Background/Objectives: To evaluate the influence of different DIXON contrasts on the quality of subtraction images in dynamic breast MRI using maximum intensity projections (MIPs). Methods: This retrospective study included n = 40 women (median age: 53.5 years, range 23–83) undergoing clinically indicated breast MRI (3T). For each MRI examination, two independent readers individually evaluated GBCA-enhanced subtraction MIPS for different timepoints (n = 5) and DIXON contrasts (n = 4) per breast, resulting in a total of 800 individual evaluations. Evaluations comprised (a) qualitative measures, using Likert-scores for artefact strength, breast parenchyma visibility, lesion visibility and reading confidence; and (b) signal intensity, measured in three regions of interest with the apparent signal-to-noise ratio (aSNR) and apparent contrast-to-noise ratio (aCNR) calculated. The evaluation results were analysed to identify differences between DIXON contrasts. Results: The “only water” DIXON contrast at ~120s after GBCA injection achieved the highest lesion conspicuity and reading confidence scores and lowest artefact scores; however, its performance was not statistically significant (p > 0.05) compared to the “in-phase” and “opposed-phase” subtractions. The aCNR at the second timepoint was slightly, but not significantly (p > 0.05), lower than the first timepoint, whilst aSNR increased significantly from the first to second timepoint in all contrasts. Conclusions: Subtraction MIPs derived from the “only water” DIXON contrast achieved the highest qualitative scoring for lesion conspicuity and confidence, with the aSNR increasing and aCNR decreasing between the first and second timepoints.

A bstract Background Brain magnetic resonance elastography (MRE) is an emerging quantitative neuroimaging technique that provides noninvasive maps of brain tissue viscoelasticity. For multi-center applications, robust cross-site reproducibility across scanner platforms is essential but remains insufficiently characterized. Purpose To evaluate cross-site reproducibility of brain multifrequency MRE measurements between two MRI scanner platforms using harmonized protocols. Study Type Prospective cross-site test-retest reproducibility study. Study Population Sixteen healthy adult volunteers (7 men, 9 women; mean age 32.2 ± 8.0 years). Field Strength/Sequence 3 T systems (Siemens MAGNETOM Cima.X and MAGNETOM Vida at two sites) with identical brain multifrequency MRE sequences, echo-planar imaging (EPI) readout, and standardized driver configuration. Assessment Each participant underwent one MRE acquisition at each site. Shear wave speed (SWS) and penetration rate (PR) were quantified in whole brain, white matter, subcortical gray matter, and cortical gray matter regions using atlas-based region-of-interest (ROI) analysis in MNI152 space. Statistical Tests Absolute relative difference (ARD), reproducibility coefficient (RDC), coefficient of variation (CV), intraclass correlation coefficient (ICC), and Bland-Altman plots were calculated to determine cross-site reproducibility. Results Cross-site reproducibility was robust for major brain regions, with region-averaged ARD values for SWS ranging from 1.38 % to 3.43 % and for PR from 3.20 % to 7.25 % across tissues. RDCs for SWS ranged from 0.02 m.s -1 to 0.07 m.s -1 , and for PR from 0.03 m.s -1 to 0.08 m.s -1 . Coefficients of variation for SWS ranged from 0.82 % to 1.93 %, and for PR from 2.21 % to 4.09 %. ICC values for SWS ranged from 0.66 to 0.84 and for PR from 0.67 to 0.88. Bland-Altman analysis showed minimal systematic bias and tight limits of agreement. Conclusion Brain multifrequency MRE demonstrates robust reproducibility across distinct 3 T platforms when using harmonized acquisition and reconstruction. These results support the use of brain MRE as a quantitative biomarker and provide benchmark reproducibility metrics for future research.

