Multi-site, multi-platform comparison of MRI T1 measurement using the system phantom

Fine-Tuning Deep Learning Model for Quantitative Knee Joint Mapping With MR Fingerprinting and Its Comparison to Dictionary Matching Method: Fine-Tuning Deep Learning Model for Quantitative MRF

Magnetic resonance fingerprinting (MRF), as an emerging versatile and noninvasive imaging technique, provides simultaneous quantification of multiple quantitative MRI parameters, which have been used to detect changes in cartilage composition and structure in osteoarthritis. Deep learning (DL)-based methods for quantification mapping in MRF overcome the memory constraints and offer faster processing compared to the conventional dictionary matching (DM) method. However, limited attention has been given to the fine-tuning of neural networks (NNs) in DL and fair comparison with DM. In this study, we investigate the impact of training parameter choices on NN performance and compare the fine-tuned NN with DM for multiparametric mapping in MRF. Our approach includes optimizing NN hyperparameters, analyzing the singular value decomposition (SVD) components of MRF data, and optimization of the DM method. We conducted experiments on synthetic data, the NIST/ISMRM MRI system phantom with ground truth, and in vivo knee data from 14 healthy volunteers. The results demonstrate the critical importance of selecting appropriate training parameters, as these significantly affect NN performance. The findings also show that NNs improve the accuracy and robustness of T1, T2, and T1ρ mappings compared to DM in synthetic datasets. For in vivo knee data, the NN achieved comparable results for T1, with slightly lower T2 and slightly higher T1ρ measurements compared to DM. In conclusion, the fine-tuned NN can be used to increase accuracy and robustness for multiparametric quantitative mapping from MRF of the knee joint.

Multi-center QA of ultrahigh-field systems

Over the past two decades, ultra-high field (UHF) magnetic resonance imaging (MRI) has evolved from pure investigational devices to now systems with CE and FDA clearance for clinical use. UHF MRI offers enhanced diagnostic value, especially in brain and musculoskeletal imaging, aiding in the differential diagnosis of conditions like multiple sclerosis and epilepsy. However, to fully harness the potential of UHF, multi-center studies and quality assurance (QA) protocols are critical for ensuring reproducibility across different systems and sites. This becomes even more vital as the UHF community comprises three generations of magnet design, and many UHF sites are currently upgrading to the latest system architecture. Hence, this review presents multi-center QA measurements that have been performed at UHF, in particular from larger consortia through their "travelling heads" studies. Despite the technical variability between different vendors and system generations, these studies have shown a high level of reproducibility in structural and quantitative imaging. Furthermore, the review highlights the ongoing challenges in QA, such as transmitter performance drift and the need for a standard reliable multi-tissue phantom for RF coil calibration, which are crucial for advancing UHF MRI in both clinical and research applications.

Paramagnetic salt and agarose recipes for phantoms with desired T1 and T2 values for low-field MRI

Tissue-mimicking reference phantoms are indispensable for the development and optimization of magnetic resonance (MR) measurement sequences. Phantoms have greatest utility when they mimic the MR signals arising from tissue physiology; however, many of the properties underlying these signals, including tissue relaxation characteristics, can vary as a function of magnetic field strength. There has been renewed interest in magnetic resonance imaging (MRI) at field strengths less than 1 T, and phantoms developed for higher field strengths may not be physiologically relevant at these lower fields. This work focuses on developing materials with specific relaxation properties for lower magnetic field strengths. Specifically, we developed recipes that can be used to create synthetic samples for target nuclear magnetic resonance relaxation values for fields between 0.0065 and 0.55 T.T 1 $$ {T}_1 $$ andT 2 $$ {T}_2 $$ mixing models for agarose-based gels doped with a paramagnetic salt (one of CuSO4, GdCl3, MnCl2, or NiCl2) were created using relaxation measurements of synthetic gel samples at 0.0065, 0.064, and 0.55 T. Measurements were evaluated for variability with respect to measurement repeatability and changing synthesis protocol or laboratory temperature. The mixing models were used to identify formulations of agarose and salt composition to approximately mimic the relaxation times of five neurological tissues (blood, cerebrospinal fluid, fat, gray matter, and white matter) at 0.0065, 0.0475, 0.05, 0.064, and 0.55 T. These mimic sample formulations were measured at each field strength. Of these samples, the GdCl3 and NiCl2 measurements were closest to the target tissue relaxation times. The GdCl3 or NiCl2 mixing model recipes are recommended for creating target relaxation samples below 0.55 T. This work can help development of MRI methods and applications for low-field systems and applications.

