In an ideal world, the new group of targeted therapies making their way through clinical trials should follow the hugely successful Herceptin(trastuzumab)-HercepTest model—they should be utilized in subsets of patient populations, to be determined by a diagnostic that correlates well with therapeutic response rates. After all, targeted therapies are intended to act by modulating specific molecules or receptors and therefore ought to work on those patients who, at a minimum, have those molecules, or, as in the case of Herceptin, a Genentech Inc. drug for breast cancer, who have too much of the targeted receptor. (The HercepTest diagnostic kit is used to distinguish those who overexpress a particular protein, Her2/neu, from those who either express it in moderate levels or don't express it at all.) Indeed, in approving Herceptin, the FDA required it to be used in conjunction with a diagnostic—the first time the agency has ever made such a stipulation. FDA officials have indicated informally that they would like companies to incorporate diagnostics into other studies of targeted therapies to best determine which patients can benefit from these drugs. (See "Dako Ponders the Expanding Universe," IN VIVO, March 2000 [A#2000800040.)
But scientists aren't yet convinced of the best ways to use the current diagnostics in pre-selecting patients who are likely...