PURPOSE:To visualize and characterize the five most common kidney stone types based on their magnetic susceptibilities in MRI using QSM. METHODS:Three water-based agar phantoms were constructed, containing a total of 53 ex vivo kidney stones of varying types and sizes. Seven different MRI multi-echo gradient echo sequences were employed at field strengths of 1.5 T and 3 T. Each acquisition was repeated three times. Susceptibility maps were reconstructed and mean susceptibility values calculated for the individual kidney stones (normalized to the agar gel). Five of the most common kidney stone types-carbonate apatite, calcium oxalate, uric acid, cystine, and struvite-were investigated with respect to different sequence parameters and field strengths. RESULTS:All examined kidney stones were reliably visualized as diamagnetic using QSM. Overall, good repeatability was observed for the mean susceptibility values and standard deviations of the individual kidney stones. Mean susceptibility values were highly consistent across acquisitions at both field strengths for kidney stones belonging to the five major types with diameters larger than 3.0 mm (66% of all stones). The overall mean susceptibility values (± SD) were - 1.25 ± 0.16 ppm for carbonate apatite, -1.08 ± 0.13 ppm for calcium oxalate, -0.94 ± 0.17 ppm for uric acid, -1.32 ± 0.16 ppm for cystine, and -1.28 ± 0.28 ppm for struvite. CONCLUSION:Our data indicate that QSM is well-suited to reliably visualize kidney stones and distinguish certain types among the five most common kidney stone compositions using MRI. This represents a crucial step toward in vivo imaging of suspected urolithiasis in clinical practice.

PURPOSE:To achieve a reliable fat/water separation (FWS) at 7 T using a new 3D radial sequence and reconstruction workflow with quasi-continuous TE sampling. METHODS:A 3D radial density-adapted sequence with quasi-continuous TE sampling was developed for a 7 T whole-body MRI system. The reconstruction workflow included an off-resonance correction to reduce chemical shift artifacts. Fat and water signals were separated using a graph cut algorithm, yielding proton density fat fraction maps. The sequence and reconstruction pipeline were tested and validated using phantom measurements and in vivo acquisitions of the lower leg of healthy volunteers. RESULTS:The proposed sequence and reconstruction pipeline enabled sampling of the fat/water oscillation curve over a range from a minimal mean TE of 0.27 ms to a maximum mean TE of 10.13 ms, with an effective TE increment of 85 μs. The off-resonant correction significantly reduced chemical shift artifacts in the fat signal. The data points of the fat/water oscillation showed good agreement with a multi-peak fit function. The quantification of phantom fat percentages was verified with a mean absolute error of 1.5%. The proposed pipeline resulted in consistent interpretation of fat and water signals without any swaps, neither in phantom nor in in vivo measurements. CONCLUSION:This study successfully demonstrated the application of a 3D radial sequence with quasi-continuous TEs for FWS at 7 T. Due to the high sampling rate (effective TE increment below 100 μs), the presented FWS workflow demonstrated high robustness against fat/water swaps, which are typical at ultra-high field strengths.

PURPOSE:High-resolution diffusion-weighted imaging (DWI) is clinically demanding. The purpose of this work is to develop an efficient self-supervised algorithm unrolling technique for submillimeter-resolution DWI. METHODS:We developed submillimeter DWI acquisition utilizing multi-band multi-shot EPI with diffusion shift encoding. We unrolled the alternating direction method of multipliers (ADMM) to perform scan-specific self-gated self-supervised DeepDWI learning for multi-shot echo planar imaging with diffusion shift encoding on a clinical 7 T scanner. RESULTS:We demonstrate that (1) ADMM unrolling is generalizable across slices, (2) ADMM unrolling outperforms multiplexed sensitivity-encoding (MUSE) and compressed sensing with locally-low rank (LLR) regularization in terms of image sharpness, tissue continuity, and motion robustness, and (3) ADMM unrolling enables clinically feasible inference time. CONCLUSION:Our proposed ADMM unrolling enables whole brain DWI of 21 diffusion volumes at 0.7 mm isotropic resolution and 10 min scan, and shows higher signal-to-noise ratio (SNR), clearer tissue delineation, and improved motion robustness, which makes it plausible for clinical translation.

