Objective To optimize a scanning protocol for dual-energy spectral lower-extremity computed tomography (CT) venography (CTV) based on a phantom study.

Methods Test plugs were placed in the cavities of an energy CT quality-control phantom to simulate clinical scenarios. A 4 mgI/mL iodine rod was used to mimic venous enhancement in the lower extremities, and duck blood clots of various sizes were placed in test tubes containing 4 mgI/mL of iodine solution to simulate thrombi of different sizes in the lower-extremity veins. Revolution CT was used to perform standard CT imaging (Group A) and spectral imaging (Group B) on phantoms containing iodine rods and test tubes. The imaging parameters for Group A were as follows: tube voltage of 120 kVp, auto tube-current technology (100–600 mA), noise index (NI) of 10, and image reconstruction using 40% posterior multimodel adaptive statistical iterative reconstruction (ASIR-V). The imaging parameters for Group B were spectral imaging (GSI) mode, instantaneous dual tube voltage of 80/140 kVp, tube current with GSI Assist technology, and three scan groups based on NI values of 10, 11, and 12. For each scan group, monoenergetic images at 40–70 keV with 10 keV intervals were reconstructed, each combined with 40%, 60%, and 80% posterior ASIR-V, which resulted in 36 image sets. Other imaging parameters for Groups A and B were consistent. The effective radiation doses (ED) for Groups A and B were recorded after scanning, and the contrast-to-noise ratio (CNR) of the iodine rods was calculated. Subjective image quality and true- and false-positive rates for thrombus identification were assessed.

Results The EDs for Group B, with NI values of 11 and 12, were 21.5% and 32.2% lower than those for Group A, respectively. In Group B, for the scan groups with NI values of 10 and 11, except for the images at 70 keV combined with 40% ASIR-V and at 60 keV combined with 40% ASIR-V, the CNR of the iodine rods was higher than that in Group A. The highest edge-sharpness scores for the iodine rods in Group B were observed in the scan group with an NI value of 10 for images at 50 keV combined with 40% and 60% ASIR-V, and in the scan group with an NI value of 11 for images at 50 keV combined with 60% ASIR-V. These three image sets scored 5 (4, 5) compared with Group A’s score of 3 (3, 4). The true- and false-positive rates for large-thrombus identification in Group A were 65.0% and 30.0%, respectively, whereas those for small-thrombus identification were 55.0% and 50.0%, respectively. In Group B, the best thrombus-identification efficacy was observed in the scan groups, with NI values of 10 and 11 for images at 50 keV combined with 60% ASIR-V. The true- and false-positive rates for large-thrombus identification were 90.0% and 5.0%, respectively, whereas those for small-thrombus identification were 80.0% and 5.0%, respectively.

Conclusions Setting the NI to 11 and reconstructing monoenergetic images at 50 keV combined with 60% ASIR-V is the optimal imaging strategy for dual-energy spectral lower-extremity CTV, which balances between image quality and radiation dose in the current phantom study.