The New England Journal of Medicine, today, retracted the 2006 paper by Anil Potti, MD, and colleagues from Duke University.
The retraction notice, written by the original study authors, indicates they were unable to replicate their own work: “We have tried and failed to reproduce results supporting the validation of the lung metagene model described in the article. We deeply regret the effect of this action on the work of other investigators.”
Somehow that apology doesn’t seem to begin to make up for the resources -- time and money -- consumed by this mess.
Keith Baggerly, PhD, and Kevin Coombes, PhD, the biostatisticians at the University of Texas MD Anderson Cancer Center who initially flagged the data, estimate that they spent more than 1,900 hours (that’s nearly a year’s work) trying to figure out what the Duke group did. Not to mention the effort by the National Cancer Institute biostatistics group and FDA auditors.
Today’s notice is the fourth retraction for the group, following on retractions from Nature Medicine, Journal of Clinical Oncology, and Lancet Oncology.
There are still numerous papers that include Potti and his mentor, Joseph Nevins, PhD, then Director of Center for Applied Genomics & Technology with Duke’s Institute for Genome Science and Policy, as authors. Two papers in the Journal of the American Medical Association list Dr. Potti as the senior author, making them particularly concerning.
But what about all of the papers that include Drs. Potti and Nevins as middle authors? Did those papers rely on faulty methods as the retracted papers did? Or were these high-profile scientists given authorship as a courtesy, a recognition of their “pioneering” efforts despite having little influence on how the experiments were conducted or reported?
I don’t know the answer to that. (I’ve asked a Duke spokesperson to find out if there are plans to retract more papers and if the methodology was different on recent studies, but I haven’t heard back.)
I do know, though, that I now look at each of the genomic signature papers coming out of Duke with a bit of skepticism and I doubt that I’m the only one doing so. That might not be fair to the fellow or postdoc whose project it is, but I realize now that I don’t have the expertise to evaluate the data adequately -- and clearly the in-house experts at Duke haven’t been doing it for me.
With that in mind, maybe Potti and Nevins and their coauthors ought to apologize not only to the community at large, but also to their younger colleagues at Duke whose work, no matter how good it is, is going to be looked at with a bit of uncertainty because of the actions by Potti et al.