OBJECTIVES:The use of prostate magnetic resonance imaging (MRI) is increasing, and coverage often captures substantial portions of the pelvis, visualizing findings outside the prostate gland. The objective of this study was to evaluate the feasibility of automated detection and segmentation of high-prevalence incidental findings in prostate MRI. MATERIALS AND METHODS:This IRB-approved, retrospective study included n=465 prostate MRI examinations (1.5 and 3.0 T), comprising n=315 internal cases from our institution and n=150 external cases from 3 independent data sets. Manual segmentations were performed for perirectal lymph nodes, sigmoid diverticulosis, urinary bladder diverticula, bladder wall thickenings, inguinal hernias, periarticular bone changes of the hip, prominent synovial compartments of the hip, and hydroceles testis on axial T2-weighted (T2w) images for n=265 internal cases (n=520 ROIs). An nnU-Net model was trained on n=213 of these cases. The remaining n=52 independent cases (quantitative test set) were used for the quantitative evaluation of model performance using Dice score, intersection over union (IoU), Hausdorff distance (HD), mean surface distance (MSD), confusion matrices, sensitivity, specificity, and accuracy. Furthermore, n=200 additional examinations (reader test set), comprising n=50 internal, n=150 from 3 independent external data sets, were evaluated by 2 radiologists in an AI-assisted evaluation. Readers assessed the presence of incidental findings (step 1) and the correctness of the nnU-Net-predicted findings (step 2), based on the AI-predicted segmentations. RESULTS:Segmentation performance varied between the incidental findings. Evaluation of the independent quantitative test set revealed that the highest mean Dice scores were achieved for sigmoid diverticulosis (0.80±0.14), hydroceles testis (0.76±0.20), and periarticular bone changes of the hip (0.70±0.07). Radiologists' evaluation of AI predictions on an independent reader test set comprising 1 internal and 3 external data sets demonstrated high agreement in AI-assisted evaluation for most incidental findings. Accuracies per data set (PROSTATEx/Ai4ar/internal/Prostate-3T) were 0.94/0.62/0.96/0.94 for perirectal lymph nodes, 0.82/0.80/0.80/0.84 for sigmoid diverticulosis, 0.86/0.86/0.98/0.98 for urinary bladder diverticula, 0.96/0.82/0.88/0.72 for bladder wall thickenings, 0.94/0.82/0.78/0.86 for inguinal hernias, 0.90/0.80/0.82/0.84 for periarticular bone changes of the hip, 0.90/0.82/0.84/0.96 for prominent synovial compartments of the hip, 0.98/0.80/0.96/0.96 for hydroceles testis. Inter-reader agreement for the AI-assisted evaluation of the presence of incidental findings on T2w images was high, with Cohen κ values ranging from 0.74 to 0.92 for most findings. CONCLUSIONS:The nnU-Net-based AI model was able to capture and segment frequent incidental findings in prostate MRI across 4 independent data sets, demonstrating potential to support radiologists in consistent reporting. This supports further research with larger, more diverse data sets, including additional annotations and clinical targets.

Scaling is an important step for achieving accurate participant-specific models when studying human motion. Scaling a generic musculoskeletal model using OpenSim is time consuming, depends on the user expertise, and requires a static pose. Recently, AddBiomechanics introduced automatic scaling of the participant-specific models independent of user expertise and static pose. However, its validation is limited to synthetic data. In this exploratory study, we compared models scaled via AddBiomechanics and OpenSim against models scaled based on magnetic resonance images (MRI). We performed an optical motion capture experiment in which we recorded walking at 0.8 m/s, 1.2 m/s, and 1.6 m/s, followed by an MRI scan of the lower extremities for 10 participants (M/F). We scaled the model using OpenSim, AddBiomechanics, and using the MRI data. In OpenSim, it was necessary for some participants to lock certain joint coordinates before scaling to avoid unrealistic postures. To evaluate the different models, we performed inverse kinematics for the three different walking trials of each participant and analyzed the joint angle trajectories and overall average root mean square error (RMSE) between the measured markers and virtual markers of the models scaled with OpenSim, AddBiomechanics and MRI. For those participants where we did not lock coordinates in the OpenSim model, the average marker RMSE was 1.657 cm for the OpenSim model, compared to 1.585 cm with AddBiomechanics and 1.471 cm for the MRI-based model. For the participants where we locked coordinates, the RMSE was 1.588 cm for the OpenSim model, compared to 1.725 cm with AddBiomechanics and 1.439 cm for the MRI-based model. The joint angles were similar, with the largest difference for the models with locked coordinates, where the maximum difference was 9.2° (ankle angle). Our exploratory study suggests that AddBiomechanics offers a practical alternative to OpenSim, showing comparable accuracy with no meaningful differences, while requiring less time and user effort.

