Optical Flow-Guided Cine MRI Segmentation With Learned Corrections.

In cardiac cine magnetic resonance imaging (MRI), the heart is repeatedly imaged at numerous time points during the cardiac cycle. Frequently, the temporal evolution of a certain region of interest such as the ventricles or the atria is highly relevant for clinical diagnosis. In this paper, we devise a novel approach that allows for an automatized propagation of an arbitrary region of interest (ROI) along the cardiac cycle from respective annotated ROIs provided by medical experts at two different points in time, most frequently at the end-systolic (ES) and the end-diastolic (ED) cardiac phases. At its core, a 3D TV- L1 -based optical flow algorithm computes the apparent motion of consecutive MRI images in forward and backward directions. Subsequently, the given terminal annotated masks are propagated by this bidirectional optical flow in 3D, which results, however, in improper initial estimates of the segmentation masks due to numerical inaccuracies. These initially propagated segmentation masks are then refined by a 3D U-Net-based convolutional neural network (CNN), which was trained to enforce consistency with the forward and backward warped masks using a novel loss function. Moreover, a penalization term in the loss function controls large deviations from the initial segmentation masks. This method is benchmarked both on a new dataset with annotated single ventricles containing patients with severe heart diseases and on a publicly available dataset with different annotated ROIs. We emphasize that our novel loss function enables fine-tuning the CNN on a single patient, thereby yielding state-of-the-art results along the complete cardiac cycle.

Development of image-based decision support systems utilizing information extracted from radiological free-text report databases with text-based transformers.

OBJECTIVES: To investigate the potential and limitations of utilizing transformer-based report annotation for on-site development of image-based diagnostic decision support systems (DDSS). METHODS: The study included 88,353 chest X-rays from 19,581 intensive care unit (ICU) patients. To label the presence of six typical findings in 17,041 images, the corresponding free-text reports of the attending radiologists were assessed by medical research assistants ("gold labels"). Automatically generated "silver" labels were extracted for all reports by transformer models trained on gold labels. To investigate the benefit of such silver labels, the image-based models were trained using three approaches: with gold labels only (MG), with silver labels first, then with gold labels (MS/G), and with silver and gold labels together (MS+G). To investigate the influence of invested annotation effort, the experiments were repeated with different numbers (N) of gold-annotated reports for training the transformer and image-based models and tested on 2099 gold-annotated images. Significant differences in macro-averaged area under the receiver operating characteristic curve (AUC) were assessed by non-overlapping 95% confidence intervals. RESULTS: Utilizing transformer-based silver labels showed significantly higher macro-averaged AUC than training solely with gold labels (N = 1000: MG 67.8 [66.0-69.6], MS/G 77.9 [76.2-79.6]; N = 14,580: MG 74.5 [72.8-76.2], MS/G 80.9 [79.4-82.4]). Training with silver and gold labels together was beneficial using only 500 gold labels (MS+G 76.4 [74.7-78.0], MS/G 75.3 [73.5-77.0]). CONCLUSIONS: Transformer-based annotation has potential for unlocking free-text report databases for the development of image-based DDSS. However, on-site development of image-based DDSS could benefit from more sophisticated annotation pipelines including further information than a single radiological report. CLINICAL RELEVANCE STATEMENT: Leveraging clinical databases for on-site development of artificial intelligence (AI)-based diagnostic decision support systems by text-based transformers could promote the application of AI in clinical practice by circumventing highly regulated data exchanges with third parties. KEY POINTS: • The amount of data from a database that can be used to develop AI-assisted diagnostic decision systems is often limited by the need for time-consuming identification of pathologies by radiologists. • The transformer-based structuring of free-text radiological reports shows potential to unlock corresponding image databases for on-site development of image-based diagnostic decision support systems. • However, the quality of image annotations generated solely on the content of a single radiology report may be limited by potential inaccuracies and incompleteness of this report.

Direct deep learning-based survival prediction from pre-interventional CT prior to transcatheter aortic valve replacement.