A signal model for fat-suppressed T1-mapping and dynamic contrast-enhanced MRI with interrupted spoiled gradient-echo readout

The conventional gradient-echo steady-state signal model is the basis of various spoiled gradient-echo (SPGR) based quantitative MRI models, including variable flip angle (VFA) MRI and dynamic contrast-enhanced MRI (DCE). However, including preparation pulses, such as fat suppression or saturation bands, disrupts the steady-state and leads to a bias in T1 and DCE parameter estimates. This work introduces a signal model that improves the accuracy of VFA T1-mapping and DCE for interrupted spoiled gradient-echo (I-SPGR) acquisitions. The proposed model was applied to a VFA T1-mapping I-SPGR sequence in the Gold Standard T1-phantom (3 T), in the brain of four healthy volunteers (3 T), and to an abdominal DCE examination (1.5 T). T1-values obtained with the proposed and conventional model were compared to reference T1-values. Bland-Altman analysis (phantom) and analysis of variance (in vivo) were used to test whether bias from both methods was significantly different (p = 0.05). The proposed model outperformed the conventional model by decreasing the bias in the phantom with respect to the phantom reference values (mean bias -2 vs. -35% at 3 T) and in vivo with respect to the conventional SPGR (-6 vs. -37% bias in T1, p < 0.01). The proposed signal model estimated approximately 48% (depending on baseline T1) higher contrast concentrations in vivo, which resulted in decreased DCE pharmacokinetic parameter estimates of up to 35%. The proposed signal model improves the accuracy of quantitative parameter estimation from disrupted steady-state I-SPGR sequences. It therefore provides a flexible method for applying fat suppression, saturation bands, and other preparation pulses in VFA T1-mapping and DCE.

T1 and T2 measurements across multiple 0.55T MRI systems using open-source vendor-neutral sequences

PURPOSE

To compare T1 and T2 measurements across commercial and prototype 0.55T MRI systems in both phantom and healthy participants using the same vendor-neutral pulse sequences, reconstruction, and analysis methods.

METHODS

Standard spin echo measurements and abbreviated protocol measurements of T1, B1, and T2 were made on two prototype 0.55 T systems and two commercial 0.55T systems using an ISMRM/NIST system phantom. Additionally, five healthy participants were imaged at each system using the abbreviated protocol for T1, B1, and T2 measurement. The phantom measurements were compared to NMR-based reference measurements to determine accuracy, and both phantom and in vivo measurements were compared to assess reproducibility and differences between the prototype and commercial systems.

RESULTS

Vendor-neutral sequences were implemented across all four systems, and the code for pulse sequences and reconstruction is freely available. For participants, there was no difference in the mean T1 and T2 relaxation times between the prototype and commercial systems. In the phantom, there were no significant differences between the prototype and commercial systems for T1 and T2 measurements using the abbreviated protocol.

CONCLUSION

Quantitative T1 and T2 measurements at 0.55T in phantom and healthy participants are not statistically different across the prototype and commercial systems.

Repeatability of 3D MR fingerprinting during scanner software upgrades

OBJECTIVE

This study aims to quantify the repeatability of a 3D Magnetic Resonance Fingerprinting (MRF) research protocol in the context of a scanner software upgrade. All of MRI assumes consistent hardware performance and raw data pre-processing on the acquisition side. Software upgrades can affect hardware specifications and reconstruction chain parameters. Understanding how vendor-provided software upgrades vary MRF-derived T1 and T2 values is crucial for its application in different settings.

MATERIALS AND METHODS

Eight healthy volunteers were imaged with an in-house developed 3D MRF pulse sequence using a 3T scanner before and after a software upgrade (VA31A to VA50A, MAGNETOM Vida, Siemens Healthineers). Online MRF reconstruction using Singular Value Decomposition (SVD) timeseries compression and B1+ correction was performed. The study involved test-retest repeatability assessment and a comparison of pre- and post-upgrade data based on automatically extracted T1 and T2 values from MNI-152 Harvard-Oxford Subcortical Structural Atlas regions.

RESULTS

Significant mismatches were found directly after the upgrade. However, after an information exchange with the vendor, the 3D-MRF sequence showed consistent repeatability in both intra-version test-retest scenarios and cross-version comparisons: - 1.16 ± 3.18% variability in T1 and - 0.54 ± 4.84% in T2 for intra-version tests, and - 0.83 ± 3.68% (T1) and - 0.05 ± 5.81% (T2) variability for cross-version comparisons.

DISCUSSION

The study shows the reliable performance of 3D MRF protocols across software upgrades is possible, but it also highlights the importance of detailed evaluation and vendor collaboration in ensuring consistency. These findings support the application of MRF in longitudinal studies and emphasize the need for systematic assessments following hardware or software modifications.