BackgroundSelecting appropriate sample sizes in magnetic resonance imaging studies is a complex process that requires to balance statistical rigor with the practical challenges of measuring a large patient population. In this Institutional Review Board approved study, we evaluate the dominant error types ("finite N" errors versus precision errors) for apparent diffusion coefficient (ADC)-based lesion characterization in diffusion-weighted magnetic resonance imaging (DWI) of the female breast in a local dataset and compare our results with current literature.MethodsFirst, in a literature review including 24 published breast DWI studies, the standard error of the area under the receiver operating characteristic curve as a measure of sample size-related errors (finite N errors) was estimated for the reported ADC values and compared to the values, derived from expert readings of a university hospital's cohort of 171 patients with suspicious breast lesions. Second, precision errors were assessed based on published analyses of the coefficient of variation of ADC values, measured in breast DWI exams.ResultsFinite N errors were dominant in the in-house study and most of the 24 reviewed studies. The median sample size at which finite N errors and precision errors were equal was determined to be n = 932.DiscussionThis analysis of dominant error types shows that the required sample sizes for the considered use case are not unreasonably large and that reducing sample sizes may not be justified based on the merits of the conducted analysis. Nonetheless, incorporating dominant error type assessments into future studies may provide valuable insights for optimizing study design and improving methodological rigor.

Q-space trajectory imaging (QTI) provides promising markers of tissue microstructure, but clinical translation requires shorter acquisitions, faster analysis, and more robust parameter estimation at high spatial resolution. To address these barriers, we trained a voxel-wise multilayer perceptron (MLP) to infer QTI-derived scalar parameters directly from the diffusion signal. We established reference QTI parameters of the brain in 18 healthy subjects using constrained fitting on 50-min QTI scans. The MLP was trained to estimate those reference parameters from a five-minute subset of the diffusion data. We compared the MLP with the constrained fit applied to the same short-protocol input, computing normalized root mean squared error, peak signal-to-noise ratio, and structural similarity with respect to the reference. Here, the MLP consistently achieved better performance metrics, with normalized root mean squared errors up to two-fold lower. For one whole-brain dataset, MLP inference reduced computation time from more than an hour with constrained fitting to a few seconds. Robustness to lower SNR was tested in a separate 1.7 mm isotropic voxel size acquisition of the full protocol, in which the MLP retained lower errors and less visually apparent noise. Finally, we demonstrate qualitative feasibility in two glioma patients scanned with the short protocol. We conclude that a simple MLP can provide high-quality QTI parameter estimates from short tensor-valued diffusion acquisitions. This enables five-minute, high-resolution QTI and may encourage further clinical studies of markers such as microscopic fractional anisotropy.

Background/Objectives: Artificial intelligence (AI) can support lesion detection in gadolinium-based contrast agent-enhanced (GBCA-enhanced) breast MRI. However, its effectiveness on virtual contrast-enhanced (vCE) images remains unclear. This feasibility study evaluated the publicly available MAMA-MIA nnU-Net model trained on GBCA-enhanced data using an independent cohort of both GBCA-enhanced and vCE breast MRI. Methods: This IRB-approved retrospective study included the publicly available nnU-Net model trained on n = 1506 MAMA-MIA breast MRI scans and a cohort of n = 2126 in-house 3T breast MRI scans. A generative adversarial network (Pix2Pix-GAN) was developed on n = 1870 of the in-house scans and used to generate vCE data on the remaining independent n = 256 in-house cases. The MAMA-MIA nnU-net was applied to both GBCA-enhanced (GBCA) and corresponding vCE images. Ground-truth segmentations of malignant lesions served to calculate the Dice score, Hausdorff distance, and lesion dimension differences. Results: The final test set comprised n = 250 cases (n = 69 malignant, n = 181 benign). Lesion detection rates were 91% (n = 63/n = 69; 95% confidence interval (CI): 82.3–96.0%) for GBCA and 84% (n = 58/n = 69; 95% CI: 73.7–90.9%) for vCE. Two lesions missed in GBCA were identified by vCE. The Hausdorff distances were similar (GBCA: 6.4 (IQR: 3.2–9.3; 95% CI: 5.2–7.8) mm; vCE: 6.7 (IQR: 3.9–9.7; 95% CI: 5.3–8.0) mm, p = 0.564). The Dice scores showed minor differences (GBCA: 0.829 (IQR: 0.723–0.900; 95% CI: 0.786–0.865) vs. vCE: 0.826 (IQR: 0.720–0.857; 95% CI: 0.770–0.836); p < 0.001). vCE images had slightly higher non-target tissue segmentation (median 6072 mm3 vs. 5754 mm3). Conclusions: A GBCA-trained algorithm demonstrated some cross-domain transferability to vCE images, albeit with a reduced case-level sensitivity of 84% (95% CI: 73.7–90.9%) vs. 91% (95% CI: 82.3–96.0%). Based on these preliminary results, further research, including larger cohorts and more diverse datasets, is warranted.

