Purpose
Local specific absorption rate (SAR) cannot be measured and is usually evaluated by offline numerical simulations using generic body models that of course will differ from the patient's ...anatomy. An additional safety margin is needed to include this intersubject variability. In this work, we present a deep learning–based method for image‐based subject‐specific local SAR assessment. We propose to train a convolutional neural network to learn a “surrogate SAR model” to map the relation between subject‐specific B1+ maps and the corresponding local SAR.
Method
Our database of 23 subject‐specific models with an 8–transmit channel body array for prostate imaging at 7 T was used to build 5750 training samples. These synthetic complex B1+ maps and local SAR distributions were used to train a conditional generative adversarial network. Extra penalization for local SAR underestimation errors was included in the loss function. In silico and in vivo validation were performed.
Results
In silico cross‐validation shows a good qualitative and quantitative match between predicted and ground‐truth local SAR distributions. The peak local SAR estimation error distribution shows a mean overestimation error of 15% with 13% probability of underestimation. The higher accuracy of the proposed method allows the use of less conservative safety factors compared with standard procedures. In vivo validation shows that the method is applicable with realistic measurement data with impressively good qualitative and quantitative agreement to simulations.
Conclusion
The proposed deep learning method allows online image‐based subject‐specific local SAR assessment. It greatly reduces the uncertainty in current state‐of‐the‐art SAR assessment methods, reducing the time in the examination protocol by almost 25%.
Abstract Background Membrane abnormalities in polyunsaturated fatty acids (PUFAs) have been reported in schizophrenia and have been associated with brain tissue loss in normal ageing. Therefore PUFA ...may be involved in the excessive brain tissue loss reported in schizophrenia. Methods A systematic MEDLINE database search was conducted to identify studies that compared PUFAs in erythrocyte membranes in patients and controls. Patients were categorized by medication regime in medication naive first-episode patients, and patients receiving typical or atypical antipsychotics. Sample Fourteen studies were included, comprising a total of 429 patients with schizophrenia and 444 healthy control subjects. Cohen's d effect sizes were calculated for PUFAs in erythrocyte membranes using the random-effects model. Combined Cohen's d was calculated separately for patients on different medication regime. Results Medication-naive patients and patients taking typical antipsychotics showed significantly (p < 0.01) decreased concentrations of arachidonic (AA), docosahexaenoic (DHA), and docosapentaenoic (DPA) acid. In addition, patients taking typical antipsychotics showed decreased linoleic (LA), dihomo-γ-linolenic acid (DGLA), eicosapentaenoic (EPA) and docosatetraenoic (DTA) acid (p < 0.01). Patients taking atypical antipsychotics showed decreased DHA (p < 0.01) only. Conclusions PUFA concentrations in erythrocyte membranes are decreased in schizophrenia. Of particular importance in patients are lower concentrations of DHA and AA, two fatty acids that are abundant in the brain and important precursors in the cell-signalling cascade.
Objectives
Outcome of endovascular treatment in acute ischemic stroke patients depends on collateral circulation to provide blood supply to the ischemic territory. We evaluated the performance of a ...commercially available algorithm for assessing the collateral score (CS) in acute ischemic stroke patients.
Methods
Retrospectively, baseline CTA scans (≤ 3-mm slice thickness) with an intracranial carotid artery (ICA), middle cerebral artery segment M1 or M2 occlusion, from the MR CLEAN Registry (
n
= 1627) were evaluated. All CTA scans were evaluated for visual CS (0–3) by eight expert radiologists (reference standard). A Web-based AI algorithm quantified the collateral circulation (0–100%) for correctly detected occlusion sides. Agreement between visual CS and categorized automated CS (0: 0%, 1: > 0– ≤ 50%, 2: > 50– < 100%, 3: 100%) was assessed. Area under the curve (AUC) values for classifying patients in having good (CS: 2–3) versus poor (CS: 0–1) collaterals and for predicting functional independence (90-day modified Rankin Scale 0–2) were computed. Influence of CTA acquisition timing after contrast material administration was reported.