Optimization and validation of low-field MP2RAGE T1 mapping on 0.35T MR-Linac: Toward adaptive dose painting with hypoxia biomarkers

BACKGROUND

Stereotactic MR-guided Adaptive Radiation Therapy (SMART) dose painting for hypoxia has potential to improve treatment outcomes, but clinical implementation on low-field MR-Linac faces substantial challenges due to dramatically lower signal-to-noise ratio (SNR) characteristics. While quantitative MRI and T1 mapping of hypoxia biomarkers show promise, T1-to-noise ratio (T1NR) optimization at low fields is paramount, particularly for the clinical implementation of oxygen-enhanced (OE)-MRI. The 3D Magnetization Prepared (2) Rapid Gradient Echo (MP2RAGE) sequence stands out for its ability to acquire homogeneous T1-weighted contrast images with simultaneous T1 mapping.

PURPOSE

To optimize MP2RAGE for low-field T1 mapping; conduct experimental validation in a ground-truth phantom; establish feasibility and reproducibility of low-field MP2RAGE acquisition and T1 mapping in healthy volunteers.

METHODS

The MP2RAGE optimization was performed to maximize the contrast-to-noise ratio (CNR) of T1 values in white matter (WM) and gray matter (GM) brain tissues at 0.35T. Low-field MP2RAGE images were acquired on a 0.35T MR-Linac (ViewRay MRIdian) using a multi-channel head coil. Validation of T1 mapping was performed with a ground-truth Eurospin phantom, containing inserts of known T1 values (400-850 ms), with one and two average (1A and 2A) MP2RAGE scans across four acquisition sessions, resulting in eight T1 maps. Mean (± SD) T1 relative error, T1NR, and intersession coefficient of variation (CV) were determined. Whole-brain MP2RAGE scans were acquired in 5 healthy volunteers across two sessions (A and B) and T1 maps were generated. Mean (± SD) T1 values for WM and GM were determined. Whole-brain T1 histogram analysis was performed, and reproducibility was determined with the CV between sessions. Voxel-by-voxel T1 difference maps were generated to evaluate 3D spatial variation.

RESULTS

Low-field MP2RAGE optimization resulted in parameters: MP2RAGETR of 3250 ms, inversion times (TI1/TI2) of 500/1200 ms, and flip angles (α1/α2) of 7/5°. Eurospin T1 maps exhibited a mean (± SD) relative error of 3.45% ± 1.30%, T1NR of 20.13 ± 5.31, and CV of 2.22% ± 0.67% across all inserts. Whole-brain MP2RAGE images showed high anatomical quality with clear tissue differentiation, resulting in mean (± SD) T1 values: 435.36 ± 10.01 ms for WM and 623.29 ± 14.64 ms for GM across subjects, showing excellent concordance with literature. Whole-brain T1 histograms showed high intrapatient and intersession reproducibility with characteristic intensity peaks consistent with voxel-level WM and GM T1 values. Reproducibility analysis revealed a CV of 0.46% ± 0.31% and 0.35% ± 0.18% for WM and GM, respectively. Voxel-by-voxel T1 difference maps show a normal 3D spatial distribution of noise in WM and GM.

CONCLUSIONS

Low-field MP2RAGE proved effective in generating accurate, reliable, and reproducible T1 maps with high T1NR in phantom studies and in vivo feasibility established in healthy volunteers. While current work is focused on refining the MP2RAGE protocol to enable clinically efficient OE-MRI, this study establishes a foundation for TOLD T1 mapping for hypoxia biomarkers. This advancement holds the potential to facilitate a paradigm shift toward MR-guided biological adaptation and dose painting by leveraging 3D hypoxic spatial distributions and improving outcomes in conventionally challenging-to-treat cancers.

Repeat it without me: Crowdsourcing the T1 mapping common ground via the ISMRM reproducibility challenge

PURPOSE

T1 mapping is a widely used quantitative MRI technique, but its tissue-specific values remain inconsistent across protocols, sites, and vendors. The ISMRM Reproducible Research and Quantitative MR study groups jointly launched a challenge to assess the reproducibility of a well-established inversion-recovery T1 mapping technique, using acquisition details from a seminal T1 mapping paper on a standardized phantom and in human brains.

METHODS

The challenge used the acquisition protocol from Barral et al. (2010). Researchers collected T1 mapping data on the ISMRM/NIST phantom and/or in human brains. Data submission, pipeline development, and analysis were conducted using open-source platforms. Intersubmission and intrasubmission comparisons were performed.

RESULTS

Eighteen submissions (39 phantom and 56 human datasets) on scanners by three MRI vendors were collected at 3 T (except one, at 0.35 T). The mean coefficient of variation was 6.1% for intersubmission phantom measurements, and 2.9% for intrasubmission measurements. For humans, the intersubmission/intrasubmission coefficient of variation was 5.9/3.2% in the genu and 16/6.9% in the cortex. An interactive dashboard for data visualization was also developed: https://rrsg2020.dashboards.neurolibre.org.

