Rapid submillimeter QSM and R2* mapping using interleaved multishot 3D-EPI at 7 and 3 Tesla.

PURPOSE: To explore the high signal-to-noise ratio (SNR) efficiency of interleaved multishot 3D-EPI with standard image reconstruction for fast and robust high-resolution whole-brain quantitative susceptibility (QSM) and R 2 ∗ $$ {R}_2^{\ast } $$ mapping at 7 and 3T. METHODS: Single- and multi-TE segmented 3D-EPI is combined with conventional CAIPIRINHA undersampling for up to 72-fold effective gradient echo (GRE) imaging acceleration. Across multiple averages, scan parameters are varied (e.g., dual-polarity frequency-encoding) to additionally correct for B 0 $$ {\mathrm{B}}_0 $$ -induced artifacts, geometric distortions and motion retrospectively. A comparison to established GRE protocols is made. Resolutions range from 1.4 mm isotropic (1 multi-TE average in 36 s) up to 0.4 mm isotropic (2 single-TE averages in approximately 6 min) with whole-head coverage. RESULTS: Only 1-4 averages are needed for sufficient SNR with 3D-EPI, depending on resolution and field strength. Fast scanning and small voxels together with retrospective corrections result in substantially reduced image artifacts, which improves susceptibility and R 2 ∗ $$ {R}_2^{\ast } $$ mapping. Additionally, much finer details are obtained in susceptibility-weighted image projections through significantly reduced partial voluming. CONCLUSION: Using interleaved multishot 3D-EPI, single-TE and multi-TE data can readily be acquired 10 times faster than with conventional, accelerated GRE imaging. Even 0.4 mm isotropic whole-head QSM within 6 min becomes feasible at 7T. At 3T, motion-robust 0.8 mm isotropic whole-brain QSM and R 2 ∗ $$ {R}_2^{\ast } $$ mapping with no apparent distortion in less than 7 min becomes clinically feasible. Stronger gradient systems may allow for even higher effective acceleration rates through larger EPI factors while maintaining optimal contrast.

PAE planning: Radiation exposure and image quality of CT and CBCT.

PURPOSE: To determine accurate organ doses, effective doses, and image quality of computed tomography (CT) compared with cone beam CT (CBCT) for correct identification of prostatic arteries. METHOD: A dual-energy CT scanner and a flat-panel angiography system were used. Dose measurements (gallbladder (g), intestine (i), bladder (b), prostate (p), testes (t), active bone marrow of pelvis (bmp) and femura (bmf)) were performed using an anthropomorphic phantom with 65 thermoluminescent dosimeters in the pelvis and abdomen region. For the calculation of the contrast-to-noise ratio (CNR) of the pelvic arteries, a patient whose weight and height were almost identical to those of the phantom was selected for each examination type. RESULTS: The effective dose of CT was 2.7 mSv and that of CBCT was 21.8 mSv. Phantom organ doses were lower for CT than for CBCT in all organs except the testes (g: 1.2 mGy vs. 3.3 mGy, i: 5.8 mGy vs. 23.9 mGy, b: 6.9 mGy vs. 19.4 mGy, p: 6.4 mGy vs. 13.2 mGy, t: 4.7 mGy vs. 2.4 mGy, bmp: 5.1 mGy vs. 18.2 mGy, bmf: 3.3 mGy vs. 6.6 mGy). For human pelvic arteries, the CNR of CT was better than that of CBCT, with the exception of one prostate artery that showed stenosis on CT. Evaluation by experienced radiologists also confirmed the better detectability of prostate arteries on CT examination. CONCLUSIONS: In our study preprocedural CT had lower organ doses and better image quality comparedd with CBCT and should be considered for the correct identification of prostatic arteries.

Predicting disease-related MRI patterns of multiple sclerosis through GAN-based image editing.

