Evaluation of image quality, radiation dose and diagnostic performance of dual-energy CT datasets in patients with hepatocellular carcinoma

Aim To evaluate image quality and diagnostic accuracy of different dual-energy computed tomography (DECT) datasets for identification of hepatocellular carcinoma (HCC), assess the reliability of virtual unenhanced (VU) images in replacing standard unenhanced (SU) images, and quantify effective dose (ED) at different tube voltages. Material and methods Thirty cirrhotic patients underwent liver contrast-enhanced DECT. Two blinded observers retrospectively evaluated conventional unenhanced and VU images, 140 kVp/80 kVp/mixed tube potential arterial datasets and conventional portal-venous/late phases in consensus. Final diagnosis was based on pathological proof or imaging criteria. Image quality, ED, sensitivity, and specificity of arterial datasets were calculated. Results Thirty-eight HCC and 18 benign lesions were detected at 80 kVp, 33 HCC and 22 benign lesions were detected at 140 kVp, and 36 HCC and 20 benign lesions were detected at mixed tube potentials. Final diagnosis confirmed 37 HCC and 20 benign lesions. There was no significant difference in diagnostic confidence between 80 kVp, 140 kVp, and mixed tube potential arterial datasets (p>0.05). Image quality was adequate for all datasets, with increased quality at higher tube potential (80 versus 140 kVp, p=0.001; mixed versus 140 kVp, p=0.001; 80 kVp versus mixed, p=0.0024). Significant ED reduction was observed between 140 and 80 kVp datasets (p<0.001). Conclusions The 140 kVp dataset provided higher image quality. The 80 kVp images were more sensitive in detecting HCC. VU images are adequate in replacing SU images. The ED of the 80 kVp dataset was significantly lower.