New York Times,
Bert Vogelstein, a cancer geneticist at Johns Hopkins University, says he is haunted by three pictures.
The first shows a man’s upper body studded with large melanomas. The second shows what happened when the man took a drug called vemurafenib. Vemurafenib belongs to a relatively new class of drugs, called targeted cancer therapy. Unlike earlier chemotherapy drugs, they attack specific molecules found only in cancer cells. In response to the vemurafenib, the tumors shrank in a matter of weeks, to the point that the man’s skin looked smooth and healthy.
The third picture is a case of déjà vu. After 16 weeks of treatment, the melanoma returned. “All the lesions reappeared — every single one,” said Dr. Vogelstein. “That struck me as nearly as amazing as the fact that they had disappeared.”
The man died several weeks later.
This short-lived reprieve is heartbreakingly common in targeted cancer therapies. To understand why — and how to get around it — Dr. Vogelstein has teamed up with mathematicians to create detailed models of cancer.
Their research demonstrates a little-appreciated but inescapable fact about cancer: It is an evolutionary disease. And their studies are provoking new thinking about ways to use drugs to kill cancerous cells. Some of their findings were published on Tuesday in a paper in the journal eLife.
As cells divide, there’s a tiny chance they will acquire a mutation. Sometimes a mutation will speed up the growth of the cell compared with surrounding cells. Natural selection then unfolds within the body. As the fast-growing cells become cancerous, mutations that enable them to grow faster take over the population inside of tumors.
Mutations can also allow cancer cells to resist targeted cancer therapy. Typically, the target protein’s structure changes, so that the drug can no longer grab onto it.
To understand the rise of resistance in tumors, Dr. Vogelstein and his colleagues turned to those mathematical models. Last year, for example, they studied 28 people whose colon cancer was treated with a drug called panitumumab. Even before the start of treatment, the scientists found, one in a million tumor cells already carried mutations making them resistant to the drug.
That might not sound like a lot of cells — until you remember that by the time tumors are detectable, they may already contain billions of cells. That means that a thousand cells in each tumor will be able to resist the drug.
The mathematical model Dr. Vogelstein and his colleagues created accurately reproduced the pace at which the resistant cells produced a new tumor. “You could predict like clockwork when they came back,” said Dr. Vogelstein. “Things rarely work out that well in research. “
The model illuminates how targeted cancer therapy can be doomed even before it starts. “It’s just a fait accompli,” said Dr. Vogelstein.
He and his colleagues then started using their mathematical models to look for ways to avoid this failure. One possibility is to use two drugs instead of one.
The most common way that oncologists use two drugs is to prescribe them in sequence. Only after the first drug fails do doctors use the second. Dr. Vogelstein and his colleagues found that resistance can defeat this strategy too.
“If you use the drugs one after the other, this is a certain recipe for treatment failure,” said a co-author of the eLife paper, Martin Nowak, the director of the Program for Evolutionary Dynamics at Harvard University.
As the tumor rebounds from the first drug, the cells inside it mutate. Hundreds or thousands of cells n become resistant to the second drug, too.
“The second one fails for the same reason as the first one,” said Dr. Nowak.
The new study indicates that the best strategy is to hit tumors with two or more targeted cancer therapies at once. The chances are vanishingly small that any cell in the tumor will already have two mutations making it resistant to both drugs.
As the cancer cells die and the tumor shrinks, there won’t be enough time for a cell resistant to one of the drugs to gain a mutation making it resistant to the other.
Even using two drugs at once isn’t a guarantee that they’ll work, though. “If there exists a single mutation in the genome that gives resistance to both drugs simultaneously, then you cure nobody,” said Dr. Nowak.
Cancer experts who were not involved in the study found the study intriguing. Maxine Sun, a public health researcher specializing in cancer at the University of Montreal, called the study “quite impressive,” while Dr. Sumanta Pal of City of Hope Comprehensive Cancer Center near Los Angeles called it “provocative.”
Dr. Pal was reluctant to embrace the conclusions without more data from trials on cancer patients, however. “Ultimately, clinical validation is more important than mathematical validation,” he said.
Dr. Sun worried that figuring out which combinations of drugs are safe and effective would be challenging. “That would be a very difficult and time-consuming task,” she said.
Dr. Vogelstein agreed. “There are serious, practical issues involved in doing this,” he said. “But there are serious, practical issues involved in not doing it. The sooner the drugs are combined, the better it will be for patients.”
Copyright 2013 The New York Times Company. Reproduced with permission.