INTRODUCTION: Multiple sclerosis (MS) is a complex neurodegenerative disorder that affects the brain and spinal cord. In this study, we applied a deep learning-based approach using the StyleGAN model to explore patterns related to MS and predict disease progression in magnetic resonance images (MRI). METHODS: We trained the StyleGAN model unsupervised using T1-weighted GRE MR images and diffusion-based ADC maps of MS patients and healthy controls. We then used the trained model to resample MR images from real input data and modified them by manipulations in the latent space to simulate MS progression. We analyzed the resulting simulation-related patterns mimicking disease progression by comparing the intensity profiles of the original and manipulated images and determined the brain parenchymal fraction (BPF). RESULTS: Our results show that MS progression can be simulated by manipulating MR images in the latent space, as evidenced by brain volume loss on both T1-weighted and ADC maps and increasing lesion extent on ADC maps. CONCLUSION: Overall, this study demonstrates the potential of the StyleGAN model in medical imaging to study image markers and to shed more light on the relationship between brain atrophy and MS progression through corresponding manipulations in the latent space.

Perfusion and T2 Relaxation Time as Predictors of Severity and Outcome in Sepsis-Associated Acute Kidney Injury: A Preclinical MRI Study.

BACKGROUND: Preventing sepsis-associated acute kidney injury (S-AKI) can be challenging because it develops rapidly and is often asymptomatic. Probability assessment of disease progression for therapeutic follow-up and outcome are important to intervene and prevent further damage. PURPOSE: To establish a noninvasive multiparametric MRI (mpMRI) tool, including T1 , T2 , and perfusion mapping, for probability assessment of the outcome of S-AKI. STUDY TYPE: Preclinical randomized prospective study. ANIMAL MODEL: One hundred and forty adult female SD rats (65 control and 75 sepsis). FIELD STRENGTH/SEQUENCE: 9.4T; T1 and perfusion map (FAIR-EPI) and T2 map (multiecho RARE). ASSESSMENT: Experiment 1: To identify renal injury in relation to sepsis severity, serum creatinine levels were determined (31 control and 35 sepsis). Experiment 2: Animals underwent mpMRI (T1 , T2 , perfusion) 18 hours postsepsis. A subgroup of animals was immediately sacrificed for histology examination (nine control and seven sepsis). Result of mpMRI in follow-up subgroup (25 control and 33 sepsis) was used to predict survival outcomes at 96 hours. STATISTICAL TESTS: Mann-Whitney U test, Spearman/Pearson correlation (r), P < 0.05 was considered statistically significant. RESULTS: Severely ill septic animals exhibited significantly increased serum creatinine levels compared to controls (70 ± 30 vs. 34 ± 9 µmol/L, P < 0.0001). Cortical perfusion (480 ± 80 vs. 330 ± 140 mL/100 g tissue/min, P < 0.005), and cortical and medullary T2 relaxation time constants were significantly reduced compared to controls (41 ± 4 vs. 37 ± 5 msec in cortex, P < 0.05, 52 ± 7 vs. 45 ± 6 msec in medulla, P < 0.05). The combination of cortical T2 relaxation time constants and perfusion results at 18 hours could predict survival outcomes at 96 hours with high sensitivity (80%) and specificity (73%) (area under curve of ROC = 0.8, Jmax = 0.52). DATA CONCLUSION: This preclinical study suggests combined T2 relaxation time and perfusion mapping as first line diagnostic tool for treatment planning. LEVEL OF EVIDENCE: 2 TECHNICAL EFFICACY STAGE: 2.

Eighteen-Month Changes in Physical Activity, Body Weight, Quadriceps Strength, and Gait Biomechanics during the COVID-19 Pandemic.

