Dual-energy CT with metal artifact reduction for post-EVAR CTA surveillance: a prospective study with image quality...

To evaluate the value of dual-energy computed tomography (DECT) combined with the iterative metal artifact reduction (iMAR) for postoperative endovascular aortic repair (EVAR) CT angiography (CTA) surveillance.

In this prospective study, post-EVAR CTA surveillance was performed using a Siemens Force CT (100/150 kVp dual-energy). Images were reconstructed as conventional mixed and 40-190 keV virtual monoenergetic images (VMIs), with and without iMAR. Image quality was assessed via quantitative parameters and 5-point subjective scores, with clinical data and radiation dose recorded simultaneously. Diagnostic performance was referenced to surgical/DSA findings and blinded expert adjudication. Statistical analysis was performed via two-way repeated-measures ANOVA with post-hoc paired t-tests or Wilcoxon tests for quantitative parameters, While subjective scores were analyzed using the Friedman test with post-hoc Wilcoxon signed-rank tests. P values were Bonferroni-corrected; corrected P < 0.05 indicated significance.

Fifty post-EVAR patients (82% male; 68.7 ± 8.9 years) received safe radiation doses; stent type yielded no quantitative differences (P > 0.05). iMAR-DECT synergistically enhanced image quality, with improvements correlating positively with artifact severity while sparing normal tissues for iMAR. Energy levels dominated the trend: most parameters declined rapidly at 40-70 keV but plateaued beyond 100 keV. iMAR significantly outperformed non-iMAR in scores of artifact reduction (plateauing at 150-190 keV) and vessel visibility (peaking at 90 keV, P < 0.01). For diagnosis, iMAR+DECT converted five false negatives into true positives (endoleaks, thrombosis, infection), showing a trend toward higher sensitivity than those for non-iMAR+DECT (P > 0.05, large effect size).

The integration of DECT-VMIs with the iMAR algorithm appears to improve image quality in post-EVAR CTA by effectively reducing stent-related metal artifacts while enhancing vascular and soft-tissue visualization. Based on our quantitative and qualitative assessments, an energy level of 90 keV (effective range 80-100 keV) may be considered optimal to optimize diagnostic evaluation, pending validation in larger, multi-center cohorts.