Synthetic MRI for Clinical Neuroimaging: Results of the Magnetic Resonance Image Compilation (MAGiC) Prospective, Multicenter, Multireader Trial

Tanenbaum, Lawrence N.; Tsiouris, Apostolos John; Johnson, Angela N.; Naidich, Thomas P.; DeLano, Mark C.; Melhem, Elias R.; Quarterman, Patrick; Parameswaran, Sajeesh; Shankaranarayanan, Ajit; Goyen, Matthias; Field, Aaron S.

doi:10.3174/ajnr.a5227

Cited by 209 publications

(241 citation statements)

References 25 publications

Supporting

Mentioning

226

Contrasting

Unclassified

Order By: Relevance

“…Several studies have been conducted to validate the clinical feasibility of synthetic MRI . These studies reported that synthetic images such as T 1 w, T 2 w, and PDw had acceptable diagnostic utilities like that of conventional imaging series.…”

mentioning

confidence: 99%

Data‐driven synthetic MRI FLAIR artifact correction via deep neural network

Ryu

Nam

Gho³

et al. 2019

Magnetic Resonance Imaging

View full text Add to dashboard Cite

Background FLAIR (fluid attenuated inversion recovery) imaging via synthetic MRI methods leads to artifacts in the brain, which can cause diagnostic limitations. The main sources of the artifacts are attributed to the partial volume effect and flow, which are difficult to correct by analytical modeling. In this study, a deep learning (DL)‐based synthetic FLAIR method was developed, which does not require analytical modeling of the signal. Purpose To correct artifacts in synthetic FLAIR using a DL method. Study Type Retrospective. Subjects A total of 80 subjects with clinical indications (60.6 ± 16.7 years, 38 males, 42 females) were divided into three groups: a training set (56 subjects, 62.1 ± 14.8 years, 25 males, 31 females), a validation set (1 subject, 62 years, male), and the testing set (23 subjects, 57.3 ± 20.4 years, 13 males, 10 females). Field Strength/Sequence 3 T MRI using a multiple‐dynamic multiple‐echo acquisition (MDME) sequence for synthetic MRI and a conventional FLAIR sequence. Assessment Normalized root mean square (NRMSE) and structural similarity (SSIM) were computed for uncorrected synthetic FLAIR and DL‐corrected FLAIR. In addition, three neuroradiologists scored the three FLAIR datasets blindly, evaluating image quality and artifacts for sulci/periventricular and intraventricular/cistern space regions. Statistical Tests Pairwise Student's t‐tests and a Wilcoxon test were performed. Results For quantitative assessment, NRMSE improved from 4.2% to 2.9% (P < 0.0001) and SSIM improved from 0.85 to 0.93 (P < 0.0001). Additionally, NRMSE values significantly improved from 1.58% to 1.26% (P < 0.001), 3.1% to 1.5% (P < 0.0001), and 2.7% to 1.4% (P < 0.0001) in white matter, gray matter, and cerebral spinal fluid (CSF) regions, respectively, when using DL‐corrected FLAIR. For qualitative assessment, DL correction achieved improved overall quality, fewer artifacts in sulci and periventricular regions, and in intraventricular and cistern space regions. Data Conclusion The DL approach provides a promising method to correct artifacts in synthetic FLAIR. Level of Evidence: 4 Technical Efficacy: Stage 1 J. Magn. Reson. Imaging 2019;50:1413–1423.

show abstract

mentioning

confidence: 99%

Data‐driven synthetic MRI FLAIR artifact correction via deep neural network

Ryu

Nam

Gho³

et al. 2019

Magnetic Resonance Imaging

View full text Add to dashboard Cite

show abstract

“…[86] The synthetic FLAIR sequence had notably lower scores on image quality and more artefacts than conventional images. This is in line with other studies [54,85], reporting artefacts in FLAIR images especially in the interface between CSF and brain tissue, which makes it necessary to carefully assess the other sequences in order to classify these changes as artefacts and not pathological changes. These kinds of artefacts could be attributed to the difficulty of segmenting voxels containing two types of tissues, CSF and brain parenchyma, which is often the case in the interface of ventricle walls and at the sulci.…”

Section: Paper Isupporting

confidence: 60%

“…A more recent study with a later version of the sequence for synthetic MR showed improved numbers, with 96% of synthetic images being rated as sufficient, compared to 98% for conventional images. [85] Tanenbaum et al had a larger patient cohort with a wider range of diagnoses than our study, and they found that overall diagnostic image quality of synthetic images was statistically noninferior to conventional images. Synthetic T1WI had higher lesion conspicuity than conventional T1WI, possibly due to the higher contrast, and this could explain the higher number of correct diagnoses in synthetic T1WI.…”

Section: Paper Imentioning

confidence: 70%

“…The Tanenbaum 2017 study had similar results, with a larger patient cohort with varied diagnoses. [85] Since synthetic MRI is dependent on only one scan, there could be concerns regarding motion artefacts corrupting the data, affecting all synthetic images. In this small cohort this was not an issue; Tanenbaum et al reported motion artefacts in 7.5% of the synthetic images, which is lower compared to previous reports on motion artefacts in conventional studies, where up to 29% of images could be affected.…”