PURPOSE: This study assessed the effects of the COVID-19 pandemic restrictions/lockdowns on physical activity levels, body mass, quadriceps strength, and gait biomechanics over 18 months. METHODS: Ten healthy men were assessed at baseline (~14 wk before first lockdown) and 17.9 ± 0.3 months later (<1 wk after second lockdown). At both times, physical activity levels, body mass, and quadriceps strength were acquired using the International Physical Activity Questionnaire, a force plate, and a dynamometer, respectively. Gait data were also acquired using a motion capture system and force plates during self-paced walking, from which spatiotemporal parameters, knee angles, and external moments were computed. Baseline and follow-up measurements were compared using two-tailed paired t -tests ( α = 0.05). RESULTS: At follow-up, participants spent less time doing vigorous physical activity (∆ = -76 ± 157 min·wk -1 , P = 0.048), exhibited a tendency toward increased sedentary time (∆ = +120 ± 162 min·d -1 , P = 0.056), weighed more (∆ = +2.5 ± 2.8 kg, P = 0.021), and showed a trend toward reduced quadriceps strength (∆ = -0.29 ± 0.45 (N·m)·kg -1 , P = 0.071) compared with baseline. At follow-up, participants walked slower (∆ = -0.09 ± 0.07 m·s -1 , P = 0.005), had greater knee flexion angles at heel strike (∆ = +2.2° ± 1.8°, P = 0.004) and during late stance (∆ = +2.2° ± 1.8°, P = 0.004), had reduced knee extension moments (∆ = -0.09 ± 0.09 (N·m)·kg -1 , P = 0.012) and knee internal rotation moments (∆ = -0.02 ± 0.02 (N·m)·kg -1 , P = 0.012) during late stance. CONCLUSIONS: Healthy men exhibited reduced physical activity levels, increased body weight, a tendency toward reduced quadriceps strength, and altered gait biomechanics over the initial 18 months of the COVID-19 pandemic-alterations that could have far-reaching health consequences.

Age-related differences of cerebellar cortex and nuclei: MRI findings in healthy controls and its application to spinocerebellar ataxia (SCA6) patients.

Understanding cerebellar alterations due to healthy aging provides a reference point against which pathological findings in late-onset disease, for example spinocerebellar ataxia type 6 (SCA6), can be contrasted. In the present study, we investigated the impact of aging on the cerebellar nuclei and cerebellar cortex in 109 healthy controls (age range: 16 - 78 years) using 3 Tesla magnetic resonance imaging (MRI). Findings were compared with 25 SCA6 patients (age range: 38 - 78 years). A subset of 16 SCA6 (included: 14) patients and 50 controls (included: 45) received an additional MRI scan at 7 Tesla and were re-scanned after one year. MRI included T1-weighted, T2-weighted FLAIR, and multi-echo T2*-weighted imaging. The T2*-weighted phase images were converted to quantitative susceptibility maps (QSM). Since the cerebellar nuclei are characterized by elevated iron content with respect to their surroundings, two independent raters manually outlined them on the susceptibility maps. T1-weighted images acquired at 3T were utilized to automatically identify the cerebellar gray matter (GM) volume. Linear correlations revealed significant atrophy of the cerebellum due to tissue loss of cerebellar cortical GM in healthy controls with increasing age. Reduction of the cerebellar GM was substantially stronger in SCA6 patients. The volume of the dentate nuclei did not exhibit a significant relationship with age, at least in the age range between 18 and 78 years, whereas mean susceptibilities of the dentate nuclei increased with age. As previously shown, the dentate nuclei volumes were smaller and magnetic susceptibilities were lower in SCA6 patients compared to age- and sex-matched controls. The significant dentate volume loss in SCA6 patients could also be confirmed with 7T MRI. Linear mixed effects models and individual paired t-tests accounting for multiple comparisons revealed no statistical significant change in volume and susceptibility of the dentate nuclei after one year in neither patients nor controls. Importantly, dentate volumes were more sensitive to differentiate between SCA6 (Cohen's d = 3.02) and matched controls than the cerebellar cortex volume (d = 2.04). In addition to age-related decline of the cerebellar cortex and atrophy in SCA6 patients, age-related increase of susceptibility of the dentate nuclei was found in controls, whereas dentate volume and susceptibility was significantly decreased in SCA6 patients. Because no significant changes of any of these parameters was found at follow-up, these measures do not allow to monitor disease progression at short intervals.

Brain Macro-Structural Alterations in Aging Rats: A Longitudinal Lifetime Approach.

