The Institute of Medicine’s committee charged with the Review of Omics-Based Tests for Predicting Patient Outcomes in Clinical Trials is holding its second meeting today and tomorrow, March 30 and 31, in Washington, DC. And the stage is set for more revelations and critical discussions.
Perhaps the most notable name on the agenda is that of Joseph Nevins, PhD, The Barbara Levine University Professor of Breast Cancer Genomics at Duke University Medical Center.
Dr. Nevins was the senior author on all of the now-retracted papers by Anil Potti, MD, and, therefore, presumably oversaw the events that ultimately led to formation of this IOM committee. (Briefly summarized here and here).
Until now, he has remained largely silent about the problems that occurred, providing little insight into how they happened in the first place or how they were allowed to continue after outside researchers -- and indeed the National Cancer Institute -- started asking questions and pointing out errors.
This is his opportunity to provide that information. We can only hope that he will take the opportunity so others can, perhaps, learn how to avoid such outrageous problems in the future.
Dr. Nevins, though, cannot be the only one held responsible for the problems. As I noted previously, there are supposed to be other checks and balances in science and clinical medicine, including institutional review boards (IRBs) and peer review.
Neither worked in this case. Both will be the subject of discussion at this week’s meeting.
And the discussions are likely to be interesting.
Keith Baggerly, PhD, one of the biostatisticians at MD Anderson Cancer Center in Houston, who started ringing the alarm bells about the Duke work, has provided the IOM committee with copies of the letters from journal editors explaining why they wouldn’t publish his critical assessment of Dr. Potti’s work.
In one case, Daniel G. Haller, MD, Editor-in-Chief of the Journal of Clinical Oncology, simply wrote “I regret to inform you that we cannot accept your correspondence for publication. Thank you for giving us the opportunity to consider your correspondence for the Journal of Clinical Oncology.” No further explanation provided.
In another case, Dr. David Collingridge, Editor of The Lancet Oncology, declined to publish Baggerly’s manuscript, citing the opinions of a statistical reviewer and other editors at the journal. Comments from the statistical reviewer include the following: “I note that you have already had a spat with Potti and colleagues, and you revisit some of the same issues here. There is some merit, of course, in revisiting this problem; but that said, you can't have much better publicity than Nature Medicine. Overall, my view is that this is one of these issues where a difference of opinion will persist no matter [what debate is undertaken].”
A difference of opinion? Baggerly and his co-author Kevin Coombes, PhD, called into question the Duke group’s entire analysis and showed that they were unable to reproduce it. Isn’t reproducibility a tenet of scientific method and advance?
As Dr. Baggerly has pointed out numerous times, including in interviews with me, these same journals had no problem publishing the flashy data from the Duke group. Yet they seem to feel no responsibility to air scientific criticisms of the work thereafter. If that is the case, then how does the community at large know that such criticisms exist or that there might be a problem?
I don’t have the answers to these questions. Nor do I think there are easy ones available. But I am heartened to see that the IOM committee has decided to address these larger ethical issues, as well as the methodologic challenges involved in the development and testing of ‘omics signatures.