OBJECTIVE: To assess the accuracy of a semiautomated 3-dimensional (3D) computed tomographic angiography software in the evaluation of lower limb stenoses compared to reader evaluation, using digital subtraction angiography (DSA) as standard of reference. MATERIALS AND METHODS: Forty patients with peripheral vascular occlusive disease underwent both DSA and computed tomographic angiography. The vascular tree was divided into 6 segments from distal aorta to popliteal artery. Each district was evaluated for significant stenosis by one experienced vascular radiologist (on axial as well as 3D images) and by a semiautomated 3D software analysis, independently. Evaluation of the vessel evaluation was then repeated by a poorly experienced radiologist twice, first without 3D software analysis and then supported by 3D software analysis. RESULTS: Both experienced radiologist and automated evaluations obtained high statistical results when compared to DSA. The analysis by poorly experienced radiologist obtained lower statistical results, which significantly improved when supported by 3D software analysis. CONCLUSIONS: Three-dimensional analysis software should be feasible to identify significant vascular stenoses and may help a poorly experienced radiologist to significantly improve diagnostic accuracy.
OBJECTIVE: To assess the accuracy of a semiautomated 3-dimensional (3D) computed tomographic angiography software in the evaluation of lower limb stenoses compared to reader evaluation, using digital subtraction angiography (DSA) as standard of reference. MATERIALS AND METHODS: Forty patients with peripheral vascular occlusive disease underwent both DSA and computed tomographic angiography. The vascular tree was divided into 6 segments from distal aorta to popliteal artery. Each district was evaluated for significant stenosis by one experienced vascular radiologist (on axial as well as 3D images) and by a semiautomated 3D software analysis, independently. Evaluation of the vessel evaluation was then repeated by a poorly experienced radiologist twice, first without 3D software analysis and then supported by 3D software analysis. RESULTS: Both experienced radiologist and automated evaluations obtained high statistical results when compared to DSA. The analysis by poorly experienced radiologist obtained lower statistical results, which significantly improved when supported by 3D software analysis. CONCLUSIONS: Three-dimensional analysis software should be feasible to identify significant vascular stenoses and may help a poorly experienced radiologist to significantly improve diagnostic accuracy.