Dual-energy, standard and low-kVp contrast-enhanced CT-cholangiography: A comparative analysis of image quality and radiation exposure

Abstract Objective Quantitative image quality assessment in terms of image noise (IN), contrast-to-noise ratio (CNR), and signal-to-noise ratio (SNR) in relation to required radiation dose (RD) for dual-energy (DE), standard and low-kVp contrast-enhanced computed-tomography (CT) cholangiography. Materials and methods For each of 22 DECT-cholangiography examinations, 3 image datasets were analyzed as independent single-source CT-acquisitions at different tube potential, i.e. 80 kVp, 120 kVp-equivalent (linear blended dataset M0.3: 30% 80 kVp, 70% 140 kVp), and 140 kVp. Analysis comprised determination of IN, CNR and SNR in regions of interest (ROI) placed in liver parenchyma and contrasted bile ducts. IN was evaluated as mean standard deviation of 3 ROI placed within liver parenchyma (segments 6/7, 5/8, 2/3); CNR was assessed as bile duct-to-liver parenchyma ratio, and SNR as bile duct-to-image noise ratio. RD in terms of CT dose index (CTDIvol ), dose-length product (DLP) and effective dose (ED) has been determined for each of the datasets, and compared to console prediction and scan summary values. Using phantom measurements of CTDIvol , a method for separating comprehensive RD values of DE-acquisitions into the original RD contribution of each tube (80 kVp/140 kVp) has been developed, enabling comparison of all 3 datasets as if independently acquired using single-source “single-energy” technique. Results Highest IN was detected for 80 kVp- (38.6 ± 5.1 HU), lowest for 120 kVp-equivalent linear blended M0.3-datasets (23.1 ± 3.4 HU) with significant differences between all datasets ( P < 0.001). Highest SNR and CNR were measured for M0.3- (SNR: 14.8 ± 4.1; CNR: 11.6 ± 3.8) and 80 kVp-datasets (SNR: 13.8 ± 4.8; CNR: 11.2 ± 4.5); lowest for 140 kVp-datasets (SNR: 9.5 ± 2.5; CNR: 7.1 ± 2.3) with significant differences between M0.3- and 140 kVp-datasets as well as between 80 kVp- and 140kVp-datasets (both P < 0.001 for both CNR, SNR). CTDIvol , DLP and ED were reduced by 50% for low-kilovoltage acquisitions (CTDIvol : 5.5 ± 1.4 mGy; DLP: 127.8 ± 40.1 mGy cm; ED: 1.9 ± 0.6 mSv) compared to comprehensive DE-acquisitions (CTDIvol : 11.0 ± 2.3 mGy; DLP: 253.8 ± 67.5 mGy cm; ED: 3.8 ± 1.0 mSv, tube contribution: 80 kVp: 44.5%; 140 kVp: 55.5%), and by 20% compared to conventional acquisitions at 120 kVp (CTDIvol : 6.71 mGy; DLP: 153.5 ± 16.9 mGy cm; ED: 2.3 ± 0.3 mSv). Conclusions Despite higher IN, low-kilovoltage CT-cholangiography reveals no significant differenc