Comparison of a deep learning-accelerated T2-weighted turbo spin echo (TSE) sequence and a standard T2-weighted TSE in prostate MRI: reduced acquisition times and comparable image quality
Comparison of a deep learning-accelerated T2-weighted turbo spin echo (TSE) sequence and a standard T2-weighted TSE in prostate MRI: reduced acquisition times and comparable image quality
Chenjiang Wu, Yudong Zhang
To investigate the feasibility of a deep learning-accelerated T2-weighted turbo spin echo (TSE) sequence (DL-T2WI) applied to prostate MRI, using standard T2-weighted TSE (ST-T2WI) as reference.
One hundred patients with prostate disease undergoing prostate MRI were prospectively enrolled between July 2022 and Dec 2022. Axial DL-T2WI and ST-T2WI of the prostate were scanned with exact spatial resolution. Quantitative image evaluation was conducted by calculating the signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR). Qualitative image evaluation including overall image quality, artifacts, boundary sharpness of the prostate gland, and lesion conspicuity was assessed by two radiologists using a 5-point Likert scale, with 5 indicating the best quality. Comparative analyses were conducted for the two sequences.
DL-T2WI resulted in a 62.2% timing reduction (109 sec for DL-T2WI and 288 sec for ST-T2WI in axial imaging, respectively). Compared to ST-T2WI, DL-T2WI had significantly higher SNR (p≤0.001) and CNR (p≤0.007). Inter-observer agreement regarding qualitative evaluation was excellent (Kendall’s W>0.75). In axial imaging, the overall image quality of DL-T2WI was not significantly different from that of ST-T2WI (P=0.417), but had fewer artifacts (p<0.001) and better lesion conspicuity (p<0.001), while ST-T2WI had higher definition of prostate borders (p<0.001).
Compared with standard TSE, deep learning-accelerated T2WI is feasible with comparable overall image quality to ST-T2WI while reducing prostate MRI acquisition time.
