Dual-layer dual-energy computed tomography for the assessment of hypovascular hepatic metastases: impact of closing k-edge on image quality and lesion detectability.
Objective: To evaluate the image quality of virtual-monoenergetic-imaging (VMI) from dual-layer dual-energy CT (DLCT) for the assessment of hypovascular liver metastases and its effect on lesion detectability.
Methods: Eighty-one patients with hypovascular-liver-metastases undergoing portal-venous-phase abdominal DLCT were included. Polyenergetic-images (PEI) and VMI at 40-200 keV (VMI40-200, 10-keV interval) were reconstructed. Image noise, tumor-to-liver contrast, and contrast-to-noise ratio (CNR) of hepatic parenchyma and metastatic nodules (n = 288) were measured to determine the optimal monoenergetic levels. Two radiologists independently and subjectively assessed the image quality (image contrast, image noise, and diagnostic confidence) of PEI and optimal VMI on 5-point scales to determine the best energy. For 38 patients having up to 10 metastases each with diameters < 25 mm (153 lesions), we compared blindly assessed lesion detectability and conspicuity between PEI and VMI at the best energy.
Results: Image noise of VMI40-200 was consistently lower than that of PEI (p < 0.01). Tumor-to-liver contrast and CNR increased as the energy decreased with CNR at VMI40-70 being higher than that observed on PEI (p < 0.01). The highest subjective score for diagnostic confidence was assigned at VMI40 followed by VMI50-70, all of which were significantly better than that of PEI (p < 0.01, kappa = 0.75). Lesion detectability at VMI40 was significantly superior to PEI, especially for lesions with diameters of < 10 mm (p < 0.01, kappa ≥ 0.6).
Conclusions: VMI40-70 provided a better subjective and objective image quality for the evaluation of hypovascular liver metastases, and the lesion detectability was improved with use of VMI40 compared with conventional PEI. Conclusions: • DLCT-VMI at 40-70 keV provides a superior subjective and objective image quality compared with conventional PEI for the assessment of hypovascular hepatic metastases during portal venous phase. • Tumor-to-liver contrast and CNR of hypovascular hepatic metastases was maximized at 40 keV without a relevant increase in the image noise. • VMI at 40 keV yields a superior lesion detectability, especially for small (< 1 cm) metastatic nodules compared with conventional PEI.