Intravoxel incoherent motion (IVIM) MRI allows for simultaneous assessment of tissue microcirculation (perfusion) and diffusion of water. In single-center studies, IVIM has shown great potential for diagnosis, treatment outcome prediction, and treatment monitoring for many different diseases and organs. However, heterogeneity in data acquisition protocols, pre-processing pipelines, and post-processing routines yields differences in reported IVIM parameters, which has constrained large-scale deployment of IVIM. Moreover, deploying IVIM protocols and analysis typically requires technical expertise, further challenging wider use, especially for clinicians. In this consensus paper, to accelerate the deployment of IVIM, we provide recommendations and harmonize protocols for brain, breast, kidney, liver, muscle, and pancreas IVIM studies. For this goal we organized multiple questionnaires and held a dedicated workshop. To ensure a level of standardized, reproducible results, without restricting innovation, we suggest a small subset of b-values to always be measured and analyzed separately, and to which more extensive b-value sampling can be added for advanced investigations. We further introduce detailed recommendations on acquisition protocols and analysis pipelines. To increase consistency, repeatability, and reproducibility, we highly recommend that these protocols and pipelines be deployed by scientists and clinicians for IVIM studies. For advanced users who desire different protocols or analysis approaches, we suggest adding results from our suggested protocols and analysis pipeline in the supplemental part of their paper to enable retrospective studies.

Purpose: To complement 1.5-minute measurements of common tensor-valued diffusion MRI (dMRI) markers with rapid constrained fitting. Methods: Fast dMRI protocols for obtaining rotational invariants of the cumulant expansion (RICE) were paired with constrained weighted linear least squares (CWLLS) to stabilize the more fragile WLLS fit. A compact constraint set was formulated, including a novel mean-dependent upper bound on total diffusional variance. Evaluation used diffusion tensor distribution (DTD) simulations, healthy-volunteer data with a resolution-dependent SNR experiment, and a glioma patient dataset. A 5-minute q-space trajectory imaging (QTI) protocol served as a reference. Results: Across experiments, CWLLS reduced unphysical estimates and fit outliers in parameters such as microscopic FA and isotropic diffusivity variance. In simulations, it narrowed error distributions most clearly in the CSF-dominant case, while some metrics showed a bias-variance trade-off. In vivo, CWLLS removed negative variance estimates, truncated out-of-bounds tails, and reduced artifacts in fluid-contaminated voxels while preserving anatomical contrast. It also retained more stable maps than WLLS at higher resolution, although both estimators degraded in the lowest-SNR setting. Notably, the new mean-dependent variance bound was violated in 15.4 Conclusion: CWLLS for fast RICE yielded high-quality parameter maps at an online-ready computational cost. This may enhance the reliability of dMRI tissue characterization and strengthen the path toward clinical translation.

