Daniel Matheson Adams MD, Presenter: Nothing to Disclose

C. Gregory Elliott MD, Abstract Co-Author: Nothing to Disclose

Scott Stevens MD, Abstract Co-Author: Nothing to Disclose

Scott Woller MD, Abstract Co-Author: Nothing to Disclose

Scott Evans PhD, Abstract Co-Author: Nothing to Disclose

Jim Lloyd BS, Abstract Co-Author: Nothing to Disclose

Greg Snow PhD, Abstract Co-Author: Nothing to Disclose

Todd Allen MD, Abstract Co-Author: Nothing to Disclose

Joseph Bledsoe MD, Abstract Co-Author: Nothing to Disclose

Todd Delton Lovelace MD, Abstract Co-Author: Nothing to Disclose

Valerie Aston RT, Abstract Co-Author: Nothing to Disclose

CT pulmonary angiography (CTPA) is frequently used for diagnosis of suspected pulmonary embolism (PE). The rate of false positive and false negative studies increases when the pretest probability for PE is discordant with the results of this test. We measured rates of inaccurate diagnosis of suspected PE for CTPA as interpreted by general vs. chest radiologists and stratified the results by pretest probability.

This retrospective study was conducted in the Emergency Departments at Intermountain Medical Center and LDS Hospital in Salt Lake City, Utah. All CTPAs from July 1, 2009 to April 30, 2010 were identified through a query of the Intermountain Enterprise Data Warehouse. Studies originally interpreted by the general radiology group were blindly re-interpreted by one of three fellowship-trained chest radiologists. The chest radiologists’ interpretations were used as the standard to measure the rates of false-positive and false-negative diagnosis. Kappa coefficients for agreement amongst the chest radiologists were calculated by overlapping 20% of the sample. Pretest probability was calculated for each patient using a Revised Geneva Score (RGS). The patients were stratified based on pretest probability and the rates of inaccurate diagnoses were reported for each group.

241 consecutive positive studies and 241 randomly selected negative studies were distributed to the study radiologists and 99 studies overlapped with two study radiologists. Thirteen/298 studies (4.4%) were false-positive and 7/283 studies (2.5%) were false-negative. The chest radiologists’ Kappa coefficients were 1.0, 0.94, and 0.94. The pretest probabilities of the study sample by RGS were low probability 25%, intermediate probability 64%, and high probability 11%. False positive rates for low, intermediate, and high probability were 3.1%, 5.3%, and 0%. False negative rates for the same groups were 6.0%, 2.2%, and 0% respectively.

CTPA carries risk for inaccurate diagnosis. Our results showed that this risk of inaccuracy varied according to the pre-test probability for PE. 

The rates of inaccurate diagnosis of pulmonary embolism with CT pulmonary angiography vary with pre-test probability for disease. A pre-test probability assessment is recommended prior to imaging.

Adams, D, Elliott, C, Stevens, S, Woller, S, Evans, S, Lloyd, J, Snow, G, Allen, T, Bledsoe, J, Lovelace, T, Aston, V, Diagnostic Accuracy of CT Pulmonary Angiography for Suspected Pulmonary Embolism in Relation to the Pre-test Probability of Disease.  Radiological Society of North America 2011 Scientific Assembly and Annual Meeting, November 26 - December 2, 2011 ,Chicago IL. http://archive.rsna.org/2011/11034293.html