CONCLUSION

The T1 intersubmission variability was twice as high as the intrasubmission variability in both phantoms and human brains, indicating that the acquisition details in the original paper were insufficient to reproduce a quantitative MRI protocol. This study reports the inherent uncertainty in T1 measures across independent research groups, bringing us one step closer to a practical clinical baseline of T1 variations in vivo.

Phantoms for Quantitative Body MRI: a review and discussion of the phantom value

In this paper, we review the value of phantoms for body MRI in the context of their uses for quantitative MRI methods research, clinical trials, and clinical imaging. Certain uses of phantoms are common throughout the body MRI community, including measuring bias, assessing reproducibility, and training. In addition to these uses, phantoms in body MRI methods research are used for novel methods development and the design of motion compensation and mitigation techniques. For clinical trials, phantoms are an essential part of quality management strategies, facilitating the conduct of ethically sound, reliable, and regulatorily compliant clinical research of both novel MRI methods and therapeutic agents. In the clinic, phantoms are used for development of protocols, mitigation of cost, quality control, and radiotherapy. We briefly review phantoms developed for quantitative body MRI, and finally, we review open questions regarding the most effective use of a phantom for body MRI.

Deep Learning Prostate MRI Segmentation Accuracy and Robustness: A Systematic Review

Purpose To investigate the accuracy and robustness of prostate segmentation using deep learning across various training data sizes, MRI vendors, prostate zones, and testing methods relative to fellowship-trained diagnostic radiologists. Materials and Methods In this systematic review, Embase, PubMed, Scopus, and Web of Science databases were queried for English-language articles using keywords and related terms for prostate MRI segmentation and deep learning algorithms dated to July 31, 2022. A total of 691 articles from the search query were collected and subsequently filtered to 48 on the basis of predefined inclusion and exclusion criteria. Multiple characteristics were extracted from selected studies, such as deep learning algorithm performance, MRI vendor, and training dataset features. The primary outcome was comparison of mean Dice similarity coefficient (DSC) for prostate segmentation for deep learning algorithms versus diagnostic radiologists. Results Forty-eight studies were included. Most published deep learning algorithms for whole prostate gland segmentation (39 of 42 [93%]) had a DSC at or above expert level (DSC ≥ 0.86). The mean DSC was 0.79 ± 0.06 (SD) for peripheral zone, 0.87 ± 0.05 for transition zone, and 0.90 ± 0.04 for whole prostate gland segmentation. For selected studies that used one major MRI vendor, the mean DSCs of each were as follows: General Electric (three of 48 studies), 0.92 ± 0.03; Philips (four of 48 studies), 0.92 ± 0.02; and Siemens (six of 48 studies), 0.91 ± 0.03. Conclusion Deep learning algorithms for prostate MRI segmentation demonstrated accuracy similar to that of expert radiologists despite varying parameters; therefore, future research should shift toward evaluating segmentation robustness and patient outcomes across diverse clinical settings. Keywords: MRI, Genital/Reproductive, Prostate Segmentation, Deep Learning Systematic review registration link: osf.io/nxaev © RSNA, 2024.

Detectability and accuracy of computational measurements of in-silico and physical representations of enlarged perivascular spaces from magnetic resonance images

BACKGROUND

Magnetic Resonance Imaging (MRI) visible perivascular spaces (PVS) have been associated with age, decline in cognitive abilities, interrupted sleep, and markers of small vessel disease. But the limits of validity of their quantification have not been established.

NEW METHOD

We use a purpose-built digital reference object to construct an in-silico phantom for addressing this need, and validate it using a physical phantom. We use cylinders of different sizes as models for PVS. We also evaluate the influence of 'PVS' orientation, and different sets of parameters of the two vesselness filters that have been used for enhancing tubular structures, namely Frangi and RORPO filters, in the measurements' accuracy.

RESULTS

PVS measurements in MRI are only a proxy of their true dimensions, as the boundaries of their representation are consistently overestimated. The success in the use of the Frangi filter relies on a careful tuning of several parameters. Alpha= 0.5, beta= 0.5 and c= 500 yielded the best results. RORPO does not have these requirements and allows detecting smaller cylinders in their entirety more consistently in the absence of noise and confounding artefacts. The Frangi filter seems to be best suited for voxel sizes equal or larger than 0.4 mm-isotropic and cylinders larger than 1 mm diameter and 2 mm length. 'PVS' orientation did not affect measurements in data with isotropic voxels.

COMPARISON WITH EXISTENT METHODS

Does not apply.

