Erschienen in:
25.02.2019 | Imaging Informatics and Artificial Intelligence
Modified calcium subtraction in dual-energy CT angiography of the lower extremity runoff: impact on diagnostic accuracy for stenosis detection
verfasst von:
Domenico De Santis, Carlo N. De Cecco, U. Joseph Schoepf, John W. Nance, Ricardo T. Yamada, Brooke A. Thomas, Katharina Otani, Brian E. Jacobs, D. Alan Turner, Julian L. Wichmann, Marwen Eid, Akos Varga-Szemes, Damiano Caruso, Katharine L. Grant, Bernhard Schmidt, Thomas J. Vogl, Andrea Laghi, Moritz H. Albrecht
Erschienen in:
European Radiology
|
Ausgabe 9/2019
Einloggen, um Zugang zu erhalten
Abstract
Objectives
To investigate the diagnostic accuracy of a modified three-material decomposition calcium subtraction (CS) algorithm for the detection of arterial stenosis in dual-energy CT angiography (DE-CTA) of the lower extremity runoff compared to standard image reconstruction, using digital subtraction angiography (DSA) as the reference standard.
Methods
Eighty-eight patients (53 males; mean age, 65.9 ± 11 years) with suspected peripheral arterial disease (PAD) who had undergone a DE-CTA examination of the lower extremity runoff between May 2014 and May 2015 were included in this IRB-approved, HIPAA-compliant retrospective study. Standard linearly blended and CS images were reconstructed and vascular contrast-to-noise ratios (CNR) were calculated. Two independent observers assessed subjective image quality using a 5-point Likert scale. Diagnostic accuracy for ≥ 50% stenosis detection was analyzed in a subgroup of 45 patients who had undergone additional DSA. Diagnostic accuracy parameters were estimated with a random-effects logistic regression analysis and compared using generalized estimating equations.
Results
CS datasets showed higher CNR (15.3 ± 7.3) compared to standard reconstructions (13.5 ± 6.5, p < 0.001). Both reconstructions showed comparable qualitative image quality scores (CS, 4.64; standard, 4.57; p = 0.220). Diagnostic accuracy (sensitivity, specificity, positive and negative predictive values) for CS reconstructions was 96.5% (97.5%, 95.6%, 90.9%, 98.1) and 93.1% (98.8%, 90.4%, 82.3%, 99.1%) for standard images.
Conclusions
A modified three-material decomposition CS algorithm provides increased vascular CNR, equivalent qualitative image quality, and greater diagnostic accuracy for the detection of significant arterial stenosis of the lower extremity runoff on DE-CTA compared with standard image reconstruction.
Key Points
• Calcified plaques may lead to overestimation of stenosis severity and false positive results, requiring additional invasive digital subtraction angiography (DSA).
• A modified three-material decomposition algorithm for calcium subtraction provides greater diagnostic accuracy for the detection of significant arterial stenosis of the lower extremity runoff compared with standard image reconstruction.
• The application of this algorithm in patients with heavily calcified vessels may be helpful to potentially reduce inconclusive CT angiography examinations and the need for subsequent invasive DSA.