Upcoming Battles in the War on Cancer

“There are a lot of great ideas out there and a lot of targets, but clinical trials take too long and cost too much money for us to pursue all of them,” says Safi R. Bahcall, PhD, founder and president of Synta Pharmaceuticals. According to widely cited estimates, the cost of bringing a new drug to market runs $1.8 billion, though some experts say the true cost is at least twice as much (Nat Rev Drug Discov 2009;8:959–68). Add in an uncertain regulatory environment and the slim chance of moving a drug from preclinical studies through phase III trials to approval—about 1%—and “the system becomes unsustainable,” Bahcall says.

Obtaining molecular profiles of patients' tumors prior to study enrollment may save time and money in the end. That's because a drug can be tested in the subset of patients most likely to respond to it, meaning that fewer patients need to be enrolled in the trial to yield a statistically significant result, explains Levi Garraway, MD, PhD, coleader of the genetics research program at the Dana-Farber/Harvard Cancer Center. This was the case with crizotinib (Xalkori; Pfizer), a drug that hit the market just 4 years after target discovery, because lung cancer patients who received the agent were tested for the EML4–ALK gene fusion, which crizotinib targets, before enrolling in a clinical trial.

Researchers may also test a panel of targeted drugs early on, see which agents elicit the strongest responses compared with placebo in given subpopulations and then proceed with a phase III trial in just one or two of them.

Finding biomarkers to help monitor cancerous activity might quicken the pace of discovery, too. Researchers studying HIV can easily determine whether an agent is working by measuring the amount of the virus in a blood sample. But for most cancers, such a biomarker doesn't exist.

“Even though we have lots of drugs, we don't always know whether they are really working,” says Garraway. Examining biopsy samples or circulating tumor cells not only at diagnosis but also throughout treatment might yield that valuable information, potentially shortening trial times or shifting trial endpoints from, say, 5-year survival to 2-year survival.

A lack of participation also plagues the clinical trials system: Only about 3% to 5% of cancer patients enroll in a study. That figure would increase dramatically if Medicare and private insurance companies played a role in clinical trials, says Debasish “Debu” Tripathy, MD, coleader of the Women's Cancer Program at the University of Southern California (USC) Norris Comprehensive Cancer Center. “Every patient in the country who qualifies for a trial would be enrolled,” he predicts. Payers could tie physician pay to data submission. This approach would also accumulate a wealth of data on cost-effectiveness of therapies. “By being part of the solution to this problem, public and private payers could really advance personalized medicine,” Tripathy says.

Researchers need to quickly disseminate their data and encourage others to build on what's been learned, says Stephen Friend, MD, PhD, president and cofounder of Seattle's Sage Bionetworks. “Sequestering all of the data until publication, which can take 6 months or longer, isn't ideal.” His nonprofit medical research organization is building an open-access, online platform called “Synapse” that will allow easier sharing of data among institutions. Researchers will tag and annotate their submissions to receive credit for their contributions and modify the platform to meet their needs, unlike other static systems. “Synapse” will be available in 2012.

Additionally, says Friend, “the scale and scope of genomic projects is much bigger, and we need to expand our information systems to accommodate that.”

Creating new infrastructure will require attention to detail, says Chicago's Maitland. The medical community will need to standardize the terminology for recording clinical information, and computer databases will need to store information in compatible ways. Researchers will also need to focus on their protocols to produce reliable results. For instance, having blood samples sit at room temperature for different amounts of time before freezing can negatively affect another researcher's ability to make accurate comparisons from one sample to another.

To minimize—or at least account for—some of that variation, the University of Chicago is adopting a blood sample tracking system developed at Indiana University (Bloomington, Ind.) that uses barcodes and scanners similar to those used for express mail services. The samples are scanned at each step in the process, explains Maitland, and all of that information is fed into a database. Other scientists can then account for how the samples were handled in conducting their own tests.

The pace of technological advancement has been astonishing. Early gel-based sequencing techniques have given way to capillary sequencing methods. To hunt for genetic amplifications, deletions, translocations, and variations in copy number, researchers can perform whole-genome sequencing, RNA sequencing, and whole-exome sequencing. Massively parallel sequencing may soon become the norm.

Integrating proteomic, epigenetic, and other forms of data will give researchers additional insights into subsets of cancers. But rather than simply latching on to additional technologies, some cancer experts point to the need to better understand the data already generated, which will require building multidisciplinary teams. “Separating wheat from chaff—and doing it quickly enough for patients to benefit—remains one of our key analytical challenges,” says Stephen Gruber, MD, PhD, MPH, director of the USC Norris Comprehensive Cancer Center. The goal is to find actionable mutations, but most mutations are passenger mutations whose function remains unknown.

“Some would argue that we're missing information contributed by proteomic variation,” he continues. “I say, ‘One step at a time.' We will get there, but we need to walk before we learn how to run.”

For more news on cancer research, visit Cancer Discovery online at www.AACR.org/CDnews.