Can Dual-energy CT-based Virtual Monoenergetic Imaging Improve the Assessment of Hypodense Liver Metastases in Patients With Hepatic Steatosis?

RATIONALE AND OBJECTIVES: To evaluate the impact of noise-optimized virtual monoenergetic imaging (VMI) on lesion demarcation and measuring accuracy of hypoattenuating liver metastases in patients with fatty liver disease compared to standard reconstructions. MATERIALS AND METHODS: Twenty-eight patients (mean age 62.2 ± 7.7 years) with fatty liver disease and hypoattenuating liver metastases who underwent unenhanced and contrast-enhanced portal-venous dual-energy CT (DECT) were enrolled. Standard linearly blended and VMI series were reconstructed in 10-keV intervals. Lesion-to-parenchyma contrast-to-noise ratio (CNR) was calculated and the best VMI series was further investigated in a subjective evaluation of overall image quality and lesion demarcation. Size measurements were performed independently by measuring all hypodense lesions (n = 58) twice in a predefined sequence. Inter- and intra-rater agreement was assessed using intra-class correlation coefficient (ICC) statistics. RESULTS: The calculated CNR was greatest at 40-keV VMI (4.3 ± 2.6), significantly higher compared to standard reconstructions (2.9 ± 1.9; p < 0.001). Subjective ratings for overall image quality showed no significant difference between the 2 reconstruction techniques (both medians 4; p = 0.147), while lesion margin demarcation was found to be superior for 40-keV VMI (median 5; p ≤ 0.001). Inter- (ICC, 0.98 for 40-keV VMI; ICC, 0.93 for standard reconstruction) and intra-rater (ICC, 0.99 for 40-keV VMI; ICC, 0.94 for standard image series) analysis showed an excellent agreement for lesion measurements in both reconstruction techniques. CONCLUSION: Noise-optimized VMI reconstructions significantly improve contrast and lesion demarcation of hypoattenuating liver metastases in patients with the fatty liver disease compared to standard reconstruction.

Dual-energy CT in patients with colorectal cancer: Improved assessment of hypoattenuating liver metastases using noise-optimized virtual monoenergetic imaging.

PURPOSE: To assess the value of the noise-optimized virtual monoenergetic imaging (VMI+) technique on quantitative and qualitative image parameters in patients with hypoattenuating liver metastases from colorectal cancer (CRC) at abdominal dual-energy CT (DECT). MATERIALS AND METHODS: Fifty-three consecutive patients (mean age, 70.3 ± 11.4 years; range, 44-86 years) with histologically proven, hypoattenuating liver metastases from CRC were retrospectively included in this IRB-approved study. DECT datasets were reconstructed as standard linearly-blended M_0.6 image series, traditional virtual monoenergetic images (VMI), and noise-optimized VMI+ series. VMI and VMI+ reconstructions were obtained at energy levels ranging from 40 to 100-keV with 10-keV increments. Signal attenuation of liver parenchyma and liver metastases was measured to calculate signal-to-noise (SNR) and contrast-to-noise (CNR) ratios. Each image series was subjectively rated by three blinded radiologists with regard to image quality, lesion delineation, and image noise using a five-point Likert scale. RESULTS: Quantitative image quality parameters peaked at 40-keV VMI+ (SNR, 8.1 ± 3.4; CNR, 6.5 ± 2.6) with statistically significant differences in comparison with standard reconstructions and all traditional VMI series (P ≤ 0.001). Qualitative image analysis revealed best rating scores for 60-keV VMI+ series (median, 5) with significant differences compared to linearly-blended M_0.6 and all traditional VMI series (P ≤ 0.001). Lesion delineation showed significantly superior ratings for 40-keV VMI+ series compared to all other reconstructions (median, 5) (P ≤ 0.001). CONCLUSION: Low-keV VMI+ reconstructions demonstrate significantly increased quantitative and qualitative image quality parameters in patients with hypoattenuating liver metastases from CRC in comparison with standard reconstructions and traditional VMI series at abdominal DECT. Best lesion delineation can be achieved at 40-keV VMI+.