Assessment of AI-accelerated T2-weighted brain MRI, based on clinical ratings and image quality evaluation.
Authors
Affiliations (4)
Affiliations (4)
- Division of Neuroradiology and Musculoskeletal Radiology, Department of Biomedical Imaging and Image-guided Therapy, Medical University of Vienna, Vienna, Austria; Comprehensive Center for Clinical Neurosciences and Mental Health, Medical University of Vienna, Vienna, Austria.
- Division of Neuroradiology and Musculoskeletal Radiology, Department of Biomedical Imaging and Image-guided Therapy, Medical University of Vienna, Vienna, Austria; NMR Research Unit, Queen Square Multiple Sclerosis Centre, Queen Square Institute of Neurology, University College London, London, United Kingdom; Department of Radiology and Nuclear Medicine, VU University Medical Centre, Amsterdam, NL, the Netherlands; National Institute for Health Research (NIHR) University College London Hospitals (UCLH) Biomedical Research Centre, United Kingdom; Centre for Medical Image Computing (CMIC), Department of Medical Physics and Biomedical Engineering, University College London, United Kingdom.
- High Field MR Centre, Department of Biomedical Imaging and Image-Guided Therapy, Medical University of Vienna, Vienna, Austria; Centre of Advanced Imaging, University of Queensland, Brisbane, Australia.
- Division of Neuroradiology and Musculoskeletal Radiology, Department of Biomedical Imaging and Image-guided Therapy, Medical University of Vienna, Vienna, Austria; Comprehensive Center for Clinical Neurosciences and Mental Health, Medical University of Vienna, Vienna, Austria; NMR Research Unit, Queen Square Multiple Sclerosis Centre, Queen Square Institute of Neurology, University College London, London, United Kingdom.
Abstract
To compare clinical ratings and signal-to-noise ratio (SNR) measures of a commercially available Deep Learning-based MRI reconstruction method (T2<sub>(DR)</sub>) against conventional T2- turbo spin echo brain MRI (T2<sub>(CN)</sub>). 100 consecutive patients with various neurological conditions underwent both T2<sub>(DR)</sub> and T2<sub>(CN)</sub> on a Siemens Vida 3 T scanner with a 64-channel head coil in the same examination. Acquisition times were 3.33 min for T2<sub>(CN)</sub> and 1.04 min for T2<sub>(DR)</sub>. Four neuroradiologists evaluated overall image quality (OIQ), diagnostic safety (DS), and image artifacts (IA), blinded to the acquisition mode. SNR and SNR<sub>eff</sub> (adjusted for acquisition time) were calculated for air, grey- and white matter, and cerebrospinal fluid. The mean patient age was 43.6 years (SD 20.3), with 54 females. The distribution of non-diagnostic ratings did not differ significantly between T2<sub>(CN)</sub> and T2<sub>(DR)</sub> (IA p = 0.108; OIQ: p = 0.700 and DS: p = 0.652). However, when considering the full spectrum of ratings, significant differences favouring T2<sub>(CN)</sub> emerged in OIQ (p = 0.003) and IA (p < 0.001). T2<sub>(CN)</sub> had higher SNR (157.9, SD 123.4) than T2<sub>(DR)</sub> (112.8, SD 82.7), p < 0.001, but T2<sub>(DR)</sub> demonstrated superior SNR<sub>eff</sub> (14.1, SD 10.3) compared to T2<sub>(CN)</sub> (10.8, SD 8.5), p < 0.001. Our results suggest that while T2<sub>(DR)</sub> may be clinically applicable for a diagnostic setting, it does not fully match the quality of high-standard conventional T2<sub>(CN)</sub>, MRI acquisitions.