Open 1.0-T versus closed 1.5-T cardiac MR: Image quality assessment

Abstract

Purpose: The aim of this paper was to compare the open 1-T (O-1T) versus the closed 1.5-T (C-1.5T) cardiac magnetic resonance (MR).

Patients/methods: The MR examinations of two concurrent cohorts (each including 100 subjects) of patients with suspected or known cardiac disease were reviewed. Such examinations were obtained using O-1T or C-1.5T MRI. The bright-blood cine, T1-weighted (T1), T2-weighed short-tau inversion recovery (T2-STIR), late gadolinium enhancement (LGE) sequences were performed. Signal-to-noise ratio of blood (SNRb) or myocardium (SNRm), and contrast-to-noise ratio of myocardium (CNRm) were calculated. Subjective image quality (SIQ) of each sequence was graded as 0 = poor, 1 = intermediate, or 2 = optimal. Each examination was considered as diagnostic when the report answered the clinical question.

Results: C-1.5T was better than O-1T on cine for SNRb(median 172 versus 452), SNRm(71 versus 160) and CNRm (107 versus 265) and on T2-STIR for SNRb(10 versus 29), SNRm(74 versus 261) and CNRm(-67 versus -233)(P < 0.001). On LGE, SNRm was higher with O-1T than for C-1.5T (312 versus 79, P < 0.001) while CNR was lower (158 versus 389; P < 0.001). No significant differences were found for SNRb on LGE and both SNRm and CNRm on T1 (P ≥ 0.215). SIQ of O-1T was not significantly different from that of C-1.5T for both R1 and R2 for cine, T1, and LGE (P ≥ 0.157); for T2-STIR, SIQ of O-1T was significantly lower (P = 0.003). R1-R2 concordance was almost perfect (κ = 0.816-0.894), and all examinations were diagnostic.

Conclusion: Even though quantitative measurements mostly favored C-1.5T, the SIQ of O-1T was not significantly different for any sequence, with the only exception of T2-STIR.

Keywords: Cardiac magnetic resonance; Claustrophobic patients; Obese patients; Open magnetic resonance; Reproducibility.