SU‐F‐J‐153: ACR MRI Phantom Test On a 1.5T MR‐Simulator with Flexible Coils Setting and Performance Comparison to Volumetric Head Coil Setting

Abstract

Purpose:To accommodate head‐and‐neck (HN) RT simulation scan with immobilization, sub‐optimized flexible radiofrequency coils have to be used on an MR‐sim rather than a volumetric head coil for diagnostic radiological (DR) applications. In this study, we present and compare ACR MRI phantom test with flexible coils setting (RT‐setting) and volumetric head coil setting (DR‐setting).Methods:Two 8‐channel array coils were wrapped close to an ACR MRI phantom placed on a flat couch‐top in RT‐setting for acquisition. ACR MRI phantom tests were conducted 10 times for each setting on a 1.5T MR‐sim. Sagittal localizer (TE/TR=20/200ms), axial T1 (TE/TR=20/500ms) and T2 scans (2TEs/TR=20,80/2000ms) were acquired (NEX=1). Acqusition and image analysis were conducted following ACR guidelines. Extra 10 RT‐setting acquisitions were obtained by enabling prescan normalization.Results:For RT‐setting without prescan normalization, all tests passed the ACR recommended criteria but image intensity uniformity. Geometric accuracy was 147.8±0.3mm and 191.0±0.4mm. Spatial resolution was 0.9mm. Slice thickness was 5.2mm±0.1mm (T1) and 5.1±0.1mm (T2). Slice position accuracy was 1.4±0.6mm, −2.2±0.3mm (T1), and 1.5±0.2mm, −2.2±0.2mm (T2). Image uniformity was 80.6%±2.6% (T1) and 80.2%±3.1% (T2) (criteria: ≥87.5% for 1.5T). Ghosting ratio was 0.0006±0.0004. Low contrast detectability was 30.7±1.7 (T1) and 22.6±3.0 (T2). Image uniformity (91.4%±4.2% and 91.8%±4.3%) and low contrast detectability (36.5±1.4 and 31.6±2.2) of DR‐setting were better than RT‐setting. Prescan normalization substantially improved image uniformity to 93.4%±0.2% and 93.4%±0.3%, but slightly compromised ghosting (0.0033±0.0004) and spatial resolution (0.9mm or 1.0mm) for RT‐setting.Conclusion:Flexible coils setting for HN‐RT simulation scan could successfully pass ACR criteria (prescan normalization enabled) and generally achieve comparable performance to volumetric coil setting, although compromise on low contrast detectability was observed probably due to the lower signal‐to‐noise ratio of the flexible coils. Our results could also serve as a reference baseline of high‐field MR‐sim QA performance under ACR guideline.