Section: Paper Imentioning

confidence: 99%

“…[54,85,86] In this setting, it might be appropriate to consider the indications for different radiology examinations and modalities, and to speculate whether a short MRI protocol with synthetic MRI might be a sufficient substitute for some of the CT examinations performed, e.g., in the work-up of patients with non-specific headache and a low level of suspicion of pathological changes. This would be beneficial for the patients with regards to radiation aspects, and if it was possible to achieve a higher throughput in the MR suite with a short MR protocol it could be manageable for the radiology department.…”

Section: Future Aspectsmentioning

confidence: 99%

See 2 more Smart Citations

Clinical Applications of Synthetic MRI of the Brain

Blystad¹

View full text Add to dashboard Cite

Magnetic Resonance Imaging (MRI) has a high soft-tissue contrast with a high sensitivity for detecting pathological changes in the brain. Conventional MRI is a time-consuming method with multiple scans that relies on the visual assessment of the neuroradiologist. Synthetic MRI uses one scan to produce conventional images, but also quantitative maps based on relaxometry, that can be used to quantitatively analyse tissue properties and pathological changes. The studies presented here apply the use of synthetic MRI of the brain in different clinical settings. In the first study, synthetic MR images were compared to conventional MR images in 22 patients. The contrast, the contrast-to-noise ratio, and the diagnostic quality were assessed. Image quality was perceived to be inferior in the synthetic images, but synthetic images agreed with the clinical diagnoses to the same extent as the conventional images. Patients with early multiple sclerosis were analysed in the second study. In patients with multiple sclerosis, contrast-enhancing white matter lesions are a sign of active disease and can indicate a need for a change in therapy. Gadolinium-based contrast agents are used to detect active lesions, but concern has been raised regarding the long-term effects of repeated use of gadolinium. In this study, relaxometry was used to evaluate whether pre-contrast injection tissue-relaxation rates and proton density can identify active lesions without gadolinium. The findings suggest that active lesions often have relaxation times and proton density that differ from non-enhancing lesions, but with some overlap. This makes it difficult to replace gadolinium-based contrast agent injection with synthetic MRI in the monitoring of MS patients. Malignant gliomas are primary brain tumours with contrast enhancement due to a defective blood-brain barrier. However, they also grow in an infiltrative, diffuse manner, making it difficult to clearly delineate them from surrounding normal brain tissue in the diagnostic workup, at surgery, and during follow-up. The contrast-enhancing part of the tumour is easily visualised, but not the diffuse infiltration. In studies three and four, synthetic MRI was used to analyse the peritumoral area of malignant gliomas, and revealed quantitative findings regarding peritumoral relaxation changes and non-visible contrast enhancement suggestive of non-visible infiltrative tumour growth. In conclusion, synthetic MRI provides quantitative information about the brain tissue and this could improve the diagnosis and treatment for patients

show abstract

Characterization of meningiomas with synthetic imaging

et al. 2022

View full text Add to dashboard Cite

Introduction: Synthetic magnetic resonance imaging (SyMRI) is a novel quantitative and qualitative technique that permits the reconstruction of multiple image contrasts and quantitative maps from a single scan, thereby providing quantitative information and reducing scan times. The purpose of this study is to characterize intracranial meningiomas using SyMRI. Methods:The study included 35 patients with meningiomas (6 males, 29 females; mean age 61 ± 17 years; range 21-90 years). Using 3T MR scanners, SyMRI was performed in addition to conventional FSET2, FLAIR, DWI, T1, and T1 with gadolinium. SyMRI software was used to generate T1, T2, and PD quantitative maps. Osirix MD was used to measure quantitative values of T1, T2, and PD using a ROI. Results:We analyzed 42 meningiomas, 8 of which were associated with edema, and 5 contained calcifications.Mean relaxivity values of meningiomas on synthetic T1, T2, and PD maps at 3T MRI were 1382.6 ± 391.7 ms, 95.6 ± 36.5 ms, and 89.1 ± 9.7 pu, respectively. Signal intensities in terms of T1, T2, and PD did not differ significantly between meningiomas with and without edema (p = .994, p = .356, and p = .221, respectively), nor between meningiomas containing and not containing calcifications (p = .840, p = .710, and p = .455, respectively). Values of T1 and T2 measured in meningiomas and the normalappearing white matter approximated reference values found in the literature with other quantitative methods. Conclusion:The presented method offers a novel approach to characterize meningiomas through their relaxation parameters measured with a SyMRI sequence.

show abstract

Synthetic MRI for Clinical Neuroimaging: Results of the Magnetic Resonance Image Compilation (MAGiC) Prospective, Multicenter, Multireader Trial

Cited by 209 publications

References 25 publications

Data‐driven synthetic MRI FLAIR artifact correction via deep neural network

Data‐driven synthetic MRI FLAIR artifact correction via deep neural network

Clinical Applications of Synthetic MRI of the Brain

Characterization of meningiomas with synthetic imaging

Contact Info

Product

Resources

About