ABSTRACT Background Myofibrillar myopathies (MFM) form a large group of clinically and genetically heterogeneous protein aggregate diseases. We investigated whether a novel quantitative MRI protocol can reveal new aspects of structural and biochemical muscle pathology in three classic MFM subtypes. Methods MRI of the lower legs was performed in nine MFM patients with filamin‐C (FLNC; n = 5), desmin (DES, n = 2) and LIM domain binding 3 (LDB3; n = 2) gene mutations, one patient with non‐MFM, filamin‐C related distal myopathy (4 males, 6 females, 51.0 ± 7.7 years) and 10 age‐matched healthy control subjects (5 males, 5 females, 50.0 ± 11.0 years). 1H MRI at 3 T addressed fatty replacement and edema‐like changes as well as quantitative measurements of proton density fat fraction (PDFF) and water T2 relaxation times. 39K/23Na MRI at 7 T was employed to determine apparent tissue potassium and tissue sodium concentrations (aTPC/aTSC). Results T1‐weighted and T2‐weighted STIR imaging showed the highest degree of fat replacement in the soleus and gastrocnemius medialis muscle regions and the highest degree of edema‐like changes in the extensor regions in all 10 myopathy patients. The lowest degree of fat replacement and edema‐like changes was present in the gastrocnemius lateralis muscles. Marked fatty replacement of peroneus muscles was also present in DES‐related MFM and FLNC‐related distal myopathy. Muscular PDFF values were significantly increased in all MFM patients (p = 0.003 ‐ < 0.001) with 60 and 35 of 63 muscles analysed showing increased mean PDFF (> 10% and > 50%). When excluding the muscles with PDFF > 50%, the median water T2 was significantly increased in all muscle regions of MFM patients with the exception of the tibialis anterior and posterior muscles. Fat‐corrected aTSC values in MFM patients were significantly increased compared to healthy controls (55.6 ± 16.3 mM vs. 23.2 ± 5.5 mM, p < 0.001) in all muscles but peroneus muscles, whereas fat‐corrected aTPC values were reduced in all muscles except for gastrocnemius lateralis, tibialis posterior and peroneus muscles (75.4 ± 13.3 mM vs. 108.9 ± 9.9 mM, p < 0.001). Conclusions Quantitative PDFF measurements and water T2 mapping serve as valuable tools to objectively quantify fat and edema‐like changes in MFM. Furthermore, changes in potassium/sodium ion balance in the lower leg muscles of MFM patients could serve as new markers to quantify the extent of biochemical changes in individual muscle regions. Further longitudinal evaluation is required to validate whether they are sensitive to changes prior to a high degree of fat replacement.

Diffusion-weighted imaging (DWI) can support lesion detection and characterization in breast magnetic resonance imaging (MRI), however especially high b-value diffusion-weighted acquisitions can be prone to intensity artifacts that can affect diagnostic image assessment. This study aims to detect both hyper- and hypointense artifacts on high b-value diffusion-weighted images (b=1500 s/mm2) using deep learning, employing either a binary classification (artifact presence) or a multiclass classification (artifact intensity) approach on a slice-wise dataset.This IRB-approved retrospective study used the single-center dataset comprising n=11806 slices from routine 3T breast MRI examinations performed between 2022 and mid-2023. Three convolutional neural network (CNN) architectures (DenseNet121, ResNet18, and SEResNet50) were trained for binary classification of hyper- and hypointense artifacts. The best performing model (DenseNet121) was applied to an independent holdout test set and was further trained separately for multiclass classification. Evaluation included area under receiver operating characteristic curve (AUROC), area under precision recall curve (AUPRC), precision, and recall, as well as analysis of predicted bounding box positions, derived from the network Grad-CAM heatmaps. DenseNet121 achieved AUROCs of 0.92 and 0.94 for hyper- and hypointense artifact detection, respectively, and weighted AUROCs of 0.85 and 0.88 for multiclass classification on single-slice high b-value diffusion-weighted images. A radiologist evaluated bounding box precision on a 1-5 Likert-like scale across 200 slices, achieving mean scores of 3.33+-1.04 for hyperintense artifacts and 2.62+-0.81 for hypointense artifacts. Hyper- and hypointense artifact detection in slice-wise breast DWI MRI dataset (b=1500 s/mm2) using CNNs particularly DenseNet121, seems promising and requires further validation.

To evaluate a DINOv2-based medical slice transformer (MST) for triaging abbreviated breast MRI by ruling out examinations with suspicious findings that are immediately actionable (Breast Imaging Reporting and Data System [BI-RADS] ≥ 4) across contrast-enhanced and non–contrast-enhanced protocols. This institutional review board–approved retrospective study included 1847 single-breast MRI examinations (377 BI-RADS ≥ 4) from an in-house dataset and 924 from an external dataset (Duke). Four abbreviated protocols were tested: T1-weighted early subtraction (T1sub), diffusion-weighted imaging with b = 1500 s/mm² (DWI1500), DWI1500 + T2-weighted (T2w), and T1sub + T2w. Performance was assessed at 90

