Abstract Archives of the RSNA, 2007
Marco Das MD, Presenter: Nothing to Disclose
Georg Mühlenbruch MD, Abstract Co-Author: Nothing to Disclose
Marcos Salganicoff PhD, Abstract Co-Author: Employee, Siemens AG
Annemarie Bakai PhD, Abstract Co-Author: Employee, Siemens AG
Rolf W. Günther, Abstract Co-Author: Nothing to Disclose
Joachim Ernst Wildberger MD, Abstract Co-Author: Nothing to Disclose
To evaluate the influence of CAD on radiologists’ detection performance for pulmonary embolism (PE) in routinely acquired multidetector-row CT (MDCT) chest examinations.
42 patients with suspected PE received a routine 16-slice MDCT scan to rule out PE (120 kV, 100 mAseff., 16x0.75 mm col., 1/0.8 mm slice reconstruction, 100 ml cm with bolus tracking). A CAD software prototype (PE CAD, Siemens Version .7) was applied to all datasets. Three radiologists (R1-3) with different levels of experience read all cases first without and then with CAD as a second reader. All thrombi were marked at the vessel level (lobar, segmental, subsegmental). All marked true positive findings were verified and were set as the standard of reference. For all three readers sensitivity for the different vessel levels with and without CAD was assessed and compared to the reference standard.
32 patients were positive for PE. The reference standard contained 215 thrombi. Sensitivity for R1 alone: all 87%, lobar 97%, segmental 93%, subsegmental 80%. R2 alone: all 82%, lobar 94%, segmental 92%, subsegmental 81%. R3 alone: Sensitivity all 77%, lobar 93%, segmental 82%, subsegmental 63%. With the use of CAD sensitivities increased significantly on the segmental and subsegmental level for all three readers (p<0.001). Overall Sensitivity of the CAD software alone: 83% (lobar: 87% segmental: 90% subsegmental: 77%) The median false positive rate per case was 4.
The use of CAD in the detection of PE shows significant improvement of radiologist’s detection performance and can be beneficial to decrease false negative diagnosis.
The use of CAD for pulmonary embolism significantly improves PE detection.
Das, M,
Mühlenbruch, G,
Salganicoff, M,
Bakai, A,
Günther, R,
Wildberger, J,
Increased Detection Performance of Pulmonary Embolism with the Use of Computer-aided Detection. Radiological Society of North America 2007 Scientific Assembly and Annual Meeting, November 25 - November 30, 2007 ,Chicago IL.
http://archive.rsna.org/2007/5010928.html