CONCLUSIONS

The in-silico and physical phantoms presented are useful for establishing the validity of quantification methods of tubular small structures.

Dissociation of Reliability, Heritability, and Predictivity in Coarse- and Fine-Scale Functional Connectomes during Development

The functional connectome supports information transmission through the brain at various spatial scales, from exchange between broad cortical regions to finer-scale, vertex-wise connections that underlie specific information processing mechanisms. In adults, while both the coarse- and fine-scale functional connectomes predict cognition, the fine scale can predict up to twice the variance as the coarse-scale functional connectome. Yet, past brain-wide association studies, particularly using large developmental samples, focus on the coarse connectome to understand the neural underpinnings of individual differences in cognition. Using a large cohort of children (age 9-10 years; n = 1,115 individuals; both sexes; 50% female, including 170 monozygotic and 219 dizygotic twin pairs and 337 unrelated individuals), we examine the reliability, heritability, and behavioral relevance of resting-state functional connectivity computed at different spatial scales. We use connectivity hyperalignment to improve access to reliable fine-scale (vertex-wise) connectivity information and compare the fine-scale connectome with the traditional parcel-wise (coarse scale) functional connectomes. Though individual differences in the fine-scale connectome are more reliable than those in the coarse-scale, they are less heritable. Further, the alignment and scale of connectomes influence their ability to predict behavior, whereby some cognitive traits are equally well predicted by both connectome scales, but other, less heritable cognitive traits are better predicted by the fine-scale connectome. Together, our findings suggest there are dissociable individual differences in information processing represented at different scales of the functional connectome which, in turn, have distinct implications for heritability and cognition.

New clinical opportunities of low-field MRI: heart, lung, body, and musculoskeletal

Contemporary whole-body low-field MRI scanners (< 1 T) present new and exciting opportunities for improved body imaging. The fundamental reason is that the reduced off-resonance and reduced SAR provide substantially increased flexibility in the design of MRI pulse sequences. Promising body applications include lung parenchyma imaging, imaging adjacent to metallic implants, cardiac imaging, and dynamic imaging in general. The lower cost of such systems may make MRI favorable for screening high-risk populations and population health research, and the more open configurations allowed may prove favorable for obese subjects and for pregnant women. This article summarizes promising body applications for contemporary whole-body low-field MRI systems, with a focus on new platforms developed within the past 5 years. This is an active area of research, and one can expect many improvements as MRI physicists fully explore the landscape of pulse sequences that are feasible, and as clinicians apply these to patient populations.

Efficiency and Accuracy Evaluation of Multiple Diffusion-Weighted MRI Techniques Across Different Scanners

BACKGROUND

The choice between different diffusion-weighted imaging (DWI) techniques is difficult as each comes with tradeoffs for efficient clinical routine imaging and apparent diffusion coefficient (ADC) accuracy.

PURPOSE

To quantify signal-to-noise-ratio (SNR) efficiency, ADC accuracy, artifacts, and distortions for different DWI acquisition techniques, coils, and scanners.

STUDY TYPE

Phantom, in vivo intraindividual biomarker accuracy between DWI techniques and independent ratings.

POPULATION/PHANTOMS

NIST diffusion phantom. 51 Patients: 40 with prostate cancer and 11 with head-and-neck cancer at 1.5 T FIELD STRENGTH/SEQUENCE: Echo planar imaging (EPI): 1.5 T and 3 T Siemens; 3 T Philips. Distortion-reducing: RESOLVE (1.5 and 3 T Siemens); Turbo Spin Echo (TSE)-SPLICE (3 T Philips). Small field-of-view (FOV): ZoomitPro (1.5 T Siemens); IRIS (3 T Philips). Head-and-neck and flexible coils.

ASSESSMENT

SNR Efficiency, geometrical distortions, and susceptibility artifacts were quantified for different b-values in a phantom. ADC accuracy/agreement was quantified in phantom and for 51 patients. In vivo image quality was independently rated by four experts.

STATISTICAL TESTS

QIBA methodology for accuracy: trueness, repeatability, reproducibility, Bland-Altman 95% Limits-of-Agreement (LOA) for ADC. Wilcoxon Signed-Rank and student tests on P < 0.05 level.

RESULTS

The ZoomitPro small FOV sequence improved b-image efficiency by 8%-14%, reduced artifacts and observer scoring for most raters at the cost of smaller FOV compared to EPI. The TSE-SPLICE technique reduced artifacts almost completely at a 24% efficiency cost compared to EPI for b-values ≤500 sec/mm2 . Phantom ADC 95% LOA trueness were within ±0.03 × 10-3  mm2 /sec except for small FOV IRIS. The in vivo ADC agreement between techniques, however, resulted in 95% LOAs in the order of ±0.3 × 10-3  mm2 /sec with up to 0.2 × 10-3  mm2 /sec of bias.

