Pushing MP2RAGE boundaries: Ultimate time-efficient parameterization combined with exhaustive T1 synthetic contrasts

PURPOSE

MP2RAGE parameter optimization is redefined to allow more time-efficient MR acquisitions, whereas the T1 -based synthetic imaging framework is used to obtain on-demand T1 -weighted contrasts. Our aim was to validate this concept on healthy volunteers and patients with multiple sclerosis, using plug-and-play parallel-transmission brain imaging at 7 T.

METHODS

A "time-efficient" MP2RAGE sequence was designed with optimized parameters including TI and TR set as small as possible. Extended phase graph formalism was used to set flip-angle values to maximize the gray-to-white-matter contrast-to-noise ratio (CNR). Several synthetic contrasts (UNI, EDGE, FGATIR, FLAWSMIN , FLAWSHCO ) were generated online based on the acquired T1 maps. Experimental validation was performed on 4 healthy volunteers at various spatial resolutions. Clinical applicability was evaluated on 6 patients with multiple sclerosis, scanned with both time-efficient and conventional MP2RAGE parameterizations.

RESULTS

The proposed time-efficient MP2RAGE protocols reduced acquisition time by 40%, 30%, and 19% for brain imaging at (1 mm)3 , (0.80 mm)3 and (0.65 mm)3 , respectively, when compared with conventional parameterizations. They also provided all synthetic contrasts and comparable contrast-to-noise ratio on UNI images. The flexibility in parameter selection allowed us to obtain a whole-brain (0.45 mm)3 acquisition in 19 min 56 s. On patients with multiple sclerosis, a (0.67 mm)3 time-efficient acquisition enhanced cortical lesion visualization compared with a conventional (0.80 mm)3 protocol, while decreasing the scan time by 15%.

CONCLUSION

The proposed optimization, associated with T1 -based synthetic contrasts, enabled substantial decrease of the acquisition time or higher spatial resolution scans for a given time budget, while generating all typical brain contrasts derived from MP2RAGE.

Time Course Changes of Synthetic Relaxation Time During Neoadjuvant Chemotherapy in Breast Cancer: The Optimal Parameter for Treatment Response Evaluation

BACKGROUND

Synthetic MRI (syMRI) has enabled quantification of multiple relaxation parameters (T1/T2 relaxation time [T1/T2], proton density [PD]), and their longitudinal change during neoadjuvant chemotherapy (NAC) promises to be valuable parameters for treatment response evaluation in breast cancer.

PURPOSE

To investigate the time course changes of syMRI parameters during NAC and evaluate their value as predictors for pathological complete response (pCR) in breast cancer.

STUDY TYPE

Retrospective, longitudinal.

POPULATION

A total of 129 women (median age, 50 years; range, 28-69 years) with locally advanced breast cancer who underwent NAC; all performed multiple conventional breast MRI examinations with added syMRI during NAC.

FIELD STRENGTH/SEQUENCE

A 3.0 T, T1-weighted dynamic contrast enhanced and syMRI acquired by a multiple-dynamic, multiple-echo sequence.

ASSESSMENT

Breast MRI was set at four time-points: baseline, after one cycle, after three or four cycles of NAC and preoperation. SyMRI parameters and tumor diameters were measured and their changes from baseline were calculated. All parameters were compared between pCR and non-pCR. Interaction between syMRI parameters and clinicopathological features was analyzed.

STATISTICAL TESTS

Mann-Whitney U tests, random effects model of repeated measurement, receiver operating characteristic (ROC) analysis, interaction analysis.

RESULTS

Median synthetic T1/T2/PD and tumor diameter generally decreased throughout NAC. Absolute T1 at early-NAC, T1, and PD at mid-NAC were significantly lower in the pCR group. After early-NAC, the T1 change was significantly higher in the pCR (median ± IQR, 18.17 ± 11.33) than the non-pCR group (median ± IQR, 10.90 ± 10.03), with the highest area under the ROC curves (AUC) of 0.769 (95% CI, 0.684-0.838). Interaction analysis showed that histological grade III patients had higher odds ratio (OR) (OR = 1.206) compared to grade II patients (OR = 1.067).

DATA CONCLUSION

Synthetic T1 changes after one cycle of NAC maybe useful for early evaluating NAC response in breast cancer during whole treatment cycles. However, its discriminative ability is significantly affected by histological grade.

EVIDENCE LEVEL

4 TECHNICAL EFFICACY: Stage 2.

The advantages of the magnetic resonance image compilation (MAGiC) method for the prognosis of neonatal hypoglycemic encephalopathy

Objectives

To explore the prognostic value of magnetic resonance image compilation (MAGiC) in the quantitative assessment of neonatal hypoglycemic encephalopathy (HE).

Methods

A total of 75 neonatal HE patients who underwent synthetic MRI were included in this retrospective study. Perinatal clinical data were collected. T1, T2 and proton density (PD) values were measured in the white matter of the frontal lobe, parietal lobe, temporal lobe and occipital lobe, centrum semiovale, periventricular white matter, thalamus, lenticular nucleus, caudate nucleus, corpus callosum and cerebellum, which were generated by MAGiC. The patients were divided into two groups (group A: normal and mild developmental disability; group B: severe developmental disability) according to the score of Bayley Scales of Infant Development (Bayley III) at 9-12 months of age. Student's t test, Wilcoxon test, and Fisher's test were performed to compare data across the two groups. Multivariate logistic regression was used to identify the predictors of poor prognosis, and receiver operating characteristic (ROC) curves were created to evaluate the diagnostic accuracy.

Results

T1 and T2 values of the parietal lobe, occipital lobe, center semiovale, periventricular white matter, thalamus, and corpus callosum were higher in group B than in group A (p < 0.05). PD values of the occipital lobe, center semiovale, thalamus, and corpus callosum were higher in group B than in group A (p < 0.05). Multivariate logistic regression analysis showed that the duration of hypoglycemia, neonatal behavioral neurological assessment (NBNA) scores, T1 and T2 values of the occipital lobe, and T1 values of the corpus callosum and thalamus were independent predictors of severe HE (OR > 1, p < 0.05). The T2 values of the occipital lobe showed the best diagnostic performance, with an AUC value of 0.844, sensitivity of 83.02%, and specificity of 88.16%. Furthermore, the combination of MAGiC quantitative values and perinatal clinical features can improve the AUC (AUC = 0.923) compared with the use of MAGiC or perinatal clinical features alone.

Conclusion

The quantitative values of MAGiC can predict the prognosis of HE early, and the prediction efficiency is further optimized after being combined with clinical features.

