A few years ago, researchers at Novartis saw a disconnect between its robust cancer drug development pipeline, and the speed and efficacy by which it tested small molecules on real tumors. So it innovated the way it conducts early-stage research.
“Our remit as a company is to bring drugs to patients as quickly and effectively as possible.”
“If we could discover whether there’s a response or not earlier,” explained Dr. August Salvado, U.S. Oncology VP of Early Development, Strategy & Innovation at Novartis, “we could accrue additional patients to the study, or declare futility and avoid expending additional time and resources.”
“And if we could target specific genetic mutations, we could test across tumor types and investigate benefits against different cancers that evidence the same markers.”
Doing so would constitute a “new way of looking at drugs,” according to Salvado, and require upending both the traditional structure and timing of clinical trials.
The classic approach is to define a patient population by tumor type and, in some studies, by genetic mutation, then recruit a set number of sites (usually medical research institutions), give them a research protocol, and then wait for patients with the appropriate characteristics to, literally, show up. By not pre-planning specific sites and assuming that physicians are already using commercial or local genetic testing to identify patients with actionable mutations, the program could establish a much wider “footprint” of potential study sites for each study.
“We were interested in finding all of the tumor types in which a specific mutation might occur, so we thought it made sense to let the physicians who see those diseases screen for it,” Dr. Salvado said. “It’s one thing to have a study site waiting for a patient, and another to have a patient looking for treatment.”
Salvado proposed the idea to Christi Shaw, Novartis’ U.S. Oncology leader at the time, who “was very willing to give us some rope,” and then pulled together a cross-functional team of experts who are involved in managing the existing process, and met weekly to figure out how to disrupt it.
“We had the maps of what needed to be done, so the question was how could we change processes to accommodate a rapid timeline,” he said. “We also needed a trial design that would allow us to explore multiple tumor types with the same genetic abnormality within a single study.”
Its deep experience in the classic approach provided the starting point. First, the team decided to centralize the IRB process in order to speed study startup, and then created site qualification via a telephone checklist to get the process going quicker. (They’d still be followed-up with physical site inspections later on.) It turned to its existing relationships with many physicians who would likely qualify for the program — possessing a clinical coordinator, locked storage for drugs, etc. — to populate its initial trial.
“We wanted to change the way doctors thought about clinical trials, but not get between them and their patients.”
The program, dubbed the Signature Clinical Trial Program, launched in 2013, and opened its first trial in about 1/10th the time it would have taken with the classic approach. Novartis has since run seven additional single agent trials, and is now exploring studies of combinations of drugs (which can yield unique benefits). Nearly 600 patients have participated via close to 400 trial sites.
“Genetic screening is fairly new, and having it more widely available is what makes our signals program possible,” said Salvado. “It lets us better allocate our resources and avoid failures in latter stage trials, and get the drugs with the most reliable potential into the formal development process sooner.”
“It lets us know we’re fishing in the right pond.”