Aging is accompanied by macro-structural alterations in the brain that may relate to age-associated cognitive decline. Animal studies could allow us to study this relationship, but so far it remains unclear whether their structural aging patterns correspond to those in humans. Therefore, by applying magnetic resonance imaging (MRI) and deformation-based morphometry (DBM), we longitudinally screened the brains of male RccHan:WIST rats for structural changes across their average lifespan. By combining dedicated region of interest (ROI) and voxel-wise approaches, we observed an increase in their global brain volume that was superimposed by divergent local morphologic alterations, with the largest aging effects in early and middle life. We detected a modality-dependent vulnerability to shrinkage across the visual, auditory, and somato-sensory cortical areas, whereas the piriform cortex showed partial resistance. Furthermore, shrinkage emerged in the amygdala, subiculum, and flocculus as well as in frontal, parietal, and motor cortical areas. Strikingly, we noticed the preservation of ectorhinal, entorhinal, retrosplenial, and cingulate cortical regions, which all represent higher-order brain areas and extraordinarily grew with increasing age. We think that the findings of this study will further advance aging research and may contribute to the establishment of interventional approaches to preserve cognitive health in advanced age.

High-resolution CINE imaging of active guided knee motion using continuously acquired golden-angle radial MRI and rotary sensor information.

To explore and extend on dynamic imaging of joint motion, an MRI-safe device guiding knee motion with an attached rotary encoder was used in MRI measurements of multiple knee flexion-extension cycles using radial gradient echo imaging with the golden-angle as azimuthal angle increment. Reproducibility of knee motion was investigated. Real-time and CINE mode anatomical images were reconstructed for different knee flexion angles by synchronizing the encoder information with the MRI data, and performing flexion angle selective gating across multiple motion cycles. When investigating the influence of the rotation angle window width on reconstructed CINE images, it was found that angle windows between 0.5° and 3° exhibited acceptable image sharpness without suffering from significant motion-induced blurring. Furthermore, due to flexible retrospective image reconstruction afforded by the radial golden-angle imaging, the number of motion cycles included in the reconstruction could be retrospectively reduced to investigate the corresponding influence of acquisition time on image quality. Finally, motion reproducibility between motion cycles and accuracy of the flexion angle selective gating were sufficient to acquire whole-knee 3D dynamic imaging with a retrospectively gated 3D cone UTE sequence.

Magnetic susceptibility anisotropy in normal appearing white matter in multiple sclerosis from single-orientation acquisition.

Quantitative susceptibility mapping (QSM) has been successfully applied to study changes in deep grey matter nuclei as well as in lesional tissue, but its application to white matter has been complicated by the observed orientation dependence of gradient echo signal. The anisotropic susceptibility tensor is thought to be at the origin of this orientation dependence, and magnetic susceptibility anisotropy (MSA) derived from this tensor has been proposed as a marker of the state and integrity of the myelin sheath and may therefore be of particular interest for the study of demyelinating pathologies such as multiple sclerosis (MS). Reconstruction of the susceptibility tensor, however, requires repeated measurements with multiple head orientations, rendering the approach impractical for clinical applications. In this study, we combined single-orientation QSM with fibre orientation information to assess apparent MSA in three white matter tracts, i.e., optic radiation (OR), splenium of the corpus callosum (SCC), and superior longitudinal fascicle (SLF), in two cohorts of 64 healthy controls and 89 MS patients. The apparent MSA showed a significant decrease in optic radiation in the MS cohort compared with healthy controls. It decreased in the MS cohort with increasing lesion load in OR and with disease duration in the splenium. All of this suggests demyelination in normal appearing white matter. However, the apparent MSA observed in the SLF pointed to potential systematic issues that require further exploration to realize the full potential of the presented approach. Despite the limitations of such single-orientation ROI-specific estimation, we believe that our clinically feasible approach to study degenerative changes in WM is worthy of further investigation.

Characterization of microparticles of iron oxide for magnetic resonance imaging.