Gestational age-related changes in relaxation times of neonatal brain by quantitative synthetic magnetic resonance imaging

OBJECTIVE

This study aimed to explore the correlation between T1 and T2 relaxation times of synthetic MRI (SyMRI) and gestational age (GA) in each hemisphere of preterm and term newborns at the initial 28 days of birth.

METHODS

Seventy preterm and full-term infants were prospectively included in this study. All subjects completed 3.0 T routine MRI and SyMRI (MAGiC) one-stop scanning within 28 days of birth (aged 34-42 W at examination). The SyMRI postprocessing software (v8.0.4) was used to measure the T1 and T2 relaxation values of each brain region. The linear regression equations of quantitative relaxation values with GA were established to compare the variation speed in each brain region.

RESULTS

A significant linear and negative correlation was found between relaxation times and GA in the neonate cerebral cortex and subcortical gray and white matter regions (All p<.05). The relaxation time of the left centrum semiovale decreased with maximum variance with increasing GA among all white matter regions (T1: b = -51.45, β = -0.65, p < .0001; T2: b = -8.77, β = -0.71, p < .0001), whereas the right posterior limb of internal capsule showed minimal variance (T1: b = -27.94, β = -0.60, p < .0001; T2: b = -3.25, β = -0.68, p < .0001). Among all gray matter regions, the right globus pallidus and thalamus indicated the most significant decreasing degree of T1 and T2 relaxation values with GA (right globus pallidus T1: b = -33.14, β = -0.64, p < .0001; right thalamus T2: b = -3.94, β = -0.81, p < .0001), and the right and left occipital lobes indicated the least significant decreasing degree of T1 and T2 relaxation values with GA, respectively (right occipital lobes T1: b = -11.18, β = -0.26, p = .028; left occipital lobes T2: b = -1.22, β = -0.27, p = .024).

CONCLUSIONS

SyMRI could quantitatively evaluate the linear changes of T1 and T2 relaxation values with GA in brain gray and white matter of preterm and term neonates.

Estimation of white matter hyperintensities with synthetic MRI myelin volume fraction in patients with multiple sclerosis and non-multiple-sclerosis white matter hyperintensities: A pilot study among the Indian population

AIM

Synthetic MRI (SyMRI) works on the MDME sequence, which acquires the relaxation properties of the brain and helps to measure the accurate tissue properties in 6 minutes. The aim of this study was to evaluate the synthetic MRI (SyMRI)-generated myelin (MyC) to white matter (WM) ratio, the WM fraction (WMF), MyC partial maps performing normative brain volumetry to investigate MyC loss in multiple sclerosis (MS) patients with white-matter hyperintensites (WMHs) and non-MS patients with WMHs in a clinical setting.

MATERIALS and METHODS

Synthetic MRI images were acquired from 15 patients with MS, and from 15 non-MS patients on a 3T MRI scanner (Discovery MR750w; GE Healthcare; Milwaukee, USA) using MAGiC, a customized version of SyntheticMR's SyMRI® IMAGE software marketed by GE Healthcare under a license agreement. Fast multi-delay multi-echo acquisition was performed with a 2D axial pulse sequence with different combinations of echo time (TEs) and saturation delay times. The total image acquisition time was 6 minutes. SyMRI image analysis was done using SyMRI software (SyMRI Version: 11.3.6; Synthetic MR, Linköping, Sweden). SyMRI data were used to generate the MyC partial maps and WMFs to quantify the signal intensities of test group and control group, andcontrol group , and their mean values were recorded. All patients also underwent conventional diffusion-weighted imaging, i.e., T1w and T2w imaging.

RESULTS

The results showed that the WMF was significantly lower in the test group than in the control group (38.8% vs 33.2%, p < 0.001). The Mann-Whitney U nonparametric t-test revealed a significant difference in the mean myelin volume between the test group and the control group (158.66 ± 32.31 vs. 138.29 ± 29.28, p = 0.044). Also, there were no significant differences in the gray matter fraction and intracranial volume between the test group and the control group.

CONCLUSIONS

We observed MyC loss in test group using quantitative SyMRI. Thus, myelin loss in MS patients can be quantitatively evaluated using SyMRI.

[Evaluation of the Accuracy of Relaxation Time Measurements Using 3D-QALAS at 3.0 T MRI and Comparison with 2D-MDME]

PURPOSE

Three-dimensional (3D) quantification using an interleaved Look-Locker acquisition sequence with a T₂ preparation pulse (QALAS) is a quantitative sequence used to measure relaxation times. The accuracy of the relaxation time measurement of 3D-QALAS at 3.0 T and the bias of 3D-QALAS have not yet been assessed. The purpose of this study was to clarify the accuracy of the relaxation time measurements using 3D-QALAS at 3.0 T MRI.

METHODS

The accuracy of the T₁ and T₂ values for 3D-QALAS was evaluated using a phantom. Subsequently, the T₁ and T₂ values and proton density of the brain parenchyma in healthy subjects were measured using 3D-QALAS and compared with those of 2D multi-dynamic multi-echo (MDME).

RESULTS

In the phantom study, the average T₁ value of 3D-QALAS was 8.3% prolonged than that for conventional inversion recovery spin-echo; the average T₂ value for 3D-QALAS was 18.4% shorter than that for multi-echo spin-echo. The in vivo assessment showed that the mean T₁ and T₂ values and PD for 3D-QALAS were prolonged by 5.3%, shortened by 9.6%, and increased by 7.0%, respectively, compared with those for 2D-MDME.

CONCLUSION

Although 3D-QALAS at 3.0 T has high accuracy T₁ value, which is less than 1000 ms, the T₁ value could be overestimated for tissues with it longer than that T₁ value. The T₂ value for 3D-QALAS could be underestimated for tissues with T₂ values, and this tendency increases with longer T₂ values.

Value of synthetic MRI quantitative parameters in preprocedural evaluation for TRUS/MRI fusion-guided biopsy of the prostate

BACKGROUND

Transrectal ultrasonography (TRUS)/magnetic resonance imaging (MRI) fusion-guided biopsy has a high clinical application value. However, this technique has some limitations, which limit its use in routine clinical practice. Therefore, the selection of suitable proatate lesions for this technique is worthy of our attention. Synthetic MRI (SyMRI) is capable of quantifying multiple relaxation parameters, which might have potential value in preprocedural evaluation for TRUS/MRI fusion-guided biopsy of the prostate. The aim of our study is to examine the value of SyMRI quantitative parameters in preprocedural evaluation for TRUS/MRI fusion-guided biopsy of the prostate.