PURPOSE: To investigate survival prediction in patients undergoing transcatheter aortic valve replacement (TAVR) using deep learning (DL) methods applied directly to pre-interventional CT images and to compare performance with survival models based on scalar markers of body composition. METHOD: This retrospective single-center study included 760 patients undergoing TAVR (mean age 81 ± 6 years; 389 female). As a baseline, a Cox proportional hazards model (CPHM) was trained to predict survival on sex, age, and the CT body composition markers fatty muscle fraction (FMF), skeletal muscle radiodensity (SMRD), and skeletal muscle area (SMA) derived from paraspinal muscle segmentation of a single slice at L3/L4 level. The convolutional neural network (CNN) encoder of the DL model for survival prediction was pre-trained in an autoencoder setting with and without a focus on paraspinal muscles. Finally, a combination of DL and CPHM was evaluated. Performance was assessed by C-index and area under the receiver operating curve (AUC) for 1-year and 2-year survival. All methods were trained with five-fold cross-validation and were evaluated on 152 hold-out test cases. RESULTS: The CNN for direct image-based survival prediction, pre-trained in a focussed autoencoder scenario, outperformed the baseline CPHM (CPHM: C-index = 0.608, 1Y-AUC = 0.606, 2Y-AUC = 0.594 vs. DL: C-index = 0.645, 1Y-AUC = 0.687, 2Y-AUC = 0.692). Combining DL and CPHM led to further improvement (C-index = 0.668, 1Y-AUC = 0.713, 2Y-AUC = 0.696). CONCLUSIONS: Direct DL-based survival prediction shows potential to improve image feature extraction compared to segmentation-based scalar markers of body composition for risk assessment in TAVR patients.

Reduction of Gadolinium-Based Contrast Agents in MRI Using Convolutional Neural Networks and Different Input Protocols: Limited Interchangeability of Synthesized Sequences With Original Full-Dose Images Despite Excellent Quantitative Performance.

OBJECTIVES: The purpose of this study was to implement a state-of-the-art convolutional neural network used to synthesize artificial T1-weighted (T1w) full-dose images from corresponding noncontrast and low-dose images (using various settings of input sequences) and test its performance on a patient population acquired prospectively. MATERIALS AND METHODS: In this monocentric, institutional review board-approved study, a total of 138 participants were included who received an adapted imaging protocol with acquisition of a T1w low dose after administration of 10% of the standard dose and acquisition of a T1w full dose after administration of the remaining 90% of the standard dose of a gadolinium-containing contrast agent. A total of 83 participants formed the training sample (51.7 ± 16.5 years, 36 women), 25 the validation sample (55.3 ± 16.4 years, 11 women), and 30 the test sample (55.0 ± 15.0 years, 9 women). Four input settings were differentiated: only the T1w noncontrast and T1w low-dose images (standard setting), only the T1w noncontrast and T1w low-dose images with a prolonged postinjection time of 5 minutes (5-minute setting), multiple noncontrast sequences (T1w, T2w, diffusion) and the T1w low-dose images (extended setting), and only noncontrast sequences (T1w, T2w, diffusion) were used (zero-dose setting). For each setting, a deep neural network was trained to synthesize artificial T1w full-dose images, which were assessed on the test sample using an objective evaluation based on quantitative metrics and a subjective evaluation through a reader-based study. Three readers scored the overall image quality, the interchangeability in regard to the clinical conclusion compared with the true T1w full-dose sequence, the contrast enhancement of lesions, and their conformity to the respective references in the true T1w full dose. RESULTS: Quantitative analysis of the artificial T1w full-dose images of the standard setting provided a peak signal-to-noise ratio of 33.39 ± 0.62 (corresponding to an average improvement of the low-dose sequences of 5.2 dB) and a structural similarity index measure of 0.938 ± 0.005. In the 4-fold cross-validation, the extended setting yielded similar performance to the standard setting in terms of peak signal-to-noise ratio ( P = 0.20), but a slight improvement in structural similarity index measure ( P < 0.0001). For all settings, the reader study found comparable overall image quality between the original and artificial T1w full-dose images. The proportion of scans scored as fully or mostly interchangeable was 55%, 58%, 43%, and 3% and the average counts of false positives per case were 0.42 ± 0.83, 0.34 ± 0.71, 0.82 ± 1.15, and 2.00 ± 1.07 for the standard, 5-minute, extended, and zero-dose setting, respectively. Using a 5-point Likert scale (0 to 4, 0 being the worst), all settings of synthesized full-dose images showed significantly poorer contrast enhancement of lesions compared with the original full-dose sequence (difference of average degree of contrast enhancement-standard: -0.97 ± 0.83, P = <0.001; 5-minute: -0.93 ± 0.91, P = <0.001; extended: -0.96 ± 0.97, P = <0.001; zero-dose: -2.39 ± 1.14, P = <0.001). The average scores of conformity of the lesions compared with the original full-dose sequence were 2.25 ± 1.21, 2.22 ± 1.27, 2.24 ± 1.25, and 0.73 ± 0.93 for the standard, 5-minute, extended, and zero-dose setting, respectively. CONCLUSIONS: The tested deep learning algorithm for synthesis of artificial T1w full-dose sequences based on images after administration of only 10% of the standard dose of a gadolinium-based contrast agent showed very good quantitative performance. Despite good image quality in all settings, both false-negative and false-positive signals resulted in significantly limited interchangeability of the synthesized sequences with the original full-dose sequences.

Variability and magnitude of brain glutamate levels in schizophrenia: a meta and mega-analysis.