Microparticles of iron oxide (MPIOs) are increasingly used for contrast generation in magnetic resonance imaging (MRI). In particular, Dynabeads® MyOne™ Tosylactivated MPIOs have enabled sensitive and targeted molecular imaging, e.g., to detect vascular inflammation. For the first time we measured the relaxivities as well as the molar susceptibility χM of these MPIOs at 7 T in agarose gels. They are r1 = 0.69 ± 0.03 s-1/mM, r2 = 220 ± 6 s-1/mM, r2* = 679 ± 14 s-1/mM, and χM = 0.66 ± 0.05 ppm/mM, when expressed with respect to the iron concentration. These material parameters are essential to optimize MRI protocols and progress toward quantitative imaging. To address the heterogeneous nature of the MPIO distributions over the size of a typical MRI voxel, we coupled the MPIOs to a fluorophore to create a bimodal phantom that can be imaged by both Light Sheet microscopy and MRI. In this phantom, the MPIOs produced contrast similar to that found in vivo . The submicron resolution of Light Sheet microscopy images provided a precise measurement of the MPIO spatial distribution in phantoms also imaged by MRI. MPIO aggregates occupying less than one MRI voxel were responsible for alterations in R2* and magnetic susceptibility χ across several MRI voxels. In these cases, the sum of R2* or χ over the affected MRI volume correlated better with the microscopically determined number of MPIOs. These findings were confirmed with simulations performed in the static dephasing regime. The microscopically determined MPIO distribution was also entered directly into the simulation framework, indicating that the bimodal phantom is a useful tool to test theoretical models against experimental measurements.

In vitro measurements of radiation exposure with different modalities (computed tomography, cone beam computed tomography) for imaging the petrous bone with a pediatric anthropomorphic phantom.

BACKGROUND: Various imaging modalities, such as multi-detector computed tomography (CT) and cone beam CT are commonly used in infants for the diagnosis of hearing loss and surgical planning of implantation hearing aid devices, with differing results. OBJECTIVE: We compared three different imaging modalities available in our institution, including a high-class CT scanner, a mid-class CT scanner and an angiography system with a cone beam CT option, for image quality and radiation exposure in a phantom study. MATERIALS AND METHODS: While scanning an anthropomorphic phantom imitating a 1-year-old child with vendor-provided routine protocols, organ doses, surface doses and effective doses were determined for these three modalities with thermoluminescent dosimeters. The image quality was evaluated using the signal difference to noise ratio (SDNR) and the spatial resolution of a line-pair insert in the phantom head. The dose efficiency, defined as the ratio of SDNR and effective dose, was also compared. RESULTS: The organ and surface doses were lowest with the high-class CT protocol, but the image quality was the worst. Image quality was best with the cone beam CT protocol, which, however, had the highest radiation exposure in this study, whereas the mid-class CT was in between. CONCLUSION: Based on our results, high-end CT should be used for surgical planning because it has the lowest dose, while the image quality is still sufficient for this purpose. However, if highest image quality is needed and required, e.g., by ENT surgeons, the other modalities should be considered.

Functional connectivity and neurotransmitter impairments of the salience brain network in chronic low back pain patients: a combined resting-state functional magnetic resonance imaging and 1 H-MRS study.

ABSTRACT: Functional reorganisation of the salience network (SN) has been proposed as one of the key pathomechanisms associated with central nociceptive processing in the chronic pain state. Being associated with an altered functional connectivity within the SN, these processes have been hypothesized to result from a loss of inhibitory function leading to node hyperexcitability and spontaneous pain. Combined resting-state BOLD functional magnetic resonance imaging (MRI) and 1 H-MR spectroscopy was applied to chronic back pain patients and healthy subjects to assess deviations from functional integrity (weighted closeness centrality [wCC], derived from resting-state functional MRI), oscillatory BOLD characteristics (spectral power), and neurotransmitter levels (GABA + , glutamate+glutamine) in 2 key SN nodes, anterior insular (aIns R ) and anterior mid-cingulate cortices. In addition, examinations were repeated in chronic back pain patients after a 4-week interdisciplinary multimodal pain treatment and in healthy subjects after 4 weeks to explore longitudinal, treatment-mediated changes in target variables. The aIns R and, to a lesser extent, the anterior mid-cingulate of patients exhibited significantly reduced wCC accompanied by a spectral power shift from a lower to a higher frequency band, indicating a desynchronization of their neuronal activity within the SN, possibly because of increased spontaneous activations. Without revealing neurotransmitter differences, patients alone showed significant positive associations between local GABA + levels and wCC in aIns R , suggesting a stronger dependence of node synchronization on the inhibitory tone in the chronic pain state. However, this needs to be explored in the future using magnetic resonance spectroscopy techniques that are more sensitive to detecting subtle neurotransmitter changes and also allow multifocal characterization of neurotransmitter tone.