METHODS

We prospectively selected 148 lesions in 137 patients who underwent prostate biopsy in our hospital. Next, 2-4 needles of TRUS/MRI fusion-guided biopsy combined with 10 needles of system biopsy (SB) were used as the protocol for prostate biopsy. Before biopsy, the MAGiC sequences of the MRI images of the enrolled patients underwent post-processing, and the longitudinal relaxation time (T1), transverse relaxation time (T2), and proton density (PD) were extracted. The biopsy pathology results were used as a gold standard to compare the differences in SyMRI quantitative parameters between benign and malignant prostate lesions in the peripheral and transitional zones. The receiver operating characteristic (ROC) curves were plotted to confirm the optimal SyMRI quantitative parameter for prostate lesion benignancy/malignancy performance, and the cutoff values of these parameters were used for grouping the lesions. The single-needle biopsy prostate cancer (PCa)-positivity rates (number of positive biopsy needles/total biopsy needles) and PCa overall detection rates by TRUS/MRI fusion-guided biopsy and SB were compared in different subgroups.

RESULTS

The T1 and T2 values can determine the benignancy/malignancy of prostate transition lesions(p < 0.01), and the T2 value has a greater diagnostic performance (p = 0.0376). The T2 value can determine the benignancy/malignancy of prostate peripheral lesions. The optimal diagnostic cutoff values for T2 were 77 and 81 ms, respectively. The single-needle PCa positivity rate of TRUS/MRI fusion-guided biopsy was higher than SB for any prostate lesions in different subgroups (p < 0.01). However, only in the subgroup of transition zone lesions with T2 ≤ 77 ms, the PCa overall detection rate of TRUS/MRI fusion-guided biopsy was significantly higher than that of SB (p = 0.031).

CONCLUSION

SyMRI-T2 value can provide a theoretical basis for the selection of suitable lesions for TRUS/MRI fusion-guided biopsy.

Personalized synthetic MR imaging with deep learning enhancements

PURPOSE

Personalized synthetic MRI (syn-MRI) uses MR images of an individual subject acquired at a few design parameters (echo time, repetition time, flip angle) to obtain underlying parametric ( ρ , T 1 , T 2 ) $$ \left(\rho, {\mathrm{T}}_1,{\mathrm{T}}_2\right) $$ maps, from where MR images of that individual at other design parameter settings are synthesized. However, classical methods that use least-squares (LS) or maximum likelihood estimators (MLE) are unsatisfactory at higher noise levels because the underlying inverse problem is ill-posed. This article provides a pipeline to enhance the synthesis of such images in three-dimensional (3D) using a deep learning (DL) neural network architecture for spatial regularization in a personalized setting where having more than a few training images is impractical.

METHODS

Our DL enhancements employ a Deep Image Prior (DIP) with a U-net type denoising architecture that includes situations with minimal training data, such as personalized syn-MRI. We provide a general workflow for syn-MRI from three or more training images. Our workflow, called DIPsyn-MRI, uses DIP to enhance training images, then obtains parametric images using LS or MLE before synthesizing images at desired design parameter settings. DIPsyn-MRI is implemented in our publicly available Python package DeepSynMRI available at: https://github.com/StatPal/DeepSynMRI.

RESULTS

We demonstrate feasibility and improved performance of DIPsyn-MRI on 3D datasets acquired using the Brainweb interface for spin-echo and FLASH imaging sequences, at different noise levels. Our DL enhancements improve syn-MRI in the presence of different intensity nonuniformity levels of the magnetic field, for all but very low noise levels.

CONCLUSION

This article provides recipes and software to realistically facilitate DL-enhanced personalized syn-MRI.

Toward deep learning replacement of gadolinium in neuro-oncology: A review of contrast-enhanced synthetic MRI

Gadolinium-based contrast agents (GBCAs) have become a crucial part of MRI acquisitions in neuro-oncology for the detection, characterization and monitoring of brain tumors. However, contrast-enhanced (CE) acquisitions not only raise safety concerns, but also lead to patient discomfort, the need of more skilled manpower and cost increase. Recently, several proposed deep learning works intend to reduce, or even eliminate, the need of GBCAs. This study reviews the published works related to the synthesis of CE images from low-dose and/or their native -non CE- counterparts. The data, type of neural network, and number of input modalities for each method are summarized as well as the evaluation methods. Based on this analysis, we discuss the main issues that these methods need to overcome in order to become suitable for their clinical usage. We also hypothesize some future trends that research on this topic may follow.

Synthetic MRI with Magnetic Resonance Spin TomogrAphy in Time-Domain (MR-STAT): Results from a Prospective Cross-Sectional Clinical Trial

BACKGROUND

Magnetic Resonance Spin TomogrAphy in Time-domain (MR-STAT) can reconstruct whole-brain multi-parametric quantitative maps (eg, T₁ , T₂ ) from a 5-minute MR acquisition. These quantitative maps can be leveraged for synthetization of clinical image contrasts.

PURPOSE

The objective was to assess image quality and overall diagnostic accuracy of synthetic MR-STAT contrasts compared to conventional contrast-weighted images.

STUDY TYPE

Prospective cross-sectional clinical trial.

POPULATION

Fifty participants with a median age of 45 years (range: 21-79 years) consisting of 10 healthy participants and 40 patients with neurological diseases (brain tumor, epilepsy, multiple sclerosis or stroke).

FIELD STRENGTH/SEQUENCE

3T/Conventional contrast-weighted imaging (T₁ /T₂ weighted, proton density [PD] weighted, and fluid-attenuated inversion recovery [FLAIR]) and a MR-STAT acquisition (2D Cartesian spoiled gradient echo with varying flip angle preceded by a non-selective inversion pulse).

ASSESSMENT

Quantitative T₁ , T₂ , and PD maps were computed from the MR-STAT acquisition, from which synthetic contrasts were generated. Three neuroradiologists blinded for image type and disease randomly and independently evaluated synthetic and conventional datasets for image quality and diagnostic accuracy, which was assessed by comparison with the clinically confirmed diagnosis.

STATISTICAL TESTS

Image quality and consequent acceptability for diagnostic use was assessed with a McNemar's test (one-sided α = 0.025). Wilcoxon signed rank test with a one-sided α = 0.025 and a margin of Δ = 0.5 on the 5-level Likert scale was used to assess non-inferiority.

RESULTS

All data sets were similar in acceptability for diagnostic use (≥3 Likert-scale) between techniques (T₁ w:P = 0.105, PDw:P = 1.000, FLAIR:P = 0.564). However, only the synthetic MR-STAT T₂ weighted images were significantly non-inferior to their conventional counterpart; all other synthetic datasets were inferior (T₁ w:P = 0.260, PDw:P = 1.000, FLAIR:P = 1.000). Moreover, true positive/negative rates were similar between techniques (conventional: 88%, MR-STAT: 84%).