DATA CONCLUSION

ZoomitPro for Siemens and TSE SPLICE for Philips resulted in a trade-off between efficiency and artifacts. Phantom ADC quality control largely underestimated in vivo accuracy: significant ADC bias and variability was found between techniques in vivo.

LEVEL OF EVIDENCE

3 TECHNICAL EFFICACY STAGE: 2.

Non-contrast based approach for liver function quantification using Bayesian-based intravoxel incoherent motion diffusion weighted imaging: A pilot study

PURPOSE

Liver cirrhosis disrupts liver function and tissue perfusion, detectable by magnetic resonance imaging (MRI). Assessing liver function at the voxel level with 13-b value intravoxel incoherent motion diffusion-weighted imaging (IVIM-DWI) could aid in radiation therapy liver-sparing treatment for patients with early impairment. This study aimed to evaluate the feasibility of IVIM-DWI for liver function assessment and correlate it with other multiparametric (mp) MRI methods at the voxel level.

METHOD

This study investigates the variability of apparent diffusion coefficient (ADC) derived from 13-b value IVIM-DWI and B1-corrected dual flip angle (DFA) T1 mapping. Experiments were conducted in-vitro with QIBA and NIST phantoms and in 10 healthy volunteers for IVIM-DWI. Additionally, 12 patients underwent an mp-MRI examination. The imaging protocol included a 13-b value IVIM-DWI sequence for generating IVIM parametric maps. B1-corrected DFA T1 pulse sequence was used for generating T1 maps, and Gadoxatate low temporal resolution dynamic contrast-enhanced (LTR-DCE) MRI was used for generating the Hepatic extraction fraction (HEF) map. The Mann-Whitney U test was employed to compare IVIM-DWI parameters (Pure Diffusion, Dslow ; Pseudo diffusion, Dfast ; and Perfusion Fraction, Fp ) between the healthy volunteer and patient groups. Furthermore, in the patient group, statistical correlations were assessed at a voxel level between LTR-DCE MRI-derived HEF, T1 post-Gadoxetate administration, ΔT1%, and various IVIM parameters using Pearson correlation.

RESULTS

For-vitro measurements, the maximum coefficient of variation of the ADC and T1 parameters was 12.4% and 16.1%, respectively. The results also showed that Fp and Dfast were able to distinguish between healthy liver function and mild liver function impairment at the global level, with p = 0.002 for Fp and p < 0.001 for Dfast . Within the patient group, these parameters also exhibited a moderate correlation with HEF at the voxel level.

CONCLUSION

Overall, the study highlighted the potential of Dfast and Fp for detecting liver function impairment at both global and pixel levels.

Challenges of implementing computer-aided diagnostic models for neuroimages in a clinical setting

Advances in artificial intelligence have cultivated a strong interest in developing and validating the clinical utilities of computer-aided diagnostic models. Machine learning for diagnostic neuroimaging has often been applied to detect psychological and neurological disorders, typically on small-scale datasets or data collected in a research setting. With the collection and collation of an ever-growing number of public datasets that researchers can freely access, much work has been done in adapting machine learning models to classify these neuroimages by diseases such as Alzheimer's, ADHD, autism, bipolar disorder, and so on. These studies often come with the promise of being implemented clinically, but despite intense interest in this topic in the laboratory, limited progress has been made in clinical implementation. In this review, we analyze challenges specific to the clinical implementation of diagnostic AI models for neuroimaging data, looking at the differences between laboratory and clinical settings, the inherent limitations of diagnostic AI, and the different incentives and skill sets between research institutions, technology companies, and hospitals. These complexities need to be recognized in the translation of diagnostic AI for neuroimaging from the laboratory to the clinic.

In vivo quantitative MRI: T1 and T2 measurements of the human brain at 0.064 T

OBJECTIVE

To measure healthy brain [Formula: see text] and [Formula: see text] relaxation times at 0.064 T.

MATERIALS AND METHODS

[Formula: see text] and [Formula: see text] relaxation times were measured in vivo for 10 healthy volunteers using a 0.064 T magnetic resonance imaging (MRI) system and for 10 test samples on both the MRI and a separate 0.064 T nuclear magnetic resonance (NMR) system. In vivo [Formula: see text] and [Formula: see text] values are reported for white matter (WM), gray matter (GM), and cerebrospinal fluid (CSF) for automatic segmentation regions and manual regions of interest (ROIs).

