ABSTRACT
<p><b>OBJECTIVE</b>To investigate the feasibility of virtual non-enhanced images in evaluating the spontaneous subarachnoid hemorrhage (SAH) by dual-energy computed tomography angiography.</p><p><b>METHODS</b>Dual-energy computed tomography angiography was performed in 43 SAH patients.Virtual non-enhanced images were obtained by using Liver VNC software. paired t-test was performed to compare the signal to noise ratio between the conventional plain scan and virtual non-enhanced images. Diagnostic accuracy for SAH by virtual non-enhanced images was calculated by using the conventional plain scan images as a gold standard. Quality score was calculated to evaluate whether virtual non-enhanced images can meet the imaging requirements of SAH.</p><p><b>RESULTS</b>The signal to noise ratio was 8.63?0.53 among plain scan images and 3.96?0.52 among virtual non-enhanced images ( t=43.18,P=0.000) . The sensitivity, specificity, and accuracy of virtual non-enhanced imaging in diagnosing the SAH were 97.05%, 100%, and 97.67% in per-patient analysis, and were 94.64%, 100%, and 98.97% in per-lesion analysis. The quality scores were 3 in six patients, 2 in 27 patients, and 1 in two patients.</p><p><b>CONCLUSIONS</b>Virtual non-enhanced images can meet the clinical requirements of diagnosis, although it has more intensive noise than conventional plain scan images. Furthermore, it has higher accuracy in evaluating SAH.</p>