Individualized Scan Protocols in Abdominal Computed Tomography

Abstract

Background: In contrast-enhanced abdominal computed tomography (CT), radiation and contrast media (CM) injection protocols are closely linked to each other, and therefore a combination is the basis for achieving optimal image quality. However, most studies focus on optimizing one or the other parameter separately. Purpose: Reducing radiation dose may be most important for a young patient or a population in need of repetitive scanning, whereas CM reduction might be key in a population with insufficient renal function. The recently introduced technical solution, in the form of an automated tube voltage selection (ATVS) slider, might be helpful in this respect. The aim of the current study was to systematically evaluate feasibility of optimizing either radiation or CM dose in abdominal imaging compared with a combined approach. Methods: Six Göttingen minipigs (mean weight, 38.9 ± 4.8 kg) were scanned on a third-generation dual-source CT. Automated tube voltage selection and automated tube current modulation techniques were used, with quality reference values of 120 kV ref and 210 mAs ref . Automated tube voltage selection was set at 90 kV semimode. Three different abdominal scan and CM protocols were compared intraindividually: (1) the standard "combined" protocol, with the ATVS slider position set at 7 and a body weight-adapted CM injection protocol of 350 mg I/kg body weight, iodine delivery rate (IDR) of 1.1 g I/s; (2) the CM dose-saving protocol, with the ATVS slider set at 3 and CM dose lowered to 294 mg I/kg, resulting in a lower IDR of 0.9 g I/s; (3) the radiation dose-saving protocol, with the ATVS slider position set at 11 and a CM dose of 441 mg I/kg and an IDR 1.3 g I/s, respectively. Scans were performed with each protocol in arterial, portal venous, and delayed phase. Objective image quality was evaluated by measuring the attenuation in Hounsfield units, signal-to-noise ratio, and contrast-to-noise ratio of the liver parenchyma. The overall image quality, contrast quality, noise, and lesion detection capability were rated on a 5-point Likert scale (1 = excellent, 5 = very poor). Protocols were compared for objective image quality parameters using 1-way analysis of variance and for subjective image quality parameters using Friedman test. Results: The mean radiation doses were 5.2 ± 1.7 mGy for the standard protocol, 7.1 ± 2.0 mGy for the CM dose-saving protocol, and 3.8 ± 0.4 mGy for the radiation dose-saving protocol. The mean total iodine load in these groups was 13.7 ± 1.7, 11.4 ± 1.4, and 17.2 ± 2.1 g, respectively. No significant differences in subjective overall image or contrast quality were found. Signal-to-noise ratio and contrast-to-noise ratio were not significantly different between protocols in any scan phase. Significantly more noise was seen when using the radiation dose-saving protocol ( P < 0.01). In portal venous and delayed phases, the mean attenuation of the liver parenchyma significantly differed between protocols ( P < 0.001). Lesion detection was significantly better in portal veno...

Individualized scan protocols for CT angiography: an animal study for contrast media or radiation dose optimization

“…Our study goes a step further and demonstrates that the CTA optimization approach is independent of the kV setting for both CM and radiation dose optimization. It thus complements the results of Martens et al [21], showing that the combination of ATVS settings and CM injection parameters can be used to optimize the radiation dose or amount of CM for parenchymal liver CT at identical kV. In comparison, greater savings were obtained in CM (-26% versus -16%) or radiation dose (-30% versus -26%) for the vascular contrast in our study.…”

A New Algorithm for Automatically Calculating Noise, Spatial Resolution, and Contrast Image Quality Metrics

“…Nevertheless, these results are encouraging and future studies can be undertaken to implement the algorithm for different anatomies and slice reconstructions, as well as to evaluate IQ metric performance when using CT scanner settings that might be expected to interfere. The algorithm can be further developed into a continuous IQ monitoring tool that may prove invaluable in radiation dose and contrast injection protocol optimization 29,30 or in retake scan decision making in daily clinical practice 31 …”

Contrast Media Reduction in Computed Tomography With Deep Learning Using a Generative Adversarial Network in an Experimental Animal Study

“…Although the vascular contrast, particularly in the pulmonary and arterial contrast phases, can be significantly reduced by a combination of adapted injection protocols and low-voltage CT, 9,21 the potential savings with these methods for parenchymal imaging are much lower because of the lower intrinsic contrast enhancement of the parenchyma. Automatic tube voltage selection can reduce the contrast dose by 17% while maintaining CNR and increasing the radiation dose 24 . Alternatively, the contrast of the liver parenchyma in the venous phases can be optimized by adjusting the injection protocol by using dual-flow protocols with different contrast media concentrations 10 …”

“…Automatic tube voltage selection can reduce the contrast dose by 17% while maintaining CNR and increasing the radiation dose. 24 Alternatively, the contrast of the liver parenchyma in the venous phases can be optimized by adjusting the injection protocol by using dual-flow protocols with different contrast media concentrations. 10 Besides a similar image contrast, the subjective image quality is important and in particular the pathological consistency of the virtual CM images.…”