William Auffermann MD, PhD, Presenter: Nothing to Disclose

Brent Little MD, Abstract Co-Author: Nothing to Disclose

Travis S. Henry MD, Abstract Co-Author: Spouse, Employee, F. Hoffmann-La Roche Ltd

Stefan Tigges MD, Abstract Co-Author: Stockholder, Microsoft Corporation Stockholder, General Electric Company

Srini Tridandapani PhD, MD, Abstract Co-Author: Nothing to Disclose

Many chest radiograph interpretation errors occur because observer's eyes do not fixate on a relevant abnormality, which may be partially due to a lack of a search strategy. This project's goal is to determine if providing a standardized chest radiograph search pattern to medical trainees improves performance on a nodule identification task.

Healthcare trainee participants included: medical and nurse practitioner students, and non-radiology residents and fellows. The image set included 50 chest radiographs; 25 normals and 25 with pulmonary nodules (one nodule per image, size range 10-20 mm). Subjects were divided into control and treatment groups, and were shown the 50 radiographs randomized into two case sets of equal size. Subjects were asked to mark nodules if present and give a rating of their diagnostic confidence using a 5 point scale. The control group received no training between case sets. The treatment group received chest radiograph search pattern training between the first and second case sets. Performance on nodule identification between control and treatment groups were compared using localization receiver operator characteristic (LROC) analysis and differences in the area under the LROC curve (D-AUC). P-values of less than 0.05 were considered statistically significant.

Control Group: There was no statistically significant difference in subject performance between case sets 1 and 2, D-AUC = 0.0521, p-value = 0.1910. Treatment group: There was a statistically significant improvement in subject performance between case sets 1 and 2, D-AUC = 0.1079, p-value = 0.0150.

Providing healthcare trainees with a chest radiograph search pattern improved performance on a nodule identification task.

Search pattern training may be useful for healthcare trainees who will evaluate chest radiographs as part of their practice.

Auffermann, W, Little, B, Henry, T, Tigges, S, Tridandapani, S, Search Pattern Training for Improving Pulmonary Nodule Identification on Chest Radiographs.  Radiological Society of North America 2014 Scientific Assembly and Annual Meeting, - ,Chicago IL. http://archive.rsna.org/2014/14011127.html