Each passing year, evidence grows that there is no single type of “breast cancer” or “lung cancer.” These broad categories simply refer to diverse tumors that just happen to originate in certain parts of the body. Scientists now know that what matters most in determining the behavior of a particular cancer (and its response to specific therapy) are the molecular pathways that drive malignant cell growth instead of where the tumor begins in the body. A tumor in the lung and a tumor in the breast, for example, may have more in common with one another (and be fueled by abnormalities in the same pathway) than two different breast cancer malignancies.
And yet, cancer drugs are still approved by the U.S. Food and Drug Administration on a disease-by-disease basis. So if researchers think that an individual drug for, say, chronic myeloid leukemia might work against a different kind of tumor because the two illnesses share similar biological underpinnings, they have to conduct a whole new clinical trial and go through the approval process all over again for the second condition.
That standard approach might be about to change. A new process, in which the FDA approves cancer drugs on the basis of which malfunctioning growth pathways they target rather than on what doctors call the resultant condition, has been quietly batted about on the sidelines of conferences or over drinks among oncologists for the past couple years. Although the FDA has not formally proposed any such change, the agency’s cancer czar, Richard Pazdur, floated the idea at a public conference on cancer care earlier this month in Washington, D.C., suggesting the idea may be gaining official traction.
If enacted, such a regulatory scheme could conceivably speed along clinical trials because it would allow investigators to draw from a large pool of people with diverse tumors that act along the same pathway instead of culling thousands of people with the same official cancer type. "Much still needs to be learned, but as we further understand cancers, it is possible we could see a shift from approving drugs based on disease type to the molecular pathways that drive them," says Stephanie Yao, an FDA spokesperson.
There is a catch: because each drug is a unique molecular entity, it has the potential to behave slightly differently in the body even if it targets the same pathway as another drug with a sterling record. Take ponatinib, a drug that targets the hallmark molecular abnormality in chronic myeloid leukemia (CML). Several other drugs that inhibit the same biological target had already been approved for frontline CML treatment. Promising early-stage trial results led to ponatinib’s application being fast-tracked, and it was approved in December 2012. Last month, however, less than a year after it received its clearance, the medication was taken off the market because the evidence revealed an uptick in life-threatening blood clots and cardiovascular issues. If ponatinib had been cleared under the proposed new scheme, however, several cancer doctors worry that the deleterious impacts may not have been picked up as quickly.
With a faster process a drug might get to market before as much safety data were collected. Once the FDA green-lights a drug and it is ushered into the hands (and IVs) of patients, there is no real incentive for a company to continue to collect information about side effects in any systematic way. When they learn about bad reactions to the drug they are legally required to report them, but they do not regularly survey patients as they do during the approval process. “There is no track record that this [new approach] would work,” says Hagop Kantarjian, the chair of the department of leukemia at the University of Texas M. D. Anderson Cancer Center. “It could be a way to clear drugs faster, but the better way to do that would be to reduce the hurdles to clear,” he says. Of course, because the proposal has not actually been put forward by FDA it is hard to tell exactly what the level of scrutiny would be for drugs that were cleared through such an avant-garde approach.