DATA CONCLUSION

MR-STAT is a quantitative technique that may provide radiologists with clinically useful synthetic contrast images within substantially reduced scan time.

EVIDENCE LEVEL

1 Technical Efficacy: Stage 2.

Synthetic MRI improves radiomics-based glioblastoma survival prediction

Glioblastoma is an aggressive and fast-growing brain tumor with poor prognosis. Predicting the expected survival of patients with glioblastoma is a key task for efficient treatment and surgery planning. Survival predictions could be enhanced by means of a radiomic system. However, these systems demand high numbers of multicontrast images, the acquisitions of which are time consuming, giving rise to patient discomfort and low healthcare system efficiency. Synthetic MRI could favor deployment of radiomic systems in the clinic by allowing practitioners not only to reduce acquisition time, but also to retrospectively complete databases or to replace artifacted images. In this work we analyze the replacement of an actually acquired MR weighted image by a synthesized version to predict survival of glioblastoma patients with a radiomic system. Each synthesized version was realistically generated from two acquired images with a deep learning synthetic MRI approach based on a convolutional neural network. Specifically, two weighted images were considered for the replacement one at a time, a T2w and a FLAIR, which were synthesized from the pairs T1w and FLAIR, and T1w and T2w, respectively. Furthermore, a radiomic system for survival prediction, which can classify patients into two groups (survival >480 days and ≤ 480 days), was built. Results show that the radiomic system fed with the synthesized image achieves similar performance compared with using the acquired one, and better performance than a model that does not include this image. Hence, our results confirm that synthetic MRI does add to glioblastoma survival prediction within a radiomics-based approach.

The Boston ASL Template and Simulator: Initial development and implementation

BACKGROUND AND PURPOSE

Templates are a hallmark of image analysis in neuroimaging. However, while numerous structural templates exist and have facilitated single-subject and large group studies, templates based on functional contrasts, such as arterial spin labeling (ASL) perfusion, are scarce, have an inherently low spatial resolution, and are not as widely distributed. Having such tools at one's disposal is desirable, for example, in the case of studies not acquiring structural scans. We here propose an initial development of an ASL adult template based on high-resolution fast spin echo acquisitions.

METHODS

High-resolution single-delay ASL, low-resolution multi-delay ASL, T₁ -weighted magnetization prepared rapid acquisition 2 gradient echoes, and T₂ fluid attenuated inversion recovery data were acquired in a cohort of 10 healthy volunteers (6 males and 4 females, 30± 7 years old). After offline reconstruction of high-resolution perfusion arterial transit time (ATT) and T1 maps, we built a multi-contrast template relying on the Advanced Normalization Toolbox multivariate template nonlinear construction framework. We offer examples for the registration of ASL data acquired with different sequences. Finally, we propose an ASL simulator based on our templates and a standard kinetic model that allows generating synthetic ASL contrasts based on user-specified parameters.

RESULTS

Boston ASL Template and Simulator (BATS) offers high-quality, high-resolution perfusion-weighted and quantitative perfusion templates accompanied by ATT and different anatomical contrasts readily available in the Montreal Neurological Institute space. In addition, examples of use for data registration and as a synthetic contrast generator show various applications in which BATS could be used.

CONCLUSIONS

We propose a new ASL template collection, named BATS, that also includes a simulator allowing the generation of synthetic ASL contrasts. BATS is available at http://github.com/manueltaso/batsasltemplate.

Qualitative and Quantitative Performance of Magnetic Resonance Image Compilation (MAGiC) Method: An Exploratory Analysis for Head and Neck Imaging

The present exploratory study investigates the performance of a new, rapid, synthetic MRI method for diagnostic image quality assessment and measurement of relaxometry metric values in head and neck (HN) tumors and normal-appearing masseter muscle. The multi-dynamic multi-echo (MDME) sequence was used for data acquisition, followed by synthetic image reconstruction on a 3T MRI scanner for 14 patients (3 untreated and 11 treated). The MDME enables absolute quantification of physical tissue properties, including T1 and T2, with a shorter scan time than the current state-of-the-art methods used for relaxation measurements. The vendor termed the combined package MAGnetic resonance imaging Compilation (MAGiC). In total, 48 regions of interest (ROIs) were analyzed, drawn on normal-appearing masseter muscle and tumors in the HN region. Mean T1 and T2 values obtained from normal-appearing muscle were 880 ± 52 ms and 46 ± 3 ms, respectively. Mean T1 and T2 values obtained from tumors were 1930 ± 422 ms and 77 ± 13 ms, respectively, for the untreated group, 1745 ± 410 ms and 107 ± 61 ms, for the treated group. A total of 1552 images from both synthetic MRI and conventional clinical imaging were assessed by the radiologists to provide the rating for T1w and T2w image contrasts. The synthetically generated qualitative T2w images were acceptable and comparable to conventional diagnostic images (93% acceptability rating for both). The acceptability ratings for MAGiC-generated T1w, and conventional images were 64% and 100%, respectively. The benefit of MAGiC in HN imaging is twofold, providing relaxometry maps in a clinically feasible time and the ability to generate a different combination of contrast images in a single acquisition.

Synthetic MRI for Radiotherapy Planning for Brain and Prostate Cancers: Phantom Validation and Patient Evaluation

Purpose

We aimed to evaluate the accuracy of T₁ and T₂ mappings derived from a multispectral pulse sequence (magnetic resonance image compilation, MAGiC^®) on 1.5-T MRI and with conventional sequences [gradient echo with variable flip angle (GRE-VFA) and multi-echo spin echo (ME-SE)] compared to the reference values for the purpose of radiotherapy treatment planning.

Methods

The accuracy of T₁ and T₂ measurements was evaluated with 2 coils [head and neck unit (HNU) and BODY coils] on phantoms using descriptive statistics and Bland–Altman analysis. The reproducibility and repeatability of T₁ and T₂ measurements were performed on 15 sessions with the HNU coil. The T₁ and T₂ synthetic sequences obtained by both methods were evaluated according to quality assurance (QA) requirements for radiotherapy. T₁ and T₂in vivo measurements of the brain or prostate tissues of two groups of five subjects were also compared.

