The Radiological Devices Panel met on February 3, 2004 at the Gaithersburg Marriott, 9751 Washingtonian Boulevard, Gaithersburg, MD. The panel met to discuss and vote on an original premarket approval (PMA) application, P030012 from R2 Technology, Inc. of Sunnyvale, California for their ImageChecker CT CAD Software System. The device is a Computer-Aided Detection (CAD) system designed to assist a physician in the identification of solid nodules in CT images of the lung. It is intended to be a “second reader” that alerts the user to regions of interest that may have been initially overlooked.
The open session of the meeting began by noting that the panel chair, Dr. Minesh Mehta was prevented by weather in the mid-west from attending in person and that Dr. Geoffrey Ibbott was appointed to take his place. Dr. Robert Phillips (Branch Chief, Radiology Devices Branch) then updated the panel regarding matters since the last Radiological Devices Panel meeting in December of 2002. This was followed with a tutorial presentation of Contemporary ROC Methods by Dr. Robert F. Wagner of the FDA’s Office of Science and Technology. The morning session was concluded with R2 Technology presenting a summary of their PMA for the ImageChecker CT CAD Software System.
The afternoon session began with the FDA presentation of their views of the data in the PMA application. This was followed by brief presentations from the two lead reviewers from the panel, Dr. David Stark – clinical aspects and Dr. Brent Blumenstein – statistics. The presentations discussed the methodologies used and the results obtained from studies performed by the sponsor. The major focus of these concerned the ROC analyses and how these relate to clinical use of the device.
The panel discussed the FDA questions and the data presented by the sponsor and the FDA. Following the discussion, the panel voted on an initial motion to recommend “not approvable” it was defeated 2 For, 5 Against and 1 Abstention. Subsequently, a motion to recommend "approvable with conditions" passed with a vote of: 7 For, 0 Against and 1 Abstention.
The conditions the panel recommended included: 1) A reanalysis of the data prior to approval to assess clinical significance, 2) The inclusion of strict rules for use in the labeling, 3) A requirement for formalized training, and 4) Post-market surveillance for more data on clinical significance.
Contact: Robert J. Doyle, Executive Secretary,
Transcripts may be purchased from: (written requests only)
Neal R. Gross & Company
1323 Rhode Island Avenue, N.W.
Washington, D.C. 20005
(202) 234-4433 (voice), (202) 387-7330 (fax)
The Food and Drug Administration
Freedom of Information Staff (FOI)
5600 Fishers Lane, HFI-35
Rockville, MD 20852
301-827-6500 (voice) or 301-443-1726 (FAX)