The Boston-based Dana-Farber Cancer Institute has notified research journals that it desires to retract six studies and correct 31 papers — including those co-authored by top executives — as part of an ongoing probe, The Wall Street Journal reported Jan. 22.
More than 50 papers are subject to review, including four co-authored by Dana-Farber President and CEO Laurie Glimcher, MD.
Papers co-authored by William Hahn, MD, PhD, chief operating officer; Irene Ghobrial, MD, director of the clinical investigator research program; and Kenneth Anderson, MD, program director of the Jerome Lipper Multiple Myeloma Center are also being investigated.
In early January, an artificial intelligence and image screening software flagged Dana-Farber research papers for data manipulation. Sholto David, a molecular biologist, published the findings in a blog post and contacted both Dana-Farber and the affiliated Harvard Medical School with a list of concerns. Some papers showed experimental results with signs of copy-and-pasting by software such as Adobe Photoshop, Mr. David alleged.
Barrett Rollins, MD, PhD, Dana-Farber's research integrity officer and chief scientific officer emeritus, told Becker's that the institution was already reviewing potential data errors in a number of manuscripts listed in the blog before its publication. In "97% of those cases," action has already been taken, Dr. Rollins said.
He also noted that 16 manuscripts contained data generated in laboratories other than those of the four Dana-Farber authors named in the blog. The data necessary to determine whether these papers can and must be corrected lies in other laboratories, whose heads have been contacted where possible, Dr. Rollins said.
Dana-Farber did not find anomalies in three of the manuscripts flagged in the blog, Dr. Rollins said. He warned about the limits of AI, telling Becker's, "While software advancements can reveal anomalies not previously detected, AI programs are not foolproof. In fact, some of the allegations recently raised by a blogger against Dana-Farber researchers are wrong."
Even if errors are present, they do not necessarily mean the authors intended to deceive, according to Dr. Rollins.
"Our experience is that errors are often unintentional and do not rise to the level of misconduct," he said.
All four authors named in the blog post have faculty appointments at Harvard Medical School, adding to a growing list of allegations brought against the institution's researchers and executives. Claudine Gay, PhD, resigned as president of Harvard University Jan. 2 amid accusations of plagiarism in her scholarly work. An internal review found that two papers needed additional citations, but did not declare research misconduct.