CAN THE SIZE-SPECIFIC DOSE ESTIMATE BE DERIVED FROM THE BODY MASS INDEX? A FEASIBILITY STUDY.

Size-specific dose estimate ($\mathbf{SSDE}$) index appears to be more suitable than the commonly used volume computed tomography dose index ($\mathbf{C}{\mathbf{TDI}}_{\mathbf{vol}}$) to estimate the dose delivered to the patient during a computed tomography (CT) scan. We evaluated whether an ${\mathbf{SSDE}}_{\mathbf{BMI}}$ can be determined from the patient's body mass index ($\mathbf{BMI}$) with sufficient reliability in the case that a $\mathbf{SSDE}$ is not given by the CT scanner. For each of the three most used examination types, CT examinations of 50 female and 50 male patients were analyzed. The $\mathbf{SSDE}$ values automatically provided by the scanner were compared with ${\mathbf{SSDE}}_{\mathbf{BMI}}$ determined from $\mathbf{C}{\mathbf{TDI}}_{\mathbf{vol}}$ and $\mathbf{BMI}$. A good accordance of ${\mathbf{SSDE}}_{\mathbf{BMI}}$ and $\mathbf{SSDE}$ was found for the chest and abdominal regions. A low correlation was observed for the head region. The presented method is a simple and practically useful surrogate approach for the chest and abdominal regions but not for the head.

Quantitative susceptibility mapping reveals alterations of dentate nuclei in common types of degenerative cerebellar ataxias.

The cerebellar nuclei are a brain region with high iron content. Surprisingly, little is known about iron content in the cerebellar nuclei and its possible contribution to pathology in cerebellar ataxias, with the only exception of Friedreich's ataxia. In the present exploratory cross-sectional study, quantitative susceptibility mapping was used to investigate volume, iron concentration and total iron content of the dentate nuclei in common types of hereditary and non-hereditary degenerative ataxias. Seventy-nine patients with spinocerebellar ataxias of types 1, 2, 3 and 6; 15 patients with Friedreich's ataxia; 18 patients with multiple system atrophy, cerebellar type and 111 healthy controls were also included. All underwent 3 T MRI and clinical assessments. For each specific ataxia subtype, voxel-based and volumes-of-interest-based group analyses were performed in comparison with a corresponding age- and sex-matched control group, both for volume, magnetic susceptiblity (indicating iron concentration) and susceptibility mass (indicating total iron content) of the dentate nuclei. Spinocerebellar ataxia of type 1 and multiple system atrophy, cerebellar type patients showed higher susceptibilities in large parts of the dentate nucleus but unaltered susceptibility masses compared with controls. Friedreich's ataxia patients and, only on a trend level, spinocerebellar ataxia of type 2 patients showed higher susceptibilities in more circumscribed parts of the dentate. In contrast, spinocerebellar ataxia of type 6 patients revealed lower susceptibilities and susceptibility masses compared with controls throughout the dentate nucleus. Spinocerebellar ataxia of type 3 patients showed no significant changes in susceptibility and susceptibility mass. Lower volume of the dentate nuclei was found to varying degrees in all ataxia types. It was most pronounced in spinocerebellar ataxia of type 6 patients and least prominent in spinocerebellar ataxia of type 3 patients. The findings show that alterations in susceptibility revealed by quantitative susceptibility mapping are common in the dentate nuclei in different types of cerebellar ataxias. The most striking changes in susceptibility were found in spinocerebellar ataxia of type 1, multiple system atrophy, cerebellar type and spinocerebellar ataxia of type 6. Because iron content is known to be high in glial cells but not in neurons of the cerebellar nuclei, the higher susceptibility in spinocerebellar ataxia of type 1 and multiple system atrophy, cerebellar type may be explained by a reduction of neurons (increase in iron concentration) and/or an increase in iron-rich glial cells, e.g. microgliosis. Hypomyelination also leads to higher susceptibility and could also contribute. The lower susceptibility in SCA6 suggests a loss of iron-rich glial cells. Quantitative susceptibility maps warrant future studies of iron content and iron-rich cells in ataxias to gain a more comprehensive understanding of the pathogenesis of these diseases.