Results
In the analyzed scans (
n
= 1024), 59% agreement was found between visual CS and automated CS. An AUC of 0.87 (95% CI: 0.85–0.90) was found for discriminating good versus poor CS. Timing of CTA acquisition did not influence discriminatory performance. AUC for predicting functional independence was 0.66 (95% CI 0.62–0.69) for automated CS, similar to visual CS 0.64 (95% CI 0.61–0.68).
Conclusions
The automated CS performs similar to radiologists in determining a good versus poor collateral score and predicting functional independence in acute ischemic stroke patients with a large vessel occlusion.
Key Points
•
Software for automated quantification of intracerebral collateral circulation on computed tomography angiography performs similar to expert radiologists in determining a good versus poor collateral score.
•
Software for automated quantification of intracerebral collateral circulation on computed tomography angiography performs similar to expert radiologists in predicting functional independence in acute ischemic stroke patients with a large vessel occlusion.
•
The timing of computed tomography angiography acquisition after contrast material administration did not influence the performance of automated quantification of the collateral status.
Animal and human autopsy studies suggest that subfields of the hippocampal formation are differentially affected by neuropsychiatric diseases. Therefore, subfield volumes may be more sensitive to ...effects of disease processes. The few human studies that segmented subfields of the hippocampal formation in vivo either assessed the subfields only in the body of the hippocampus, assessed only three subfields, or did not take the differential angulation of the head of the hippocampus into account. We developed a protocol using 7 Tesla MRI with isotropic voxels to reliably delineate the entorhinal cortex (ERC), subiculum (SUB), CA1, CA2, CA3, dentate gyrus (DG)&CA4 along the full-length of the hippocampus. Fourteen subjects (aged 54-74 years, 2 men and 12 women) were scanned with a 3D turbo spin echo (TSE) sequence with isotropic voxels of 0.7 × 0.7 × 0.7 mm(3) on a 7 T MRI whole body scanner. Based on previous protocols and extensive anatomic atlases, a new protocol for segmentation of subfields of the hippocampal formation was formulated. ERC, SUB, CA1, CA2, CA3 and DG&CA4 were manually segmented twice by one rater from coronal MR images. Good-to-excellent consistency was found for all subfields (Intraclass Correlation Coefficient's (ICC) varying from 0.74 to 0.98). Accuracy as measured with the Dice Similarity Index (DSI) was above 0.82 for all subfields, with the exception of the smaller subfield CA3 (0.68-0.70). In conclusion, this study shows that it is possible to delineate the main subfields of the hippocampal formation along its full-length in vivo at 7 T MRI. Our data give evidence that this can be done in a reliable manner. Segmentation of subfields in the full-length of the hippocampus may bolster the study of the etiology neuropsychiatric diseases.
Both hemodynamics and aneurysm wall thickness are important parameters in aneurysm pathophysiology. Our aim was to develop a method for semi-quantitative wall thickness assessment on in vivo 7T MR ...images of intracranial aneurysms for studying the relation between apparent aneurysm wall thickness and wall shear stress.
Wall thickness was analyzed in 11 unruptured aneurysms in 9 patients who underwent 7T MR imaging with a TSE-based vessel wall sequence (0.8-mm isotropic resolution). A custom analysis program determined the in vivo aneurysm wall intensities, which were normalized to the signal of nearby brain tissue and were used as measures of apparent wall thickness. Spatial wall thickness variation was determined as the interquartile range in apparent wall thickness (the middle 50% of the apparent wall thickness range). Wall shear stress was determined by using phase-contrast MR imaging (0.5-mm isotropic resolution). We performed visual and statistical comparisons (Pearson correlation) to study the relation between wall thickness and wall shear stress.
