Anil Potti

Trial participants and their representatives brought eight lawsuits against Duke and Potti, alleging that patients had been exposed to unnecessary or harmful chemotherapy based on flawed research. The cases were settled in 2015 on confidential terms.

State of the evidence

Settlement

Attribution

Accused

The HHS Office of Research Integrity concluded that Potti had committed scientific misconduct in NIH-funded research. Findings included reversed treatment-response statuses, switched recurrence phenotypes and altered values, as well as false data in grant applications and nine publications.

State of the evidence

Documented

Attribution

Directly implicated

The North Carolina Medical Board issued a reprimand and consent order concerning, among other matters, inaccuracies in his professional background; Missouri subsequently issued its own reprimand based on the North Carolina action.

State of the evidence

Documented

Attribution

Directly implicated

Potti resigned from Duke on 19 November 2010. The three trials with suspended enrolment were closed and retraction procedures began, including for the Nature Medicine paper; several major publications were subsequently withdrawn.

State of the evidence

Documented

Attribution

Directly implicated

Duke launched several phase II trials in which genomic signatures linked to the Potti/Nevins research were used to guide treatment. A total of 117 patients were eventually enrolled in the three Duke trials concerned.

Attribution

Directly implicated

Duke temporarily suspended the trials in 2009 following concerns raised by the NCI, then resumed them after an external statistical review. Subsequent Institute of Medicine/National Academies analyses found this review seriously deficient, as reviewers had not received several essential pieces of information.

State of the evidence

Documented

The Cancer Letter established that Potti had repeatedly described himself as a 1995 or 1996 Rhodes Scholar in biographical sketches and grant applications, although the Rhodes Trust had no record of such a scholarship. An American Cancer Society grant worth $729,000 contained this claim.

Attribution

Directly implicated

Duke placed Potti on administrative leave on 16 July 2010. Two days later, investigators suspended new enrolments in three trials; in August, the university announced that its review of his CV had identified substantial problems and led to sanctions.

State of the evidence

Documented

Attribution

Directly implicated

From 2006, Potti and his colleagues published research promising to predict sensitivity to different chemotherapy drugs. Keith Baggerly and Kevin Coombes quickly identified incorrect gene lists and figures, problems separating training and test datasets, and an inability to reproduce the reported performance.

State of the evidence

Documented

Attribution

Directly implicated