Virtual Monoenergetic Imaging and Iodine Perfusion Maps Improve Diagnostic Accuracy of Dual-Energy Computed Tomography Pulmonary Angiography With Suboptimal Contrast Attenuation.
Abstract
OBJECTIVES:
The aim of this study was to investigate the impact of virtual monoenergetic imaging (VMI+) and dual-energy computed tomography perfusion maps (DECT-PMs) on reader confidence and diagnostic accuracy in dual-energy computed tomography pulmonary angiography (DE-CTPA) studies with suboptimal contrast attenuation, compared with standard linearly blended reconstruction series.
MATERIALS AND METHODS:
Dual-energy computed tomography pulmonary angiography examinations with suboptimal contrast attenuation of 68 patients with suspected pulmonary embolism (PE) were included in this institutional review board-approved retrospective study. Virtual monoenergetic imaging series at 40 keV, DECT-PM, and linearly blended images (M_0.6, 60% 90-kV spectrum) were reconstructed. Contrast-to-noise ratio and signal-to-noise ratio within the pulmonary trunk were calculated. Four independent radiologists assessed the presence of PE and their diagnostic confidence using 3 DE-CTPA reconstruction protocols: protocol 1, M_0.6 images; protocol 2, M_0.6 series and DECT-PM; and protocol 3, M_0.6, DECT-PM, and VMI+ series. Receiver operating characteristic (ROC) analysis was performed.
RESULTS:
Fourteen patients showed central and 29 segmental PE. Greater contrast-to-noise ratio and signal-to-noise ratio values were measured in VMI+ series at 40 keV in comparison to M_0.6 images (P < 0.001). Diagnostic accuracy for segmental PE detection was as follows: protocol 1 (69.1%); protocol 2 (86.8%); and protocol 3 (92.6%). Protocol 3 resulted in a significantly greater area under the curve for diagnosing segmental PE (0.991, P ≤ 0.033), compared with protocol 1 and 2 (0.897 and 0.951, respectively), and provided the highest diagnostic confidence (P < 0.001).
CONCLUSIONS:
A reconstruction protocol including 40-keV VMI+ series and DECT-PM improves reader confidence and diagnostic accuracy for segmental PE detection compared with standard M_0.6 images in DE-CTPA with suboptimal contrast attenuation.