There was no way the marketing team could have projected accurate sales data because kinase inhibition had never been attempted. It was a completely new kind of medicine. No one knew what the duration of treatment would be, the exact patient population for which it would be appropriate, or how quickly doctors would adopt the drug into their standard of care. These calculations were simply outside their scope of reference because this drug was an entirely new way of treating cancer.
Matter was happy to report to Druker that at last the evidence and science behind the drug had prevailed over the business hesitations. A pre-IND meeting had been scheduled with the FDA for sometime in the next two months. “So, with delays, we are still alive,” Matter wrote.
Finally, on December 23, 1997, Druker received a letter from John Ford, now the senior clinical research physician for Novartis, via Federal Express. The clinical group had finally managed to persuade the toxicology group to at least discuss the possibility of starting a clinical trial with the FDA. The process, Ford couldn’t help but mention, was likely advanced by the resignation of the head of the toxicology group. Vasella had approved moving the drug into clinical trials. The pre-IND meeting had been requested, and would likely happen in January or February. If the FDA gave the program its okay, then a clinical trial would begin around May. If the FDA requested changes to the proposed protocol, the trial could begin in June. If the FDA requested additional toxicity data, then the trial would have to wait until the last quarter of the year. Whichever option panned out, STI-571 would be in trials that year.
MEANWHILE, THE MONKEY studies continued. In the high-dose group, the animals vomited or had diarrhea. Many of them lost their appetite and became dehydrated. One of the female monkeys in the high-dose group died after losing 15 percent of her body weight, but it wasn’t clear to the researchers that her illness was due to the drug. Many of the animals had various forms of Plasmodium, a malarial parasite common in monkeys bred in Africa and Asia. The parasite was the more likely culprit behind the changes in red blood cell counts, but the drug may have allowed the parasite to proliferate. The lowest dose of the drug given to the monkeys did not cause any problems in their blood or gastrointestinal systems. All of the male monkeys had testicular changes as a result of the drug. The company concluded that the drug was “tolerated well by the animals with minimal toxicity,” according to the report eventually submitted to the FDA. The high dose, however, would cause “definitive toxicological/pharmacological signs.”
To examine the effects of a particularly high dose of the drug, a group of monkeys received 1,800 milligrams per day for six days before an excess of toxicity led the scientists to reduce it to 1,200 milligrams per day for the remainder of the study. One of the animals had to be euthanized after it developed severe kidney disease.
All told, STI-571 was studied for thirty-nine weeks in rats, thirty-nine weeks in one group of monkeys, thirteen weeks in another group of monkeys, thirty-nine weeks or more in mice, at least twenty-six weeks in rats, and at least two weeks in rabbits. It was studied in rat fetuses, cells from Chinese hamster ovaries, and bone marrow from rats. It was mixed with Salmonella and E. coli in tests designed to show whether the drug would cause genetic mutations. A special additional study to further investigate toxicity in dogs concluded that, although liver problems remained a concern, the drug was safe.
Theoretically, the early indications of the potential efficacy of STI-571 should have been enough to launch a study within half a year after Druker reported his cell-line and marrow data to Matter and his team at Ciba-Geigy. The mandatory toxicology studies could have taken as little as three to six months. In principal, pharmaceutical companies can move ahead incredibly quickly if all signs point toward go. “Let’s assume everything has been done to go into humans and you have a hypothesis [that] you believe in, and you have a patient population who is potentially going to profit from it. It’s only the decision [that remains],” said Vasella. “It shouldn’t take months. It should not.” And yet STI-571 lingered in its preclinical state for nearly three years after the completion of the cell-line studies, and for about two years after the initial toxicology studies and the business merger. The kinase inhibition program had started in 1984. The lead compound had been synthesized by 1990. Seven years later, the phase I study had yet to begin.
PART III
Human Trials
1998–2001
• • •
Behind every new drug entering clinical testing lies a principle justifying why it should work, a mechanism proved by similar drugs already on the market. But when a new kind of medication enters trials, that proof of principle is nonexistent. Testing the drug is the only way to test the principle.
To prove the principle that cancer could be treated by targeting the abnormality driving it, STI-571 had to be given to patients with CML. Everyone involved—the investigators, the company executives, and the patients—had to accept the risks of testing this novel drug. As solid as the rationale behind the design was, proof could come only from evidence. Success was far from guaranteed.
24
_______
THE QUICKEST ANSWER
The phase I trial of STI-571 was a first for kinase inhibition, but also for Druker: He’d never experimented on patients before. It was a moment he’d envisioned for years, picturing how he would be with the patients and what he needed to explain to them. Their lives would be in his hands, and his job was not only to keep them safe, but also to make sure they knew what they were in for. He knew that giving them the best possible care meant not raising their hopes that this untested drug would do any good.