Results

The phantom results showed good agreement (mean bias, 8.4%) between the two measurement methods for T₁ values between 490 and 2,385 ms and T₂ values between 25 and 400 ms. MAGiC^® gave discordant results for T₁ values below 220 ms (bias with the reference values, from 38% to 1,620%). T₂ measurements were accurately estimated below 400 ms (mean bias, 8.5%) by both methods. The QA assessments are in agreement with the recommendations of imaging for contouring purposes for radiotherapy planning. On patient data of the brain and prostate, the measurements of T₁ and T₂ by the two quantitative MRI (qMRI) methods were comparable (max difference, <7%).

Conclusion

This study shows that the accuracy, reproducibility, and repeatability of the multispectral pulse sequence (MAGiC^®) were compatible with its use for radiotherapy treatment planning in a range of values corresponding to soft tissues. Even validated for brain imaging, MAGiC^® could potentially be used for prostate qMRI.

Uncertainty-Aware and Lesion-Specific Image Synthesis in Multiple Sclerosis Magnetic Resonance Imaging: A Multicentric Validation Study

Generative adversarial networks (GANs) can synthesize high-contrast MRI from lower-contrast input. Targeted translation of parenchymal lesions in multiple sclerosis (MS), as well as visualization of model confidence further augment their utility, provided that the GAN generalizes reliably across different scanners. We here investigate the generalizability of a refined GAN for synthesizing high-contrast double inversion recovery (DIR) images and propose the use of uncertainty maps to further enhance its clinical utility and trustworthiness. A GAN was trained to synthesize DIR from input fluid-attenuated inversion recovery (FLAIR) and T1w of 50 MS patients (training data). In another 50 patients (test data), two blinded readers (R1 and R2) independently quantified lesions in synthetic DIR (synthDIR), acquired DIR (trueDIR) and FLAIR. Of the 50 test patients, 20 were acquired on the same scanner as training data (internal data), while 30 were scanned at different scanners with heterogeneous field strengths and protocols (external data). Lesion-to-Background ratios (LBR) for MS-lesions vs. normal appearing white matter, as well as image quality parameters were calculated. Uncertainty maps were generated to visualize model confidence. Significantly more MS-specific lesions were found in synthDIR compared to FLAIR (R1: 26.7 ± 2.6 vs. 22.5 ± 2.2 p < 0.0001; R2: 22.8 ± 2.2 vs. 19.9 ± 2.0, p = 0.0005). While trueDIR remained superior to synthDIR in R1 [28.6 ± 2.9 vs. 26.7 ± 2.6 (p = 0.0021)], both sequences showed comparable lesion conspicuity in R2 [23.3 ± 2.4 vs. 22.8 ± 2.2 (p = 0.98)]. Importantly, improvements in lesion counts were similar in internal and external data. Measurements of LBR confirmed that lesion-focused GAN training significantly improved lesion conspicuity. The use of uncertainty maps furthermore helped discriminate between MS lesions and artifacts. In conclusion, this multicentric study confirms the external validity of a lesion-focused Deep-Learning tool aimed at MS imaging. When implemented, uncertainty maps are promising to increase the trustworthiness of synthetic MRI.

Generation of quantification maps and weighted images from synthetic magnetic resonance imaging using deep learning network

OBJECTIVE

The generation of quantification maps and weighted images in synthetic MRI techniques is based on complex fitting equations. This process requires longer image generation times. The objective of this study is to evaluate the feasibility of deep learning method for fast reconstruction of synthetic MRI.

APPROACH

A total of 44 healthy subjects were recruited and random divided into a training set (30 subjects) and a testing set (14 subjects). A multiple-dynamic, multiple-echo (MDME) sequence was used to acquire synthetic MRI images. Quantification maps (T1, T2, and proton density (PD) maps) and weighted (T1W, T2W, and T2W FLAIR) images were created with MAGiC software and then used as the ground truth images in the deep learning (DL) model. An improved multichannel U-Net structure network was trained to generate quantification maps and weighted images from raw synthetic MRI imaging data (8 module images). Quantitative evaluation was performed on quantification maps. Quantitative evaluation metrics, as well as qualitative evaluation were used in weighted image evaluation. Nonparametric Wilcoxon signed-rank tests were performed in this study.

MAIN RESULTS

The results of quantitative evaluation show that the error between the generated quantification images and the reference images is small. For weighted images, no significant difference in overall image quality or SNR was identified between DL images and synthetic images. Notably, the DL images achieved improved image contrast with T2W images, and fewer artifacts were present on DL images than synthetic images acquired by T2W FLAIR.

SIGNIFICANCE

The DL algorithm provides a promising method for image generation in synthetic MRI techniques, in which every step of the calculation can be optimized and faster, thereby simplifying the workflow of synthetic MRI techniques.

Simple parameters of synthetic MRI for assessment of bone density in patients with spinal degenerative disease

OBJECTIVE

Good bone quality is the key to avoiding osteoporotic fragility fractures and poor outcomes after lumbar instrumentation and fusion surgery. Although dual-energy x-ray absorptiometry (DEXA) screening is the current standard for evaluating osteoporosis, many patients lack DEXA measurements before undergoing lumbar spine surgery. The present study aimed to investigate the utility of using simple quantitative parameters generated with novel synthetic MRI to evaluate bone quality, as well as the correlations of these parameters with DEXA measurements.

METHODS

This prospective study enrolled patients with symptomatic lumbar degenerative disease who underwent DEXA and conventional and synthetic MRI. The quantitative parameters generated with synthetic MRI were T1 map, T2 map, T1 intensity, proton density (PD), and vertebral bone quality (VBQ) score, and these parameters were correlated with T-score of the lumbar spine.

RESULTS

There were 62 patients and 238 lumbar segments eligible for analysis. PD and VBQ score moderately correlated with T-score of the lumbar spine (r = -0.565 and -0.651, respectively; both p < 0.001). T1 intensity correlated fairly well with T-score (r = -0.411, p < 0.001). T1 and T2 correlated poorly with T-score. Receiver operating characteristic curve analysis demonstrated area under the curve values of 0.808 and 0.794 for detecting osteopenia/osteoporosis (T-score ≤ -1.0) and osteoporosis (T-score ≤ -2.5) with PD (both p < 0.001).

CONCLUSIONS

PD and T1 intensity values generated with synthetic MRI demonstrated significant correlation with T-score. PD has excellent ability for predicting osteoporosis and osteopenia.

The Effect of Intraventricular Hemorrhage on Brain Development in Premature Infants: A Synthetic MRI Study

Objectives: Synthetic MRI can obtain multiple parameters in one scan, including T1 and T2 relaxation time, proton density (PD), brain volume, etc. This study aimed to investigate the parameter values T1 and T2 relaxation time, PD, and volume characteristics of intraventricular hemorrhage (IVH) newborn brain, and the ability of synthetic MRI parameters T1 and T2 relaxation time and PD to diagnose IVH.

