BACKGROUND: Metal artifacts caused by endovascular coils in intracranial aneurysms pose challenges in the accurate assessment of adjacent vessels. This study evaluated the performance of S-MAR, a novel post-processing technology that combines the advantages of three traditional techniques
that is, virtual monochromatic imaging (VMI) of spectral computed tomography (CT), metal artifact reduction (MAR), and subtraction algorithm. This study aimed to investigate its performance under different conditions and validate its accuracy. METHODS: In total, 94 patients who had undergone cerebral computed tomography angiography (CTA) with dual-layer detector spectral CT were enrolled in this retrospective study. A total of 106 coils were found, and 67 patients had accompanying stents. Fifty patients underwent both digital subtraction angiography (DSA) and CTA within 6 months. Conventional CTA images, VMI [range, 40-110 kilo electron volt (keV)], and virtual non-enhanced images were generated and then post-processed to S-MAR, using the MAR technique and a subtraction algorithm. The contrast-to-noise ratio (CNR) and background noise were calculated. The maximum diameter, minimum diameter, and mean diameter of the adjacent vessels were measured. The coil artifact (CA) score was qualitatively assessed by two radiologists independently. RESULTS: Compared to conventional images, S-MAR (40-70 keV) had significantly reduced metal artifacts, improved CNR, and lower CA scores (P<
0.002). In S-MAR, a range of 60-70 keV is more suitable than 40-50 keV for coils with a diameter >
8 mm. S-MAR also provides more accurate luminal quantitative measurements (maximum, minimum and mean diameter) and shows good consistency with DSA [intraclass correlation coefficients: 0.845 (0.783, 0.884), 0.947(0.876, 0.954), and 0.956 (0.875, 0.962)]. CONCLUSIONS: S-MAR enhances vessel visualization and measurement accuracy. Our findings support its use in clinical practice for evaluating intracranial aneurysms post-embolization.