Patients enrolling in phase I clinical trials are always informed that the goal of the study is not to help them but to help advance medicine. This phase, the first of three that are required for FDA approval of a new drug, is also known as a dose-finding study. Patients are started at an extremely low dose of the experimental medication, and that amount is gradually increased. The idea is to find out how much of the drug a patient can safely tolerate, based on the principle that the more drug that is given, the more effective it will be. The highest dose at which the drug can be given safely is called the “maximum tolerated dose.” Identifying that dose is the goal of a phase I study. These are safety trials, designed to ensure that the drug is not too dangerous for human consumption.
For this reason, the cancer patients who enroll in phase I studies for new drugs are usually those who have run out of treatment options. A newly diagnosed patient who has not yet received the standard treatment for their disease would not be asked to try an experimental drug. Rather, phase I trials are populated by people whose cancer has progressed after trying every other treatment. These patients must choose between accepting palliative care to ease the pain until they die or trying an experimental drug, if an appropriate one exists, in the name of helping science. When they enroll in phase I trials, cancer patients sign consent forms acknowledging their understanding of what the study is offering. They aren’t supposed to expect the drug to work or to hold out hope that it will.
But the design of the STI-571 trial was different from most. Phase I studies are always small, and this one followed suit: The study would enroll about thirty patients and would last six to twelve months. The distinguishing feature of this phase I trial was the decision of which CML patients would be permitted to enroll.
Despite the endless delays, many features of the phase I protocol had already been decided. From the earliest meetings at Ciba-Geigy in 1996, the team agreed that the trial would be restricted to CML only. That approach was a departure from the typical phase I trial design, in which new drugs are tested on multiple cancer types at once, in case the agent works for many malignancies, or for one tumor type but not another. That approach makes sense for drugs that attack cancer in a general way, working against mechanisms that are, for example, present in normal cells but heightened in cancerous ones. Often, it’s not until the later phases of clinical trial
s that enrollment is restricted to one specific cancer type, the one the company deems most likely to win FDA approval based on the broader phase I investigation.
But this was a new kind of drug. Aside from tamoxifen, which targeted the higher levels of estrogen present in some breast cancers, cancer drugs weren’t aimed at specific peculiarities of specific cancers. With the kinase inhibitor, restricting testing to only CML made sense because the drug blocked a protein linked to a genetic mutation specific to this cancer type.
Next they had to figure out what stage of CML to test. Focusing on blast crisis, the final stage of the disease, would have been in keeping with the traditional phase I structure, but the company didn’t want to do that. By the spring of 1996, when Ciba-Geigy held the second trial planning meeting with outside oncologists, Ford and the other executives had boldly decided to test the drug not on the most advanced patients, nor on those in the middle, accelerated stage, but rather on those in earlier stages of the disease. “My initial response was, ‘Can you do that?’” said Druker, who knew that experimental drugs were tested on only the sickest cancer patients as a matter of medical ethics. Yes, he was assured, they could.
That change wasn’t the most radical one brought to the table. Ford and Gratwohl had even discussed testing the drug on newly diagnosed CML patients as a one-month treatment before they went on to receive the standard therapy, although that idea was quickly dismissed. “In 1996, no one would have considered that,” said Druker. The FDA and the medical establishment consider it unethical to give an unproven drug to patients who have not tried the current medication. Enrolling earlier-stage patients whose disease had progressed with interferon was different enough. “We already knew we were doing something unusual in a clinical trial,” said Druker, who took a stance against enrolling newly diagnosed patients.
When the time finally came to open the trial in 1998, Novartis maintained those early decisions. The study would enroll CML patients only, but those in blast crisis—the very sickest—were not allowed to join. The study would be open only to patients who’d failed interferon but were still in the early stages of the disease.
At first glance, it could appear that Novartis designed the phase I trial to maximize the possibility of benefiting the CML patients who enrolled in the trial. Didn’t limiting the trial to only this disease signal the company’s understanding of the CML-specific nature of the drug, and Novartis’s commitment to helping patients in need? And enrolling patients who were in better shape than those in the final stages of the disease gave the drug more time to work before patients died.
But these weren’t the reasons, or at least not the main ones, behind the phase I protocol choices. Rather, the design was aimed at getting the earliest possible answer about whether the drug worked. “They were very concerned about committing to a trial that didn’t have a clear and quick readout,” Druker said of the trial planners at the company. Yes, the decision to limit the study to CML only was based on the fact that the Bcr/Abl kinase was so prevalent in this cancer and had not been found in any others. But the reason why the study was limited to CML wasn’t so that patients who needed the drug could get it, but because that was the only way to determine the drug’s potential efficacy. If the drug was not tolerated in a study with multiple cancer types, then clinicians might argue for a phase I study limited to CML only. But if the drug was not tolerated in a study of CML patients only, then it could be discontinued. And even though the objective of the phase I study was to find the maximum tolerated dose, the company wanted to know as soon as possible what the drug did in patients with CML.
Opening the trial to patients in the earlier stages of CML, rather than restricting it to the sickest patients, followed similar logic. Ford and others at the company thought that blast crisis patients would never respond to the drug. Druker had seen the drug work on cells from blast crisis CML and so thought there was a chance the drug would work on these end-stage patients, but the company didn’t want to risk a vague study outcome that would leave the drug in limbo.
