Kyeong Hwa Ryu1, Hye Jin Baek2, Jin Il Moon1, Bo Hwa Choi1, Sung Eun Park1, Ji Young Ha1, Kyung Nyeo Jeon1, Kyungsoo Bae1, Dae Seob Choi3, Soo Buem Cho4, Yedaun Lee5, Young Jin Heo6. 1. Department of Radiology, Gyeongsang National University School of Medicine, Gyeongsang National University Changwon Hospital, 11 Samjeongja-ro, Seongsan-gu, Changwon 51472, Republic of Korea. 2. Department of Radiology, Gyeongsang National University School of Medicine, Gyeongsang National University Changwon Hospital, 11 Samjeongja-ro, Seongsan-gu, Changwon 51472, Republic of Korea. Electronic address: sartre81@gmail.com. 3. Department of Radiology, Gyeongsang National University School of Medicine, Gyeongsang National University Hospital, 79 Gangnam-ro, Jinju 52727, South Korea. 4. Department of Radiology, Ewha Womans University College of Medicine, 1071 Anyangcheon-ro, Yangcheon-gu, Seoul 07985, South Korea. 5. Department of Radiology, Inje University College of Medicine, Haeundae Paik Hospital, 875 Haeundae-ro, Haeundae-gu, Busan 48108, South Korea. 6. Department of Radiology, Inje University College of Medicine, Busan Paik Hospital, 75 Bokji-ro, Busanjin-gu, Busan, 47392, South Korea.
Abstract
BACKGROUND AND PURPOSE: We investigated the clinical feasibility of synthetic MRI with a 4-min single scan using a 48-channel head coil as a routine neuroimaging protocol in daily practice by assessing its diagnostic image quality. METHODS: We retrospectively reviewed the imaging data of 89 patients who underwent routine brain MRI using synthetic MRI acquisition between February 2017 and April 2017. Image quality assessments were performed by two independent readers on synthetic T1 fluid-attenuated inversion recovery (FLAIR), T2-weighted, T2 FLAIR, and phase-sensitive inversion recovery sequences acquired using multiple-dynamic multiple-echo imaging. Interobserver reliability between the two readers was assessed using kappa (κ) statistics. RESULTS: On a 4-point assessment scale, the overall image quality and anatomical delineation provided by synthetic brain MRI were found to be good with scores of more than 3 points for all sequences except for the T2 FLAIR sequence. The synthetic T2 FLAIR sequence provided sufficient image quality but showed more pronounced artifacts, especially the CSF pulsation artifact and linear hyperintensity along the brain surface. Interobserver agreement for evaluating image quality of all synthetic sequences was good to excellent (κ, 0.61-0.99; P < 0.001). CONCLUSION: Synthetic MRI can be acceptable as a routine clinical neuroimaging protocol with a short scan time. It can be helpful to design customized and flexible neuroimaging protocols for each institution.
BACKGROUND AND PURPOSE: We investigated the clinical feasibility of synthetic MRI with a 4-min single scan using a 48-channel head coil as a routine neuroimaging protocol in daily practice by assessing its diagnostic image quality. METHODS: We retrospectively reviewed the imaging data of 89 patients who underwent routine brain MRI using synthetic MRI acquisition between February 2017 and April 2017. Image quality assessments were performed by two independent readers on synthetic T1 fluid-attenuated inversion recovery (FLAIR), T2-weighted, T2 FLAIR, and phase-sensitive inversion recovery sequences acquired using multiple-dynamic multiple-echo imaging. Interobserver reliability between the two readers was assessed using kappa (κ) statistics. RESULTS: On a 4-point assessment scale, the overall image quality and anatomical delineation provided by synthetic brain MRI were found to be good with scores of more than 3 points for all sequences except for the T2 FLAIR sequence. The synthetic T2 FLAIR sequence provided sufficient image quality but showed more pronounced artifacts, especially the CSF pulsation artifact and linear hyperintensity along the brain surface. Interobserver agreement for evaluating image quality of all synthetic sequences was good to excellent (κ, 0.61-0.99; P < 0.001). CONCLUSION: Synthetic MRI can be acceptable as a routine clinical neuroimaging protocol with a short scan time. It can be helpful to design customized and flexible neuroimaging protocols for each institution.
Authors: K H Ryu; H J Baek; S Skare; J I Moon; B H Choi; S E Park; J Y Ha; T B Kim; M J Hwang; T Sprenger Journal: AJNR Am J Neuroradiol Date: 2020-02-06 Impact factor: 3.825