RESULTS

[Formula: see text] sample measurements on the MRI system were within 10% of the NMR measurement for 9 samples, and one sample was within 11%. Eight [Formula: see text] sample MRI measurements were within 25% of the NMR measurement, and the two longest [Formula: see text] samples had more than 25% variation. Automatic segmentations generally resulted in larger [Formula: see text] and [Formula: see text] estimates than manual ROIs.

DISCUSSION

[Formula: see text] and [Formula: see text] times for brain tissue were measured at 0.064 T. Test samples demonstrated accuracy in WM and GM ranges of values but underestimated long [Formula: see text] in the CSF range. This work contributes to measuring quantitative MRI properties of the human body at a range of field strengths.

T1 relaxation time of ISMRM/NIST T1 phantom spheres at 7 T

T1 relaxation times of the 14 T1 phantom spheres that make up the standard International Society for Magnetic Resonance in Medicine (ISMRM)/National Institute of Standards and Technology (NIST) system phantom are reported at 7 T. T1 values of six of the 14 T1 spheres at 7 T (with T1 > 270 ms) have been reported previously, but, to the best of our knowledge, not all of the T1s of the 14 T1 spheres at 7 T have been reported before. Given the increasing number of 7-T MRI systems in clinical settings and the increasing need for T1 phantoms that cover a wide range of T1 relaxation times to evaluate rapid T1 mapping techniques at 7 T, it is of high interest to obtain accurate T1 values for all the ISMRM/NIST T1 spheres at 7 T. In this work, T1 relaxation time was measured on a 7-T MRI scanner using an inversion-recovery spin-echo pulse sequence and derived by curve fitting to a signal equation that exhibits insensitivity to B 1 + inhomogeneity. Day-to-day reproducibility was within 0.4% and differences between two different RF coils within 1.5%. T1s of a subset of the 14 spheres were also measured by NMR at 7 T for comparison, and the T1 results were consistent between the MRI and NMR measurements. T1 measurements performed at 3 T on the same 14 spheres using the same sequence and fitting method yielded good agreement (mean percentage difference of -0.4%) with the reference T1 values available from the NIST, reflecting the accuracy of the reported technique despite being without the standard phantom housing. We found that the T1 values of all 14 NiCl₂ spheres are consistently lower at 7 T than at 3 T. Although our results were well reproduced, this study represents initial work to quantify the 7-T T1 values of all 14 NIST T1 spheres outside of the standard housing and does not warrant reproducibility of the ISMRM/NIST system phantom as a whole. A future study to assess the T1 values of a version of the ISMRM/NIST system phantom that fits inside typical commercial coils at 7 T will be very helpful. Nonetheless, the details on our acquisition and curve-fitting methods reported here allow the T1 measurements to be reproduced elsewhere. The T1 values of all 14 spheres reported here will be valuable for the development of quantitative MR fingerprinting and rapid T1 mapping for a large variety of research projects, not only in neuroimaging but also in body MRI, musculoskeletal MRI, and gadolinium contrast-enhanced MRI, each of which is concerned with much shortened T1.

A correlational study between microstructural, macrostructural and functional age-related changes in the human visual cortex

Age-related changes in the human brain can be investigated from either structural or functional perspectives. Analysis of structural and functional age-related changes throughout the lifespan may help to understand the normal brain development process and monitor the structural and functional pathology of the brain. This study, combining dedicated electroencephalography (EEG) and magnetic resonance imaging (MRI) approaches in adults (20-78 years), highlights the complex relationship between micro/macrostructural properties and the functional responses to visual stimuli. Here, we aimed to relate age-related changes of the latency of visual evoked potentials (VEPs) to micro/macrostructural indexes and find any correlation between micro/macrostructural features, as well. We studied age-related structural changes in the brain, by using the MRI and diffusion-weighted imaging (DWI) as preferred imaging methods for extracting brain macrostructural parameters such as the cortical thickness, surface area, folding and curvature index, gray matter volume, and microstructural parameters such as mean diffusivity (MD), radial diffusivity (RD) and axial diffusivity (AD). All the mentioned features were significantly correlated with age in V1 and V2 regions of the visual cortex. Furthermore, we highlighted, negative correlations between structural features extracted from T1-weighted images and DWI. The latency and amplitude of the three dominants peaks (C1, P1, N1) of the VEP were considered as the brain functional features to be examined for correlation with age and structural features of the corresponding age. We observed significant correlations between mean C1 latency and GM volume averaged in V1 and V2. In hierarchical regression analysis, the structural index did not contribute to significant variance in the C1 latency after regressing out the effect of age. However, the age explained significant variance in the model after regressing out the effect of structural feature.

Clinical translation of quantitative magnetic resonance imaging biomarkers - An overview and gap analysis of current practice

PURPOSE

This overview of the current landscape of quantitative magnetic resonance imaging biomarkers (qMR IBs) aims to support the standardisation of academic IBs to assist their translation to clinical practice.