Testing STI-571 on the healthiest CML patients in phase I would give the clearest indicator about whether or not the drug worked. If the drug was poorly tolerated among chronic-phase patients or showed absolutely no anti-CML activity in those patients, then perhaps the drug could be quietly shelved. A phase III trial—the costliest of the three clinical studies—for a rare disease with a limited market could be safely avoided.
Even at phase I, the investment was formidable. Novartis would have to cover almost all the costs of the trial: the medication, the battery of tests necessary to monitor its safety, biopsies, and any other procedure related to the care of each patient’s CML. There were expenses for recording test results and all the other data, for analyzing blood samples, for treating side effects. The costs were no different from those of developing any other new drug, though making this one was incredibly labor-intensive, requiring an arduous twelve-step process involving hazardous materials; producing just the single kilogram that was used for phase I took months. The major difference here was that, unlike industry sponsors for most phase I trials, the company was still not entirely committed to bringing this drug to market.
Druker and the other investigators were completely on board with the plan. They knew that this unusual phase I study design was the best way to prove whether the principle of kinase inhibition was valid. “Phase I is for safety, we knew that,” says Druker, “but the reality is we also wanted to plan a study to see if this thing worked or not.” The unconventional approach would enable the clinicians to see if the drug could be tolerated, indicating that it was, most likely, selective for the cancer-inducing kinase. And it would reveal, at least in a preliminary fashion, whether the drug affected the cancer.
The design also enabled them to investigate whether the concept of kinase inhibition was a viable approach for treating cancer. If the drug showed promise, then the results of the phase I study would set the stage for more profitable kinase drugs to enter the development pipeline. If the drug didn’t work, then Novartis would have minimized its losses by conducting just this single, small phase I study before canceling the kinase program altogether. The study had been designed to end the project as quickly as possible, if an ending was warranted.
THERE WAS MORE to decide than just who would take the drug. The team also had to figure out how the drug would be given. Even though the doses were experimental, the frequency with which the medicine was taken could have a huge hand in how well patients tolerated it and whether it affected the cancer. Clinicians from outside the company were pretty certain that the drug would have to be taken every day. Druker’s lab experiments had led him to believe that the kinase would have to be shut down for at least 16 continuous hours; this understanding had provided the impetus for the oral formulation years earlier. This cycle meant patients had to take a pill each day.
Novartis was pushing for a break in the regimen. That was how most chemotherapy regimens were done; treatments were frequent but not every day. Fearing toxicity, people at the company thought that a recovery time should be built into the trial. It took several rounds of discussions to settle on the once-daily approach. As everyone came to terms with the notion of daily treatment, the availability of the pill formulation—thanks to Peter Graf’s persistence years earlier—now seemed all the more fortunate. Phase I patients would need to be at the clinics regularly, but they would not have to sit for hours in chemotherapy chairs with an experimental fluid dripping into their veins.
As for how long the trial would last, the plan was for the clinicians to gather data for about six months, following the mold of typical phase I cancer drug trials. However, those six months included the treatment plus a few months of follow-up time to see how patients did afterward. Everyone assumed that the treatment itself would last just a month or two. Again, the history of chemotherapy was shaping their thinking. Most cancer treatments lasted a month, and then they were over. With chemotherapy, the goal was
to kill all the cancer cells with the poison in a short amount of time, and then stop before the poison did irreparable harm to the rest of the body. Whether in a standard clinical setting or on a clinical trial, patients didn’t remain on chemotherapy indefinitely. That wasn’t how such medications worked. Nobody imagined this drug would be any different.
25
_______
GETTING TO 200 MILLIGRAMS
As the trial finally got under way, Druker remembered Bud Romine’s letter from two years earlier asking to be his first patient if the medicine he’d read about in the Oregonian was ever tested on people. Romine, who lived in Tillamook, Oregon, was 64 when he was diagnosed with CML in 1994 and told he had about three years to live. Two years later, he saw the front-page article on Druker in the Portland daily newspaper and wanted to try the drug that had so effectively killed the malignant cells in bone marrow samples. When it was time to start enrolling patients, Druker remembered the bold request.
The study opened at three sites: OHSU, where Druker was the lead investigator; UCLA, headed by Charles Sawyers; and MD Anderson Cancer Center, in Houston, Texas, headed by Moshe Talpaz. Talpaz was famous in the oncology community for his work developing interferon, which, despite its harshness, was an extraordinary advance for CML. Of the three, he had the most experience with clinical trials. In fact, he was the only one who had led clinical trials; the company had first considered asking Talpaz to run the trial entirely, back when Druker had feared he’d be omitted. Left out of the equation was John Goldman, whose group at Hammersmith Hospital in London had been the first to confirm Druker’s 1996 results and had been asking Novartis to be part of the trials ever since. Wanting to study this wholly new drug in the most controlled way possible, the company had opted to restrict the phase I study to the United States.
The Philadelphia Chromosome Page 19
