So-called biomarkers already are used in a variety of diseases to determine whether a patient should stick with one therapy or try another, and to speed drug development.
But while scientists have discovered many potentially useful biomarkers, even the few in use today aren't fully understood. A shrinking tumor, for instance, doesn't necessarily predict longer survival. The Food and Drug Administration sometimes approves new drugs based on such promising signs, but is supposed to force manufacturers to track patients in longer studies to be sure they truly benefit.
On the other hand, it's not always a bad sign if a tumor doesn't appear to shrink quickly; cells inside the tumor may be dying, but the X-rays typically used to measure can't see that.
In the project announced Tuesday, the FDA, National Cancer Institute and Medicare will collaborate to find better biomarkers for cancer, or better ways to measure and use them.
First on the list is non-Hodgkin's lymphoma. A study will measure whether a special type of PET scan — FDG-PET, which measures cellular metabolism — can predict very early whether a patient is responding to treatment.