| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Letters to the Editor |
Cancer Research Center, The University of Chicago, Chicago, Illinois
Healthcare Investment Banking, Merrill Lynch & Co., New York, New York
In Response: We are pleased to respond to the letter by Banerji and colleagues about our recent editorial on biomarkers in phase I oncology trials (1). We maintain that these studies have not been broadly useful in defining the best dose or schedule, or as a basis for termination of development. Like many others, Banerji and colleagues basically argue for the pervasive utility of biomarkers in early clinical cancer drug development on teleologic grounds (i.e., biomarkers are intuitively useful) but without regard for the evidence against the utility of these studies, as analyzed by Goulart et al. (2) and ourselves. Additionally, Banerji and colleagues cite two anecdotes that they claim support their thesis that "biomarkers have an important role in phase I trials." We believe, however, that these two examples are consistent with our conclusion.
Their first example is the poly(ADP-ribose) polymerase inhibitor KU-0059436, which was shown to have significant activity in ovarian cancer during its phase I development, with the conclusion that the maximally tolerated dose is 400 mg twice daily (3). In addition, there was evidence of poly(ADP-ribose) polymerase inhibition at doses as low as 40 mg daily, and without evidence for greater poly(ADP-ribose) polymerase inhibition at doses above 100 mg twice daily. It is unclear, however, to us how these biomarker results have been used to further the development of the drug. Selective enrollment of BRCA-mutated patients in the dose expansion phase (as was used in this study) is not consistent with the objectives of phase I trials but is a reasonable approach for phase II studies, an issue separate from those discussed in our editorial, which focused on pharmacodynamic biomarkers.
The second example cited by Banerji and colleagues (4) is the Hsp90 inhibitor tanespimycin (17-allylamino-17-demethoxygeldanamycin). In addition to the cited study, there are six other published phase I studies of this agent (including two in children), involving a total of 197 patients (5–10). Notably, all studies have shown effects on biomarkers, and no study reported any partial responses. Using an average incremental cost of $6,675 per patient (11), the estimated cost of these biomarker studies might exceed $1,000,000. As the value of these biomarker studies is unclear from the literature, their cost-effectiveness cannot be calculated.
Although we disagree that the studies to date have been useful, we do encourage the use of biomarkers when toxicities are present that may not be due to target inhibition, particularly if there has been no evidence of antitumor activity. If there were also no evidence of a biomarker effect, this would be grounds for termination of development, at least on that schedule. This is unlikely, however, to have a significant effect on the failure rate of drugs in phase III trials, which is most likely due to the current paradigms for phase II oncology trials, rather than a failure to incorporate sufficient biomarkers into phase I studies (12, 13).
Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.
References
Related Article
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
