The situation for Anil Potti, MD, and Duke University officials continues to deteriorate.
What started as questions about the quality of Dr. Potti’s work mushroomed in the last year to allegations of data manipulation and led to numerous retractions.
Now, it appears there may also have been inappropriate handling of patient data in at least one of three clinical trials that were based on the group’s work, according to discussions during a meeting of the Institute of Medicine’s committee to Review of Omics-Based Tests for Predicting Patient Outcomes in Clinical Trials yesterday and today.
Although she declined to reveal details, Lisa McShane, PhD, an NCI biostatistician who has been investigating Dr. Potti’s work, alluded to potential problems with how the clinical trial data were being stored and who had access to them.
When asked directly by a committee member if the trials were conducted as expected by standard guidelines, with patient outcomes recorded according to good clinical practice and kept in a locked clinical database that is kept confidential from the investigators, Dr. McShane said it was possible they were not.
“I am aware of allegations that they were not followed in at least one of the trials,” she said. “And Duke officials are aware of that and have investigated. I don’t know the details of exactly what was found in the investigation, but that is a real issue.”
Because the community relies on such trials to prove or disprove the value of a predictive signature, the integrity of the data is critical. “If you cannot rely on the data in the trial having been collected in a way that is free of errors, free of potential manipulation, that is a very serious problem,” Dr. McShane said.
When pressed for further details of the allegations, Dr. McShane requested that the committee first request the information from Duke officials. “I would urge you to first try to get the information from Duke because they will know not only what the allegations were, but whether they were able to confirm or refute the allegations.”
She noted that just because someone might have had access to the data who shouldn’t have, “doesn’t necessarily mean there were actual alterations or that the information was used in any inappropriate way.”
Further, Duke University officials, particularly Robert M. Califf, MD, Director of the Duke Translational Medicine Institute and Vice Chancellor for Clinical Research, would know the details. “I know for a fact that he is well aware of this,” she said.
“In terms of our releasing the information I received, there is quite frankly a lot of detail in there -- names named, very troubling accusations made -- and I would prefer not to release that because it is so sensitive,” she continued. If the committee were unable to get further information from Duke, Dr McShane said she would consult NCI leadership and legal counsel regarding what she should and could share.
Complexity Obscures Standards
The chair of the committee, Gilbert S. Omenn, MD, PhD, Director of the Center for Computational Medicine and Biology at the University of Michigan Medical School, pointed out that the complexity of ‘omics sometimes seems to blind some investigators to the necessities of clinical trial procedures.
“I think it is ironic that there has been so much discussion about how the new ‘omics exploratory analysis is unfamiliar to some people, a lot of people, and is considered to be complex,” he said, “and, therefore, needs to be developed through special genomics entities, whether companies, departments, divisions, centers, or institutions.
“And then people lose track that if you are going to do clinical trials, there is a very well established structure, process, oversight, rules and regulations, and standards of performance for clinical trials,” he said.
“Maybe one of the biggest recommendations we need to make is that new technology platforms need to be brought into existing -- preferably existing, well-run -- clinical trial operations,” Dr. Omenn said.
Duke IRB Alerted Early
During her discussion with the committee yesterday, which ran for just over two hours, Dr. McShane said that her group alerted Duke leadership early on that there might be problems.
“When we made our original contact with Duke, it was to Kim Lyerly, the head of the Cancer Center,” Dr. McShane said. “We did that because we assumed they would be cancer center-funded trials. In fact, they weren’t. So what Kim did immediately was to contact the relevant officials at Duke University Medical School, namely the deans, and they put us in touch with the IRB chair there.”
That comment prompted a quick response from Dr. Omenn. “That is important, because we have been wondering all along if all of the communication went to just Potti and [his mentor and co-author Joseph] Nevins, or to higher ups, or other co-authors,” he said. “Here is an example where it went to the IRB, not to the investigators.”
Although that statement makes clear that the institutional review board (IRB), which is charged with ensuring patient safety, knew about potential problems, neither Dr. McShane nor committee members spoke further during the public session yesterday, regarding what the IRB should have done with that information – or what might have prevented them from pursuing it more rigorously than they did.
I’ve mentioned in previous posts that institutional, as well as personal, conflict of interest could have played a role in slowing Duke’s response.
Committee member Nathan Price, PhD, of the University of Illinois in Urbana-Champaign, made that point emphatically during a discussion regarding what government agencies have oversight responsibilities and opportunities in ‘omics-based trials.
“The thing that seems to me to be the big problem with the Duke situation overall is that we have the same people on both sides of the situation,” Dr. Price said.
“The people who are deciding whether or not this [predictor] has been developed correctly and is an independent test and should go forward [to a clinical trial], are the same people who stand to earn millions of dollars off of the fact that it works out.”
Further Information from Dr. McShane Expected
This was the second time Dr. McShane spoke before the full ‘omics committee, but her third contribution overall. In December, she described the work she’d done to try to nail down the details and problems with Dr. Potti’s work. However, she also met with a small group of committee members on June 21, in what Dr. Omenn described as an “extraordinary meeting.”
Unfortunately further details about that discussion are not available. An Institute of Medicine media representative described it as an opportunity for a small group of committee members and Dr. McShane to discuss background material related to this week’s presentation. A written document is expected to be made public from these two presentations in the next few weeks.
Will Duke Weigh In?
At the previous committee meeting in March, Dr. Omenn mentioned that he hoped Dr. Califf, who was in attendance at the time, would address the committee at this week’s meeting. That did not happen. Whether he or another Duke official will do so is unclear.
The media representative who responded to email inquiries declined to say anything specific about future meetings and none have been publicly announced thus far.