On a sunny afternoon last June, Dr. Keith Flaherty stood before a large room packed with oncologists from around the world and described the extraordinary recovery of the melanoma patients in the experimental drug trial he was leading.
It was a moment he had looked forward to for months. Beyond a breakthrough for melanoma, the results were a promising sign for an approach to treatment for all forms of cancer that he and others had championed as more effective and less toxic than standard chemotherapy.
But even as he flashed the slide of his favorite graph, showing tumors shrinking in nearly every patient, his mind was on what had happened to them since.
In the weeks leading up to the annual oncologists' conference here, several of the patients on the trial of the drug known as PLX4032 had relapsed. One had died. Another, Christopher Nelson, who had made what seemed like a miraculous recovery in March, had lost his appetite again. Dr. Flaherty feared what he might see on Mr. Nelson's scan when he returned to his office at the University of Pennsylvania.
The drug's ability to stop the melanoma, on average, he told the crowd, "appears to be approximately six months."
"I was hoping we'd get more time," said Dr. Grant McArthur, one of the six oncologists on the trial team, voicing the thought on everybody's mind when the group met at the conference. None of them had a financial stake in the drug.
Dr. Flaherty, whose perpetual optimism about this kind of treatment, known as targeted therapy, raised eyebrows among some colleagues, declined to dwell on the drug's limitations. However briefly, PLX4032 had held off the cancer by blocking a particular protein in its cells that was spurring them to multiply. If such targeted drugs were ever to provide a lasting benefit, many oncologists believed they would need to be combined with others, much as cocktails of protease inhibitors have worked against H.I.V.
"We just need," Dr. Flaherty said, "to find the right combination."
If they acted quickly enough, they might even be able to help the trial's participants. Many were still in remission. Those who had relapsed were searching for another treatment, acutely aware that their time was running out: most melanoma patients die within a year after the cancer spreads.
The problem, which had bedeviled targeted therapies for other cancers, was that while PLX4032 blocked the protein made by one mutated gene, a second mutation now seemed to be driving the cancer's growth. If that mutation could be identified, they believed, its protein could also be blocked, in a game of biological Whac-a-Mole that just might be possible to win.
The most expedient approach would be to test PLX4032 in combination with other experimental drugs that targeted other mutations, including those seen in Dr. Flaherty's relapsing patients.
But a drug that gave a patient even a few months of life could generate billions in revenue. And the standard practice among pharmaceutical companies, which say they typically invest nearly a billion dollars developing and testing a single drug, is to get each drug approved individually before testing it with others, especially those of competitors that are still experimental. Even small Phase 1 trials can cost over a million dollars. And one drug that was safe and effective, they worried, might be tainted by association with another that proved to have toxic side effects.
As Roche, the pharmaceutical giant that had licensed PLX4032, made plans to test the drug in larger trials in hopes of quick Food and Drug Administration approval, Dr. Flaherty's colleagues in the laboratory would search for the new mutation in the tumor samples of patients who had relapsed, trying to understand why the drug had stopped working.
For his part, the doctor would try to keep his patients alive. And he would work to convince the pharmaceutical industry that the fastest path to finding a combination that really worked would require changing their standard operating procedure.