Materials and methods: The study included 50 premature babies scanned with conventional and synthetic MRI. Premature infants were allocated to the case group (n = 15) and NON IVH (n = 35). The T1, T2, PD values, and brain volume were obtained by synthetic MRI. Then we assessed the impact of IVH on these parameters.

Results: In the posterior limbs of the internal capsule (PLIC), genu of the corpus callosum (GCC), central white matter (CWM), frontal white matter (FWM), and cerebellum (each p < 0.05), the T1 and T2 relaxation times of the IVH group were significantly prolonged. There were significant differences also in PD. The brain volume in many parts were also significantly reduced, which was best illustrated in gray matter (GM), cerebrospinal fluid and intracranial volume, and brain parenchymal fraction (BPF) (each p < 0.001, t = −5.232 to 4.596). The differential diagnosis ability of these quantitative values was found to be excellent in PLIC, CWM, and cerebellum (AUC 0.700–0.837, p < 0.05).

Conclusion: The quantitative parameters of synthetic MRI show well the brain tissue characteristic values and brain volume changes of IVH premature infants. T1 and T2 relaxation times and PD contribute to the diagnosis and evaluation of IVH.

Automated delineation of head and neck organs at risk using synthetic MRI-aided mask scoring regional convolutional neural network

PURPOSE

Auto-segmentation algorithms offer a potential solution to eliminate the labor-intensive, time-consuming, and observer-dependent manual delineation of organs-at-risk (OARs) in radiotherapy treatment planning. This study aimed to develop a deep learning-based automated OAR delineation method to tackle the current challenges remaining in achieving reliable expert performance with the state-of-the-art auto-delineation algorithms.

METHODS

The accuracy of OAR delineation is expected to be improved by utilizing the complementary contrasts provided by computed tomography (CT) (bony-structure contrast) and magnetic resonance imaging (MRI) (soft-tissue contrast). Given CT images, synthetic MR images were firstly generated by a pre-trained cycle-consistent generative adversarial network. The features of CT and synthetic MRI were then extracted and combined for the final delineation of organs using mask scoring regional convolutional neural network. Both in-house and public datasets containing CT scans from head-and-neck (HN) cancer patients were adopted to quantitatively evaluate the performance of the proposed method against current state-of-the-art algorithms in metrics including Dice similarity coefficient (DSC), 95th percentile Hausdorff distance (HD95), mean surface distance (MSD), and residual mean square distance (RMS).

RESULTS

Across all of 18 OARs in our in-house dataset, the proposed method achieved an average DSC, HD95, MSD, and RMS of 0.77 (0.58-0.90), 2.90 mm (1.32-7.63 mm), 0.89 mm (0.42-1.85 mm), and 1.44 mm (0.71-3.15 mm), respectively, outperforming the current state-of-the-art algorithms by 6%, 16%, 25%, and 36%, respectively. On public datasets, for all nine OARs, an average DSC of 0.86 (0.73-0.97) were achieved, 6% better than the competing methods.

CONCLUSION

We demonstrated the feasibility of a synthetic MRI-aided deep learning framework for automated delineation of OARs in HN radiotherapy treatment planning. The proposed method could be adopted into routine HN cancer radiotherapy treatment planning to rapidly contour OARs with high accuracy.

Accelerated multicontrast reconstruction for synthetic MRI using joint parallel imaging and variable splitting networks

PURPOSE

Synthetic magnetic resonance imaging (MRI) requires the acquisition of multicontrast images to estimate quantitative parameter maps, such as T₁ , T₂ , and proton density (PD). The study aims to develop a multicontrast reconstruction method based on joint parallel imaging (JPI) and joint deep learning (JDL) to enable further acceleration of synthetic MRI.

METHODS

The JPI and JDL methods are extended and combined to improve reconstruction for better-quality, synthesized images. JPI is performed as a first step to estimate the missing k-space lines, and JDL is then performed to correct and refine the previous estimate with a trained neural network. For the JDL architecture, the original variable splitting network (VS-Net) is modified and extended to form a joint variable splitting network (JVS-Net) to apply to multicontrast reconstructions. The proposed method is designed and tested for multidynamic multiecho (MDME) images with Cartesian uniform under-sampling using acceleration factors between 4 and 8.

RESULTS

It is demonstrated that the normalized root-mean-square error (nRMSE) is lower and the structural similarity index measure (SSIM) values are higher with the proposed method compared to both the JPI and JDL methods individually. The method also demonstrates the potential to produce a set of synthesized contrast-weighted images that closely resemble those from the fully sampled acquisition without erroneous artifacts.

CONCLUSION

Combining JPI and JDL enables the reconstruction of highly accelerated synthetic MRIs.

Investigating the Image Quality and Utility of Synthetic MRI in the Breast

Purpose

Synthetic MRI reconstructs multiple sequences in a single acquisition. In the present study, we aimed to compare the image quality and utility of synthetic MRI with that of conventional MRI in the breast.

Methods

We retrospectively collected the imaging data of 37 women (mean age: 55.1 years; range: 20–78 years) who had undergone both synthetic and conventional MRI of T2-weighted, T1-weighted, and fat-suppressed (FS)-T2-weighted images. Two independent breast radiologists evaluated the overall image quality, anatomical sharpness, contrast between tissues, image homogeneity, and presence of artifacts of synthetic and conventional MRI on a 5-point scale (5 = very good to 1 = very poor). The interobserver agreement between the radiologists was evaluated using weighted kappa.

Results

For synthetic MRI, the acquisition time was 3 min 28 s. On the 5-point scale evaluation of overall image quality, although the scores of synthetic FS-T2-weighted images (4.01 ± 0.56) were lower than that of conventional images (4.95 ± 0.23; P < 0.001), the scores of synthetic T1- and T2-weighted images (4.95 ± 0.23 and 4.97 ± 0.16) were comparable with those of conventional images (4.92 ± 0.27 and 4.97 ± 0.16; P = 0.484 and 1.000, respectively). The kappa coefficient of conventional MRI was fair (0.53; P < 0.001), and that of conventional MRI was fair (0.46; P < 0.001).

Conclusion

The image quality of synthetic T1- and T2-weighted images was similar to that of conventional images and diagnostically acceptable, whereas the quality of synthetic T2-weighted FS images was inferior to conventional images. Although synthetic MRI images of the breast have the potential to provide efficient image diagnosis, further validation and improvement are required for clinical application.

