David R. Holmes PhD, Presenter: Nothing to Disclose

Rickey Carter PhD, Abstract Co-Author: Nothing to Disclose

Cynthia H. McCollough PhD, Abstract Co-Author: Research Grant, Siemens AG

Maria Shiung, Abstract Co-Author: Nothing to Disclose

Lifeng Yu PhD, Abstract Co-Author: Nothing to Disclose

Kurt Ernest Augustine MS, Abstract Co-Author: Nothing to Disclose

Joel Garland Fletcher MD, Abstract Co-Author: Grant, Siemens AG

Norbert Gilles Joseph Campeau MD, Abstract Co-Author: Nothing to Disclose

Amy Louise Kotsenas MD, Abstract Co-Author: Nothing to Disclose

Sudhakar Kundapur Venkatesh MD, FRCR, Abstract Co-Author: Nothing to Disclose

Naoki Takahashi MD, Abstract Co-Author: Nothing to Disclose

David Maitland Hough MD, Abstract Co-Author: Nothing to Disclose

Anne-Marie Gisele Sykes MD, Abstract Co-Author: Nothing to Disclose

Phillip Edwards, Abstract Co-Author: Nothing to Disclose

Vance Lehman MD, Abstract Co-Author: Nothing to Disclose

Observer performance studies in low dose CT are hampered primarily by the expense and time required for radiologist evaluation. We sought to determine if a workstation designed to facilitate reader studies in low dose CT would result in reader times of less than 5 minutes for common diagnostic tasks in CT.

We constructed an open source workstation that permits automatic anonymization & randomization of CT datasets; customizable reader surveys for organ location, diagnosis, and image quality; tools for circumscription of pathological findings; viewing and correlation of axial and non-axial images; local image database for efficient retrieval; and server-based results database for statistical export. For each of 3 diagnostic tasks (abdominal CT for liver metastasis, head CT for acute neurologic deficit, chest CT for nodule detection), 3 subspecialized radiologists examined CT images of varying dose levels and reconstruction methods as part of an ongoing study. Readers were required to circumscribe all pathologic findings, give presumptive diagnoses and levels of confidence, and complete a brief image quality survey for each case. Reader times were automatically recorded for each case, with mean reader times and interquartile ranges (IQR) reported for each diagnostic task.

Three subspecialized radiologists interpreting abdominal, head and chest CT each interpreted a mean of 339, 354, and 234 cases, respectively. Mean reading times for abdominal and neurological radiologists was 2.80 ± 0.82 and 3.05 ± 0.63 minutes, respectively (abdominal IQR: 1.37 – 3.08; head IQR: 1.92 – 2.48). Mean reading times for chest radiologists was 4.48 ± 1.81 minutes (IQR: 2.5 – 5.5), with 2/3 radiologists having mean interpretation times between 5 and 6 minutes. For each diagnostic task, reading times varied significantly between pairs of readers (p < 0.004).

A workstation that minimizes radiologist effort can drastically reduce reader times to 5 minutes or less for common diagnostic tasks in low dose CT, and facilitate performance of multi-reader, multi-case studies.  Interpretation times can still vary markedly depending on diagnostic task and individual readers. 

Dedicated workstations designed for low dose CT studies can reduce interpretation times to approximately 5 minutes or less and facilitate rapid completion and substantial cost savings for observer performance studies.

Holmes, D, Carter, R, McCollough, C, Shiung, M, Yu, L, Augustine, K, Fletcher, J, Campeau, N, Kotsenas, A, Venkatesh, S, Takahashi, N, Hough, D, Sykes, A, Edwards, P, Lehman, V, Reader Efficiency Using a Dedicated Workstation for Rapid Observer Performance Studies in Low-dose CT.  Radiological Society of North America 2014 Scientific Assembly and Annual Meeting, - ,Chicago IL. http://archive.rsna.org/2014/14017887.html