Artefacts induced by coiled intracranial aneurysms on 3.0-Tesla versus 1.5-Tesla MR angiography—An in vivo and in vitro study

Abstract Objective To compare metal-induced artefacts from coiled intracranial aneurysms on 3.0-Tesla and 1.5-Tesla magnetic resonance angiography (MRA), since concerns persist on artefact enlargement at 3.0 Tesla. Materials and methods We scanned 19 patients (mean age 53; 16 women) with 20 saccular aneurysms treated with coils only, at 1.5 and 3.0 Tesla according to standard clinical 3D TOF-MRA protocols containing a shorter echo-time but weaker read-out gradient at 3.0 Tesla in addition to intra-arterial digital subtraction angiography (IA-DSA). Per modality two neuro-radiologists assessed the occlusion status, measured residual flow, and indicated whether coil artefacts disturbed this assessment on MRA. We assessed relative risks for disturbance by coil artefacts, weighted kappa's for agreement on occlusion levels, and we compared remnant sizes. For artefact measurements, a coil model was created and scanned with the same protocols followed by 2D MR scans with variation of echo-time and read-out gradient strength. Results Coil artefacts disturbed assessments less frequently at 3.0 Tesla than at 1.5 Tesla (RR: 0.3; 95%CI: 0.1–0.8). On 3.0-Tesla MRA, remnants were larger than on 1.5-Tesla MRA (difference: 0.7 mm; 95%CI: 0.3–1.1) and larger than on IA-DSA (difference: 1.0 mm; 95%CI: 0.6–1.5) with similar agreement on occlusion levels with IA-DSA for both field strengths ( κ 0.53; 95%CI: 0.23–0.84 for 1.5-Tesla MRA and IA-DSA; κ 0.47; 95%CI: 0.19–0.76 for 3.0-Tesla MRA and IA-DSA). Coil model artefacts were smaller at 3.0 Tesla than at 1.5 Tesla. The echo-time influenced artefact size more than the read-out gradient. Conclusions Artefacts were not larger, but smaller at 3.0 Tesla because a shorter echo-time at 3.0 Tesla negated artefact enlargement. Despite smaller artefacts and larger remnants at 3.0 Tesla, occlusion levels were similar for both field strengths.