Maximum intensity projections (MIPs) facilitate rapid lesion detection both for contrast-enhanced (CE) and diffusion-weighted imaging (DWI) breast magnetic resonance imaging (MRI). We evaluated the feasibility of AI-based virtual CE subtraction MIPs as a reading approach. This Institutional Review Board-approved retrospective study includes 540 multi-parametric breast MRI examinations (performed from 2017 to 2020), including multi-b-value DWI (50, 750, and 1,500 s/mm²). A 2D U-Net was trained using unenhanced (UnE) images as inputs to generate virtual abbreviated CE (VAbCE) subtractions. Two radiologists evaluated lesion suspicion, image quality, and artifacts for UnE, VACE, and abbreviated CE (AbCE) images. Lesion conspicuity was compared between VAbCE and AbCE MIPs. Cancer detection rates for UE, VAbCE, and AbCE MIPs were 90.0

Multiparametric MRI with gadolinium-based contrast agents demonstrates high sensitivity for detecting lesions in the breast, particularly in women with denser breast tissue. However, its use is limited by increased costs, time and contraindications in certain patients. This study explores a pix2pix generative adversarial network (GAN) to create virtual contrast-enhanced (vCE) MRI from un-enhanced T1w, T2w, and DWI sequences and compares it with a U-Net model. The vCE GAN achieved an SSIM of 80.75 and PSNR of 21.90, while the vCE U-Net scored 87.39 and 24.39, respectively. A multi-reader Turing test showed that 45.89

After acute lesions in the central nervous system (CNS), the interaction of microglia, astrocytes, and infiltrating immune cells decides over their resolution or chronification. However, this CNS-intrinsic cross-talk is poorly characterized. Analyzing cerebrospinal fluid (CSF) samples of Multiple Sclerosis (MS) patients as well as CNS samples of female mice with experimental autoimmune encephalomyelitis (EAE), the animal model of MS, we identify microglia-derived TGFα as key factor driving recovery. Through mechanistic in vitro studies, in vivo treatment paradigms, scRNA sequencing, CRISPR-Cas9 genetic perturbation models and MRI in the EAE model, we show that together with other glial and non-glial cells, microglia secrete TGFα in a highly regulated temporospatial manner in EAE. Here, TGFα contributes to recovery by decreasing infiltrating T cells, pro-inflammatory myeloid cells, oligodendrocyte loss, demyelination, axonal damage and neuron loss even at late disease stages. In a therapeutic approach in EAE, blood-brain barrier penetrating intranasal application of TGFα attenuates pro-inflammatory signaling in astrocytes and CNS infiltrating immune cells while promoting neuronal survival and lesion resolution. Together, microglia-derived TGFα is an important mediator of glial-immune crosstalk, highlighting its therapeutic potential in resolving acute CNS inflammation.

Breast magnetic resonance imaging (MRI) protocols often include T2-weighted fat-saturated (T2w-FS) sequences, which support tissue characterization but significantly increase scan time. This study aims to evaluate whether a 2D-U-Net neural network can generate virtual T2w-FS (VirtuT2w) images from routine multiparametric breast MRI images. This IRB-approved, retrospective study included 914 breast MRI examinations from January 2017 to June 2020. The dataset was divided into training (n = 665), validation (n = 74), and test sets (n = 175). The U-Net was trained using different input protocols consisting of T1-weighted, diffusion-weighted, and dynamic contrast-enhanced sequences to generate VirtuT2. Quantitative metrics were used to evaluate the different input protocols. A qualitative assessment by two radiologists was used to evaluate the VirtuT2w images of the best input protocol. VirtuT2w images demonstrated the best quantitative metrics compared to original T2w-FS images for an input protocol using all of the available data. A high level of high-frequency error norm (0.87) indicated a strong blurring presence in the VirtuT2 images, which was also confirmed by qualitative reading. Radiologists correctly identified VirtuT2 images with at least 96

Background: Magnetic resonance imaging (MRI) has high sensitivity for breast cancer detection, but interpretation is time-consuming. Artificial intelligence may aid in pre-screening. Purpose: To evaluate the DINOv2-based Medical Slice Transformer (MST) for ruling out significant findings (Breast Imaging Reporting and Data System [BI-RADS] >=4) in contrast-enhanced and non-contrast-enhanced abbreviated breast MRI. Materials and Methods: This institutional review board approved retrospective study included 1,847 single-breast MRI examinations (377 BI-RADS >=4) from an in-house dataset and 924 from an external validation dataset (Duke). Four abbreviated protocols were tested: T1-weighted early subtraction (T1sub), diffusion-weighted imaging with b=1500 s/mm2 (DWI1500), DWI1500+T2-weighted (T2w), and T1sub+T2w. Performance was assessed at 90