Puente: Getting the right patient into the right trial is an elusive goal in cancer. The field is increasingly being looked at as a collection of rare diseases, each with different evidence requirements. It follows that we have a big problem in convincing the registration authorities to act on the basis of a very small data sample, where you often find many inconsistencies. This is why, in my view, it can be productive to jump start the process by working in close concert with the authorities. Pfizer experienced this recently when we were approached by a consortium of Canadian academics to participate in a biomarker testing program for 10 rare diseases, in which we would supply the drugs while they funded the trial network of investigators. Initially, we balked because of the regulatory hurdles, but then Health Canada entered the picture and together we defined a protocol that set a 30 percent positive response rate as the benchmark for allowing Pfizer to obtain commercialization rights for the discoveries. So we were able to share the cost, define our regulatory risk exposure, and create a potentially remunerative business opportunity.

Looney: Are there other "rogue theories" about cancer discovery that could revolutionize treatment?

Stein: There is resurgent interest in immunotherapy, influenced by the pace of discovery against HIV. Companies are finally starting to invest significant resources in this area.

Rutstein: Cytotoxic chemotherapy, a traditional form of treatment, is being updated and improved. One trend we see is cytotoxics being reengineered to have a better safety profile.

Mehmud: We should expect many more small molecule targeted therapies. Precision medicine will be key to realizing their true potential. We have to keep the momentum around generating evidence with biomarkers and companion diagnostics, persuading more companies to take the up-front risk and invest there, even with currently available therapies.