Here's an interesting profile of Bert Vogelstein, who has had a major impact on oncology over the years, especially in the area of cancer-associated genetic mutations. Some of his recent work bears on the question of how useful some of the newer drugs are:
Vogelstein seems to enjoy pricking balloons. Recently, he has focused on a new target: exuberance over targeted cancer drugs. He says he got interested after seeing a paper last year on melanoma therapy. It included photos of the torso of a man with melanoma who had received a new drug aimed at a mutated gene called BRAF. Before treatment, the patient's skin was riddled with metastatic tumors; soon after treatment, the tumors vanished, and the man looked perfectly healthy. Five months later, the tumors reappeared in exactly the same locations. The photos “blew my mind,” Vogelstein says. “Why do the tumors all return at roughly the same time? It's almost as miraculous as when they disappear.”
Targeted drugs for other cancers usually stop working after about the same number of months, presumably because rare resistant cells in the tumors continue to grow and ultimately proliferate. To investigate, Luis Diaz and others in the Vogelstein-Kinzler lab drew on a sensitive technique they had developed for detecting mutations in the very small amount of tumor DNA present in a cancer patient's blood. They collected a series of these “liquid biopsy” measurements from patients with advanced colorectal cancer whose tumors had become resistant to a targeted cancer drug. With Harvard University computational biologist Martin Nowak, they devised a model showing that even before the patient begins treatment, some tumor cells always carry genes with random mutations that can support resistance to targeted drugs. This form of resistance, they wrote last month in Nature, is therefore “a fait accompli.”
But the modeling study also suggested that this resistance can be delayed by combining two drugs that target different pathways. Indeed, Vogelstein and colleagues suggest that once a targeted drug has passed initial safety trials, it's so clear that single-drug therapy will fail that they consider it unethical to give patients just one such drug. “Why shouldn't you design a large, very expensive trial to incorporate more than one agent?” Vogelstein asks.
There are a lot of labs working on this "liquid biopsy" idea, and it's the sort of thing that you could only imagine doing with modern DNA sequencing technology (and modern DNA sequencing costs). A big worry, as with any screening technology, is the false positive rate. As you make finer and finer distinctions among different tumor types, the incidence of any given one in the population gets lower and lower, and thus your test has to be more and more reliable in order to avoid overdiagnosing hordes of panicked patients.
Interestingly, when I talk to people outside of the medical research field, they seem less worried about overdiagnosis than underdiagnosis (false positives versus false negatives). Psychologically, I can see how that happens - they don't want to the test to miss anyone. But being told that you do have cancer, when you really don't, is not a good outcome, considering what the therapy will put you through. And this is what makes things like the PSA test recommendation (and mammograms in younger patients) so controversial. In the push to make sure that you find every patient, you can end up harming more people than you help. "But if you just save one life. . ." goes the phrase, at least goes the phrase from people who don't realize that they might be ending the sentence with ". . .it's worth killing off a few more".
I hope that the blood test idea works out; it would be a great advance. But a less-than-optimal one could be worse than having none at all. Look for plenty of arguments about this in the coming years - I'll fill in some of the talking points in advance: "The FDA is holding back medical progress by not approving this new test". "The FDA has given in to commercial pressures by approving this faulty new test". "This test will end up hurting more people than it helps". "How can you be against cancer screening? Isn't it always worth looking?". "This is all just a disguised cost-cutting effort; they're approving this test because it's cheaper than doing better screening". "This is all just a disguised cost-cutting effort; they're not approving this test because they're afraid that too many people will be diagnosed with cancer". And so on.