METHODS

We used three complementary approaches to investigate qMR IB use and quality management practices within the UK: 1) a literature search of qMR and quality management terms during 2011-2015 and 2016-2020; 2) a database search for clinical research studies using qMR IBs during 2016-2020; and 3) a survey to ascertain the current availability and quality management practices for clinical MRI scanners and associated equipment at research institutions across the UK.

RESULTS

The analysis showed increased use of all qMR methods between the periods 2011-2015 and 2016-2020 and diffusion-tensor MRI and volumetry to be popular methods. However, the "translation ratio" of journal articles to clinical research studies was higher for qMR methods that have evidence of clinical translation via a commercial route, such as fat fraction and T2 mapping. The number of journal articles citing quality management terms doubled between the periods 2011-2015 and 2016-2020; although, its proportion relative to all journal articles only increased by 3.0%. The survey suggested that quality assurance (QA) and quality control (QC) of data acquisition procedures are under-reported in the literature and that QA/QC of acquired data/data analysis are under-developed and lack consistency between institutions.

CONCLUSIONS

We summarise current attempts to standardise and translate qMR IBs, and conclude by outlining the ideal quality management practices and providing a gap analysis between current practice and a metrological standard.

Vendor-neutral sequences and fully transparent workflows improve inter-vendor reproducibility of quantitative MRI

PURPOSE

We developed an end-to-end workflow that starts with a vendor-neutral acquisition and tested the hypothesis that vendor-neutral sequences decrease inter-vendor variability of T1, magnetization transfer ratio (MTR), and magnetization transfer saturation-index (MTsat) measurements.

METHODS

We developed and deployed a vendor-neutral 3D spoiled gradient-echo (SPGR) sequence on three clinical scanners by two MRI vendors. We then acquired T1 maps on the ISMRM-NIST system phantom, as well as T1, MTR, and MTsat maps in three healthy participants. We performed hierarchical shift function analysis in vivo to characterize the differences between scanners when the vendor-neutral sequence is used instead of commercial vendor implementations. Inter-vendor deviations were compared for statistical significance to test the hypothesis.

RESULTS

In the phantom, the vendor-neutral sequence reduced inter-vendor differences from 8% to 19.4% to 0.2% to 5% with an overall accuracy improvement, reducing ground truth T1 deviations from 7% to 11% to 0.2% to 4%. In vivo, we found that the variability between vendors is significantly reduced (p = 0.015) for all maps (T1, MTR, and MTsat) using the vendor-neutral sequence.

CONCLUSION

We conclude that vendor-neutral workflows are feasible and compatible with clinical MRI scanners. The significant reduction of inter-vendor variability using vendor-neutral sequences has important implications for qMRI research and for the reliability of multicenter clinical trials.

Bias, Repeatability and Reproducibility of Liver T1 Mapping With Variable Flip Angles

Background

Three‐dimensional variable flip angle (VFA) methods are commonly used for T₁ mapping of the liver, but there is no data on the accuracy, repeatability, and reproducibility of this technique in this organ in a multivendor setting.

Purpose

To measure bias, repeatability, and reproducibility of VFA T₁ mapping in the liver.

Study Type

Prospective observational.

Population

Eight healthy volunteers, four women, with no known liver disease.

Field Strength/Sequence

1.5‐T and 3.0‐T; three‐dimensional steady‐state spoiled gradient echo with VFAs; Look‐Locker.

Assessment

Traveling volunteers were scanned twice each (30 minutes to 3 months apart) on six MRI scanners from three vendors (GE Healthcare, Philips Medical Systems, and Siemens Healthineers) at two field strengths. The maximum period between the first and last scans among all volunteers was 9 months. Volunteers were instructed to abstain from alcohol intake for at least 72 hours prior to each scan and avoid high cholesterol foods on the day of the scan.

Statistical Tests

Repeated measures ANOVA, Student t‐test, Levene's test of variances, and 95% significance level. The percent error relative to literature liver T₁ in healthy volunteers was used to assess bias. The relative error (RE) due to intrascanner and interscanner variation in T₁ measurements was used to assess repeatability and reproducibility.

Results

The 95% confidence interval (CI) on the mean bias and mean repeatability RE of VFA T₁ in the healthy liver was 34 ± 6% and 10 ± 3%, respectively. The 95% CI on the mean reproducibility RE at 1.5 T and 3.0 T was 29 ± 7% and 25 ± 4%, respectively.

Data Conclusion

Bias, repeatability, and reproducibility of VFA T₁ mapping in the liver in a multivendor setting are similar to those reported for breast, prostate, and brain.

Level of Evidence

1

Technical Efficacy Stage

1