Investigation of Synthetic Relaxometry and Diffusion Measures in the Differentiation of Benign and Malignant Breast Lesions as Compared to BI-RADS

BACKGROUND

Breast cancer is the most common malignant tumor in women and a quantitative contrast-free method is highly desirable for its diagnosis.

PURPOSE

To investigate the performance of quantitative MRI in differentiating malignant from benign breast lesions and to compare with the Breast Imaging Reporting and Data System (BI-RADS).

STUDY TYPE

Retrospective.

SUBJECTS

Eighty patients (56 with malignant lesions and 24 with benign lesions).

FIELD STRENGTH/SEQUENCE

Diffusion-weighted imaging (DWI) with a single-shot echo planar sequence and synthetic MRI with magnetic resonance image compilation (MAGiC) were performed at 3T.

ASSESSMENT

T₁ relaxation time (T₁ ), T₂ relaxation time (T₂ ), and proton density (PD) from synthetic MRI and apparent diffusion coefficient (ADC) from DWI were analyzed by two radiologists (Reader A, Reader B). Univariable and multivariable models were developed to optimize differentiation between malignant and benign lesions and their performances compared to BI-RADS.

STATISTICAL TESTS

The diagnostic performance was evaluated using multivariate logistic regression analysis and area under the receiver operating characteristic (ROC) curves (AUC).

RESULTS

T₂ , PD, and ADC values for malignant lesions were significantly lower than those in benign breast lesions for both radiologists (all P < 0.05). The combined T₂ , PD, and ADC model had the best performance for differentiating malignant and benign lesions with AUC, sensitivity, specificity, positive predictive value, and negative predictive values of 0.904, 94.6%, 87.5%, 94.6%, and 87.5%, respectively. The corresponding results for BI-RADS were no AUC, 94.6%, 75.0%, 89.8%, and 85.7%, respectively.

DATA CONCLUSION

The approach that combined synthetic MRI and DWI outperformed BI-RADS in the differential diagnosis of malignant and benign breast lesions and was achieved without contrast agents. This approach may serve as an alternative and effective strategy for the improvement of breast lesion differentiation.

LEVEL OF EVIDENCE

3.

TECHNICAL EFFICACY STAGE

3.

Synthetic MRI: Technologies and Applications in Neuroradiology

Synthetic MRI is a technique that synthesizes contrast-weighted images from multicontrast MRI data. There have been advances in synthetic MRI since the technique was introduced. Although a number of synthetic MRI methods have been developed for quantifying one or more relaxometric parameters and for generating multiple contrast-weighted images, this review focuses on several methods that quantify all three relaxometric parameters (T₁ , T₂ , and proton density) and produce multiple contrast-weighted images. Acquisition, quantification, and image synthesis techniques are discussed for each method. We discuss the image quality and diagnostic accuracy of synthetic MRI methods and their clinical applications in neuroradiology. Based on this analysis, we highlight areas that need to be addressed for synthetic MRI to be widely implemented in the clinic. LEVEL OF EVIDENCE: 5 TECHNICAL EFFICACY STAGE: 1.

Enhanced Masses on Contrast-Enhanced Breast: Differentiation Using a Combination of Dynamic Contrast-Enhanced MRI and Quantitative Evaluation with Synthetic MRI

BACKGROUND

The addition of synthetic MRI might improve the diagnostic performance of dynamic contrast-enhanced MRI (DCE-MRI) in patients with breast cancer.

PURPOSE

To evaluate the diagnostic value of a combination of DCE-MRI and quantitative evaluation using synthetic MRI for differentiation between benign and malignant breast masses.

STUDY TYPE

Retrospective, observational.

POPULATION

In all, 121 patients with 131 breast masses who underwent DCE-MRI with additional synthetic MRI were enrolled.

FIELD STRENGTH/SEQUENCE

3.0 Tesla, T₁ -weighted DCE-MRI and synthetic MRI acquired by a multiple-dynamic, multiple-echo sequence.

ASSESSMENT

All lesions were differentiated as benign or malignant using the following three diagnostic methods: DCE-MRI type based on the Breast Imaging-Reporting and Data System; synthetic MRI type using quantitative evaluation values calculated by synthetic MRI; and a combination of the DCE-MRI + Synthetic MRI types. The diagnostic performance of the three methods were compared.

STATISTICAL TESTS

Univariate (Mann-Whitney U-test) and multivariate (binomial logistic regression) analyses were performed, followed by receiver-operating characteristic curve (AUC) analysis.

RESULTS

Univariate and multivariate analyses showed that the mean T₁ relaxation time in a breast mass obtained by synthetic MRI prior to injection of contrast agent (pre-T₁ ) was the only significant quantitative value acquired by synthetic MRI that could independently differentiate between malignant and benign breast masses. The AUC for all enrolled breast masses assessed by DCE-MRI + Synthetic MRI type (0.83) was significantly greater than that for the DCE-MRI type (0.70, P < 0.05) or synthetic MRI type (0.73, P < 0.05). The AUC for category 4 masses assessed by the DCE-MRI + Synthetic MRI type was significantly greater than that for those assessed by the DCE-MRI type (0.74 vs. 0.50, P < 0.05).

DATA CONCLUSION

A combination of synthetic MRI and DCE-MRI improves the accuracy of diagnosis of benign and malignant breast masses, especially category 4 masses. Level of Evidence 4 Technical Efficacy Stage 2 J. MAGN. RESON. IMAGING 2021;53:381-391.

Pelvic multi-organ segmentation on cone-beam CT for prostate adaptive radiotherapy

BACKGROUND AND PURPOSE

The purpose of this study is to develop a deep learning-based approach to simultaneously segment five pelvic organs including prostate, bladder, rectum, left and right femoral heads on cone-beam CT (CBCT), as required elements for prostate adaptive radiotherapy planning.

MATERIALS AND METHODS

We propose to utilize both CBCT and CBCT-based synthetic MRI (sMRI) for the segmentation of soft tissue and bony structures, as they provide complementary information for pelvic organ segmentation. CBCT images have superior bony structure contrast and sMRIs have superior soft tissue contrast. Prior to segmentation, sMRI was generated using a cycle-consistent adversarial networks (CycleGAN), which was trained using paired CBCT-MR images. To combine the advantages of both CBCT and sMRI, we developed a cross-modality attention pyramid network with late feature fusion. Our method processes CBCT and sMRI inputs separately to extract CBCT-specific and sMRI-specific features prior to combining them in a late-fusion network for final segmentation. The network was trained and tested using 100 patients' datasets, with each dataset including the CBCT and manual physician contours. For comparison, we trained another two networks with different network inputs and architectures. The segmentation results were compared to manual contours for evaluations.