The differential association between local neurotransmitter levels and whole-brain resting-state functional connectivity in two distinct cingulate cortex subregions.

We examined the association between rsFC and local neurotransmitter levels in the pregenual anterior cingulate cortex (pgACC) and the anterior mid-cingulate cortex (aMCC) by varying rsFC-strengths at the whole-brain level. Our results showed region-dependent directionality of associations in the investigated ACC subdivisions.

A novel multipurpose device for guided knee motion and loading during dynamic magnetic resonance imaging.

INTRODUCTION: This work aimed to develop a novel multipurpose device for guided knee flexion-extension, both passively using a motorized pneumatic system and actively (muscle-driven) with the joint unloaded or loaded during dynamic MRI. Secondary objectives were to characterize the participant experience during device use, and present preliminary dynamic MRI data to demonstrate the different device capabilities. MATERIAL AND METHODS: Self-reported outcomes were used to characterize the pain, physical exertion and discomfort levels experienced by 10 healthy male participants during four different active knee motion and loading protocols using the novel device. Knee angular data were recorded during the protocols to determine the maximum knee range of motion achievable. Dynamic MRI was acquired for three healthy volunteers during passive, unloaded knee motion using 2D Cartesian TSE, 2D radial GRE and 3D UTE sequences; and during active, unloaded and loaded knee motion using 2D radial GRE imaging. Because of the different MRI sequences used, spatial resolution was inherently lower for active knee motion than for passive motion acquisitions. RESULTS: Depending on the protocol, some participants reported slight pain, mild discomfort and varying levels of physical exertion. On average, participants achieved ∼40° of knee flexion; loaded conditions can create knee moments up to 27Nm. High quality imaging data were obtained during different motion and loading conditions. Dynamic 3D data allowed to retrospectively extract arbitrarily oriented slices. CONCLUSION: A novel multipurpose device for guided, physiologically relevant knee motion and loading during dynamic MRI was developed. Device use was well tolerated and suitable for acquiring high quality images during different motion and loading conditions. Different bone positions between loaded and unloaded conditions were likely due to out-of-plane motion, particularly because image registration was not performed. Ultimately, this device could be used to advance our understanding of physiological and pathological joint mechanics.

Investigation of biases in convolutional neural networks for semantic segmentation using performance sensitivity analysis.

The application of deep neural networks for segmentation in medical imaging has gained substantial interest in recent years. In many cases, this variant of machine learning has been shown to outperform other conventional segmentation approaches. However, little is known about its general applicability. Especially the robustness against image modifications (e.g., intensity variations, contrast variations, spatial alignment) has hardly been investigated. Data augmentation is often used to compensate for sensitivity to such changes, although its effectiveness has not yet been studied. Therefore, the goal of this study was to systematically investigate the sensitivity to variations in input data with respect to segmentation of medical images using deep learning. This approach was tested with two publicly available segmentation frameworks (DeepMedic and TractSeg). In the case of DeepMedic, the performance was tested using ground truth data, while in the case of TractSeg, the STAPLE technique was employed. In both cases, sensitivity analysis revealed significant dependence of the segmentation performance on input variations. The effects of different data augmentation strategies were also shown, making this type of analysis a useful tool for selecting the right parameters for augmentation. The proposed analysis should be applied to any deep learning image segmentation approach, unless the assessment of sensitivity to input variations can be directly derived from the network.