Three clinical trials have been placed on hold after news broke that a researcher involved with the studies may have lied on his resume.
Cancer researcher Anil Potti of the Duke University School of Medicine has been placed on “administrative leave” pending an investigation into false claims, listed on applications for research funds, that he was a Rhodes Scholar. According to the New York Times, the American Cancer Society has also suspended payment of a $729,000 grant awarded to Potti based on an application that referenced the Rhodes scholarship.
The cancer trials aimed to test whether patterns of gene expression in tumours could be used to predict how those tumours respond to specific cancer drugs. Such predictions have been dreamt of since the earliest days of genome-wide gene expression analysis, and Potti and his coworkers appeared to take a significant step toward that goal in a 2006 Nature Medicine paper.
In the paper, Potti and his team reported gene expression profiles and drug sensitivity of cancer cells grown in culture. From these results, the team derived computer algorithms that predicted, based on which genes were turned on or off, which cell lines would be sensitive to commonly used chemotherapies including docetaxel, cisplatin, and cyclophosphamide. The paper has been cited 226 times, according to ISI.
Unfortunately, attempts to replicate the work failed. In 2007, Kevin Coombes of the University of Texas M.D. Anderson Cancer Center in Houston laid out a series of critiques, noting that his team had been unable to reproduce the drug sensitivity results. Furthermore, the gene expression analysis was fatally flawed, Coombes wrote, by inappropriate statistical maneuvers and even a simple indexing error that caused, for example, gene #1881 to be listed as gene #1882. Potti and his coworkers responded that Coombes’s team did not accurately reproduce the drug tests, that their statistical methods were sound, and that the indexing error had since been corrected.
Clinical trials based on this work were previously suspended and then resumed after a panel convened by Duke University reviewed the data. But on 16 July, Cancer Letter published a story about Potti’s false Rhodes credentials and by Monday, a veritable who’s who list of biostatisticians, some of whom pioneered the analysis of genome-wide gene expression data (Quackenbush, Storey, Speed, and Tibshirani, to name a few), had prepared a letter (available online thanks to Science Insider) asking US National Cancer Institute chief Harold Varmus to intervene and halt the trials until a “fully independent” review is conducted.
Duke University announced yesterday that it had halted the trials.
Image: Duke University