RESULTS

For the proposed method, dice similarity coefficients and mean surface distances between the segmentation results and the ground truth were 0.96 ± 0.03, 0.65 ± 0.67 mm; 0.91 ± 0.08, 0.93 ± 0.96 mm; 0.93 ± 0.04, 0.72 ± 0.61 mm; 0.95 ± 0.05, 1.05 ± 1.40 mm; and 0.95 ± 0.05, 1.08 ± 1.48 mm for bladder, prostate, rectum, left and right femoral heads, respectively. As compared to the other two competing methods, our method has shown superior performance in terms of the segmentation accuracy.

CONCLUSION

We developed a deep learning-based segmentation method to rapidly and accurately segment five pelvic organs simultaneously from daily CBCTs. The proposed method could be used in the clinic to support rapid target and organs-at-risk contouring for prostate adaptive radiation therapy.

Diagnosis and Grading of Prostate Cancer by Relaxation Maps From Synthetic MRI

BACKGROUND

The interpretation system for prostate MRI is largely based on qualitative image contrast of different tissue types. Therefore, a fast, standardized, and robust quantitative technique is necessary. Synthetic MRI is capable of quantifying multiple relaxation parameters, which might have potential applications in prostate cancer (PCa).

PURPOSE

To investigate the use of quantitative relaxation maps derived from synthetic MRI for the diagnosis and grading of PCa.

STUDY TYPE

Prospective.

SUBJECTS

In all, 94 men with pathologically confirmed PCa or benign pathological changes.

FIELD STRENGTH/SEQUENCE

T₁ -weighted imaging, T₂ -weighted imaging, diffusion-weighted imaging, and synthetic MRI at 3.0T.

ASSESSMENT

Four kinds of tissue types were identified on pathology, including PCa, stromal hyperplasia (SH), glandular hyperplasia (GH), and noncancerous peripheral zone (PZ). PCa foci were grouped as low-grade (LG, Gleason score ≤6) and intermediate/high-grade (HG, Gleason score ≥7). Regions of interest were manually drawn by two radiologists in consensus on parametric maps according to the pathological results.

STATISTICAL TESTS

Independent sample t-test, Mann-Whitney U-test, and receiver operating characteristic curve analysis.

RESULTS

T₁ and T₂ values of PCa were significantly lower than SH (P = 0.015 and 0.002). The differences of T₁ and T₂ values between PCa and noncancerous PZ were also significant (P ≤ 0.006). The area under the curve (AUC) of the apparent diffusion coefficient (ADC) value was significantly higher than T₁ , T₂ , and proton density (PD) values in discriminating PCa from SH and noncancerous PZ (P ≤ 0.025). T₂ , PD, and ADC values demonstrated similar diagnostic performance in discriminating LG from HG PCa (AUC = 0.806 [0.640-0.918], 0.717 [0.542-0.854], and 0.817 [0.652-0.925], respectively; P ≥ 0.535).

DATA CONCLUSION

Relaxation maps derived from synthetic MRI were helpful for discriminating PCa from other benign pathologies. But the overall diagnostic performance was inferior to the ADC values. T₂ , PD, and ADC values performed similarly in discriminating LG from HG PCa lesions.

LEVEL OF EVIDENCE

2 Technical Efficacy Stage: 2 J. Magn. Reson. Imaging 2020;52:552-564.

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

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.

[Consideration on SNR in Synthetic MRI]

Synthetic magnetic resonance imaging (MRI) can create different contrast weighted images by quantifying the T₁, T₂, and proton density values of the subjects from a single series of scan data. It has not been clarified how the signal to noise ratio (SNR) of the synthesized image varies depending on imaging parameters. We investigated the change of SNR in synthesized MR images by the experiment using self-made phantom. The SNR ratio of synthesized image by synthetic MRI showed the same tendency as the theoretical values due to parameter change in N_y, N_x, slice thickness, number of excitations. However, as for BW, the SNR ratio tended to be different from the theoretical values in some cases. In addition, it was suggested that the SNR of the composite image has relevance to the quantitative accuracy of the T₁, T₂, and proton density values. We thought that this is due to the image acquisition process by synthetic MRI.

[The Effect of Scan Parameters on the Synthetic MRI]

Synthetic MRI can provide proton density (PD), T₁ value, T₂ value for each pixel by only one data acquisition and can create various contrast-weighted images. The aim of this study is to evaluate the effect on the calculation of the T₁·T₂ value when changing the scan parameters for synthetic MRI. In the phantom study, when changing 1st TE/2nd TE/TR/TSE factor, the effect on the T₁·T₂ value calculated by synthetic MRI was examined. In the volunteer study, the brain was imaged and compared with known T₁·T₂ value. In phantom study, the effect on the T₂ value by the 1st TE/2nd TE/TSE factor was shown. In volunteer study, there was no problem in the calculated value of brain parenchyma. However, the T₂ value of cerebrospinal fluid had the error of known value. The results show that it is necessary to set appropriate scan parameters on synthetic MRI.

Contrast-enhanced synthetic MRI for the detection of brain metastases

Background

Synthetic magnetic resonance imaging (MRI), a technique that enables creation of various contrast-weighted images from a single MRI quantification scan, is a useful clinical tool. However, there are currently no reports examining the use of contrast-enhanced synthetic MRI for detecting brain metastases.

Purpose

To assess whether contrast-enhanced synthetic MRI is suitable for detecting brain metastases.

Material and Methods

Ten patients with a combined total of 167 brain metastases who underwent quantitative MRI and conventional T1-weighted inversion recovery fast spin-echo (conventional T1IR) MRI before and after administration of a contrast agent were included in the study. Synthetic T1IR and T1-weighted (synthetic T1W) images were produced after parameter quantification. Lesion-to-white matter contrast and contrast-to-noise ratio were calculated for each image. The number of visible lesions in each image was determined by two neuroradiologists.

Results

The mean lesion-to-white matter contrast and mean contrast-to-noise ratio of the synthetic T1IR images were significantly higher than those of the synthetic T1W (P < 0.001 and P < 0.001, respectively) and conventional T1IR (P = 0.04 and P = 0.002, respectively) images. Totals of 130 and 124 metastases were detected in the synthetic T1IR images by the first and second radiologists, respectively. The corresponding numbers were 91 and 85 in the synthetic T1W images and 119 and 119 in the conventional T1IR images. Statistical significance was not found among detected numbers of lesions.

Conclusion

Synthetic T1IR imaging created better contrast compared with synthetic T1W or conventional T1IR imaging. The ability to detect brain metastases was comparable among these imaging.