3D colored apparent wall thickness maps of the aneurysms showed spatial apparent wall thickness variation, which ranged from 0.07 to 0.53, with a mean variation of 0.22 (a variation of 1.0 roughly means a wall thickness variation of 1 voxel 0.8 mm). In all aneurysms, apparent wall thickness was inversely related to wall shear stress (mean correlation coefficient, -0.35; P < .05).
A method was developed to measure the wall thickness semi-quantitatively, by using 7T MR imaging. An inverse correlation between wall shear stress and apparent wall thickness was determined. In future studies, this noninvasive method can be used to assess spatial wall thickness variation in relation to pathophysiologic processes such as aneurysm growth and rupture.
Several studies have attempted to characterize intracranial atherosclerotic plaques by using MR imaging sequences. However, dedicated validation of these sequences with histology has not yet been ...performed. The current study assessed the ability of ultra-high-resolution 7T MR imaging sequences with different image contrast weightings to image plaque components, by using histology as criterion standard.
Five specimens of the circle of Wills were imaged at 7T with 0.11 × 0.11 mm in-plane-resolution proton attenuation-, T1-, T2-, and T2*-weighted sequences (through-plane resolution, 0.11-1 mm). Tissue samples from 13 fiducial-marked locations (per specimen) on MR imaging underwent histologic processing and atherosclerotic plaque classification. Reconstructed MR images were matched with histologic sections at corresponding locations.
Forty-four samples were available for subsequent evaluation of agreement or disagreement between plaque components and image contrast differences. Of samples, 52.3% (n = 23) showed no image contrast heterogeneity; this group comprised solely no lesions or early lesions. Of samples, 25.0% (n = 11, mostly advanced lesions) showed good correlation between the spatial organization of MR imaging heterogeneities and plaque components. Areas of foamy macrophages were generally seen as proton attenuation-, T2-, and T2*- hypointense areas, while areas of increased collagen content showed more ambiguous signal intensities. Five samples showed image-contrast heterogeneity without corresponding plaque components on histology; 5 other samples showed contrast heterogeneity based on intima-media artifacts.
MR imaging at 7T has the image contrast capable of identifying both focal intracranial vessel wall thickening and distinguishing areas of different signal intensities spatially corresponding to plaque components within more advanced atherosclerotic plaques.
Purpose
Diffusion-weighted imaging (DWI) b0 may be able to substitute T2*-weighted gradient echo (GRE) or susceptibility-weighted imaging (SWI) in case of comparable detection of intracranial ...hemorrhage (ICH), thereby reducing MRI examination time. We evaluated the diagnostic accuracy of DWI b0 compared to T2*GRE or SWI for detection of ICH after reperfusion therapy for ischemic stroke.
Methods
We pooled 300 follow-up MRI scans acquired within 1 week after reperfusion therapy. Six neuroradiologists each rated DWI images (b0 and b1000; b0 as index test) of 100 patients and, after a minimum of 4 weeks, T2*GRE or SWI images (reference standard) paired with DWI images of the same patients. Readers assessed the presence of ICH (yes/no) and type of ICH according to the Heidelberg Bleeding Classification. We determined the sensitivity and specificity of DWI b0 for detection of any ICH, and the sensitivity for detection of hemorrhagic infarction (HI1 & HI2) and parenchymal hematoma (PH1 & PH2).
Results
We analyzed 277 scans of ischemic stroke patients with complete image series and sufficient image quality (median age 65 years interquartile range, 54–75, 158 57% men). For detection of any ICH on DWI b0, the sensitivity was 62% (95% CI: 50–76) and specificity 96% (95% CI: 93–99). The sensitivity of DWI b0 was 52% (95% CI: 28–68) for detection of hemorrhagic infarction and 84% (95% CI: 70–92) for parenchymal hematoma.
Conclusion
DWI b0 is inferior for detection of ICH compared to T2*GRE/SWI, especially for smaller and more subtle hemorrhages. Follow-up MRI protocols should include T2*GRE/SWI for detection of ICH after reperfusion therapy.