leading the field of cancer care into this new era of personalized medicine. Here we talk with Dr. Garcia about what is sometimes the overlooked middle child of personalized medicine: the tests that match disease with treatment.
C3: When you think about cancer care, it’s easy to focus on causes and cures. How did you get interested in this bridge of testing that connects the two?

Garcia: In the 1990s we didn’t know about genetically targeted drugs. But we knew cancers had genetic differences and that patients with some tumor subtypes did better than others. Originally, we were looking at tumor genetics just for their prognostic value, or just as an adjunct to diagnosis.

C3: And how did prognosis move to treatment?

Garcia: In the early 2000s, we showed a high copy number of the gene EGFR in some breast and lung cancer patients, and we developed a test for EGFR copy number. Patients with amplification of EGFR had poor prognosis, that is, shorter survival than others – an abnormally high number of EGFR copies was driving the cancer. In 2003, the drug gefitinib was in clinical trials to treat lung cancer, and we also helped to show that EGFR copy number was predictive of response to the drug. Simultaneously, mutations in the EGFR gene were discovered and proved even more effective to select patients for gefitinib and then erlotinib, the first truly genetically targeted treatments in lung cancer.

C3: So now we test tumors for EGFR mutations?

Garcia: Not just EGFR, but many genetic mutations. The significant changes in the past ten years have been based in three major points: first, new discoveries of the genetics that activate cancers; second, development of molecular tests to look for these activating changes; and third, emergence of novel drugs to target them.

qa 2C3: How do these tests work?

Garcia: You can look for genes or you can look for the proteins they encode. It’s the protein that causes the disease – if a gene isn’t expressed as a protein, it doesn’t do anything. But then life is never easy because proteins are very unstable and most of the time, we can’t be sure that just because we don’t detect a protein it means that it was never expressed. Then again, just because a genetic abnormality is present, we don’t know that it’s being expressed. So we have the ability to look at it from the gene side or the protein side.

C3: And after looking in these ways, you can tell if a tumor is positive or negative for a certain genetic abnormality and then prescribe the right drug?

Garcia: In most cases, yes. But again, it’s not always that easy. Biology is complex. When we apply a test, we try to categorize the result as yes or no, but commonly the result is continuous. We put a threshold that says below this number we’ll call it low or negative and above this number we’ll call it high or positive, but this is artificial.

C3: So it’s not as simple as using a test to match a tumor with the right drug?

Garcia: Well, for example, we see that about 40 percent of patients with a rearrangement of the gene ALK don’t respond to the drug crizotinib that targets ALK. Why is that? Maybe they were diagnosed too late – maybe the tumor is past some point of irreversibility. Maybe the rearranged ALK gene is present but it’s not being expressed – and the drug is trying to work against a protein that is not being made. Or maybe, in addition to the ALK rearrangement, the tumor is being driven by something else that we didn’t test for, or don’t know how to test for, or don’t even recognize has the power to drive cancer. What we label an “ALK-positive tumor” indeed may represent many scenarios.

C3: You’re talking about subtypes of these subtypes…

Garcia: We’re trained to simplify because then we can act. And I don’t disagree that we have to simplify at the beginning. For example, we need to call a lung tumor ALK-positive or ALK-negative so that we know whether or not to use crizotinib. But then we can’t lose track of the complexity of the tumor. If we find ALK and then we stop, what else might be driving the tumor? Now that we know some of the causes of cancer, some of the tests, and some of the drugs, our challenge is to look beyond the simplicity of these one-to-one matches and ask what’s next? Every year we’re learning to answer this question for new cancer subtypes. New tests are essential in this process.