深層学習画像再構成を用いた1.5T腹部MRIにおける薄層・高速T1強調GREシーケンスの応用

Deep-learning (DL)-based image reconstruction (DLR) is a key technique for reducing acquisition time (TA) and increasing morphologic resolution in abdominal magnetic resonance imaging (MRI). We aim to compare the performance of a standard ( ${\mathrm{VIBE}}_{\mathrm{Std}}$ ) gradient echo (GRE) sequence with Dixon fat separation versus an accelerated ultra-fast ( ${\mathrm{VIBE}}_{\mathrm{UF}}$ ) and high-resolution ( ${\mathrm{VIBE}}_{\mathrm{HR}}$ ) T1-weighted GRE sequence with Dixon fat separation and DLR.

A total of 50 patients with an abdominal 1.5T MRI, with a mean age of $59\pm 11$ years, were prospectively included from January to July 2023. Each examination protocol included ${\mathrm{VIBE}}_{\mathrm{Std}}$ , ${\mathrm{VIBE}}_{\mathrm{UF}}$ , and ${\mathrm{VIBE}}_{\mathrm{HR}}$ . Both DL sequences use more aggressive parallel imaging and partial Fourier sampling to reduce TA (slice thickness ${\mathrm{VIBE}}_{\mathrm{Std}}$ and ${\mathrm{VIBE}}_{\mathrm{UF}}$ 3 mm, ${\mathrm{VIBE}}_{\mathrm{HR}}$ 2 mm). Evaluation of each contrast-enhanced datasets for noise, artifacts, sharpness/contrast, overall image quality, and diagnostic confidence was performed independently by four radiologists using a Likert scale of 1 to 5 (5 = best).

${\mathrm{VIBE}}_{\mathrm{UF}}$ significantly reduced TA (mean 7.3 s versus 15.0 s ( ${\mathrm{VIBE}}_{\mathrm{Std}}$ ) and 14.5 s ( ${\mathrm{VIBE}}_{\mathrm{HR}}$ ); $p<0.001$ ). Both DL sequences provided significantly better sharpness/contrast for all organs compared with ${\mathrm{VIBE}}_{\mathrm{Std}}$ (median 5 versus 4; $p<0.001$ ). ${\mathrm{VIBE}}_{\mathrm{UF}}$ showed less noise than ${\mathrm{VIBE}}_{\mathrm{Std}}$ (median 5 versus 4; $p<0.001$ ), but ${\mathrm{VIBE}}_{\mathrm{Std}}$ was less artifact-affected than both DL sequences (median 5 versus 4; $p<0.001$ ). Overall image quality was superior in both DL sequences compared with ${\mathrm{VIBE}}_{\mathrm{Std}}$ (median 5 versus 4; $p<0.001$ ). Diagnostic confidence and lesion detectability were not significantly different ( $p>0.05$ ).

DL-based image reconstruction significantly improves overall image quality for ${\mathrm{VIBE}}_{\mathrm{UF}}$ and ${\mathrm{VIBE}}_{\mathrm{HR}}$ , with ${\mathrm{VIBE}}_{\mathrm{UF}}$ reducing TA by $\sim 50\%$ .