Glutamatergic dysfunction is implicated in schizophrenia pathoaetiology, but this may vary in extent between patients. It is unclear whether inter-individual variability in glutamate is greater in schizophrenia than the general population. We conducted meta-analyses to assess (1) variability of glutamate measures in patients relative to controls (log coefficient of variation ratio: CVR); (2) standardised mean differences (SMD) using Hedges g; (3) modal distribution of individual-level glutamate data (Hartigan's unimodality dip test). MEDLINE and EMBASE databases were searched from inception to September 2022 for proton magnetic resonance spectroscopy (1H-MRS) studies reporting glutamate, glutamine or Glx in schizophrenia. 123 studies reporting on 8256 patients and 7532 controls were included. Compared with controls, patients demonstrated greater variability in glutamatergic metabolites in the medial frontal cortex (MFC, glutamate: CVR = 0.15, p < 0.001; glutamine: CVR = 0.15, p = 0.003; Glx: CVR = 0.11, p = 0.002), dorsolateral prefrontal cortex (glutamine: CVR = 0.14, p = 0.05; Glx: CVR = 0.25, p < 0.001) and thalamus (glutamate: CVR = 0.16, p = 0.008; Glx: CVR = 0.19, p = 0.008). Studies in younger, more symptomatic patients were associated with greater variability in the basal ganglia (BG glutamate with age: z = -0.03, p = 0.003, symptoms: z = 0.007, p = 0.02) and temporal lobe (glutamate with age: z = -0.03, p = 0.02), while studies with older, more symptomatic patients associated with greater variability in MFC (glutamate with age: z = 0.01, p = 0.02, glutamine with symptoms: z = 0.01, p = 0.02). For individual patient data, most studies showed a unimodal distribution of glutamatergic metabolites. Meta-analysis of mean differences found lower MFC glutamate (g = -0.15, p = 0.03), higher thalamic glutamine (g = 0.53, p < 0.001) and higher BG Glx in patients relative to controls (g = 0.28, p < 0.001). Proportion of males was negatively associated with MFC glutamate (z = -0.02, p < 0.001) and frontal white matter Glx (z = -0.03, p = 0.02) in patients relative to controls. Patient PANSS total score was positively associated with glutamate SMD in BG (z = 0.01, p = 0.01) and temporal lobe (z = 0.05, p = 0.008). Further research into the mechanisms underlying greater glutamatergic metabolite variability in schizophrenia and their clinical consequences may inform the identification of patient subgroups for future treatment strategies.

Deep learning enables automated MRI-based estimation of uterine volume also in patients with uterine fibroids undergoing high-intensity focused ultrasound therapy.

BACKGROUND: High-intensity focused ultrasound (HIFU) is used for the treatment of symptomatic leiomyomas. We aim to automate uterine volumetry for tracking changes after therapy with a 3D deep learning approach. METHODS: A 3D nnU-Net model in the default setting and in a modified version including convolutional block attention modules (CBAMs) was developed on 3D T2-weighted MRI scans. Uterine segmentation was performed in 44 patients with routine pelvic MRI (standard group) and 56 patients with uterine fibroids undergoing ultrasound-guided HIFU therapy (HIFU group). Here, preHIFU scans (n = 56), postHIFU imaging maximum one day after HIFU (n = 54), and the last available follow-up examination (n = 53, days after HIFU: 420 ± 377) were included. The training was performed on 80% of the data with fivefold cross-validation. The remaining data were used as a hold-out test set. Ground truth was generated by a board-certified radiologist and a radiology resident. For the assessment of inter-reader agreement, all preHIFU examinations were segmented independently by both. RESULTS: High segmentation performance was already observed for the default 3D nnU-Net (mean Dice score = 0.95 ± 0.05) on the validation sets. Since the CBAM nnU-Net showed no significant benefit, the less complex default model was applied to the hold-out test set, which resulted in accurate uterus segmentation (Dice scores: standard group 0.92 ± 0.07; HIFU group 0.96 ± 0.02), which was comparable to the agreement between the two readers. CONCLUSIONS: This study presents a method for automatic uterus segmentation which allows a fast and consistent assessment of uterine volume. Therefore, this method could be used in the clinical setting for objective assessment of therapeutic response to HIFU therapy.

Simplified intravoxel incoherent motion DWI for differentiating malignant from benign breast lesions.

BACKGROUND: To evaluate simplified intravoxel incoherent motion (IVIM) diffusion-weighted imaging (DWI) for differentiating malignant versus benign breast lesions as (i) stand-alone tool and (ii) add-on to dynamic contrast-enhanced magnetic resonance imaging. METHODS: 1.5-T DWI data (b = 0, 50, 250, 800 s/mm2) were retrospectively analysed for 126 patients with malignant or benign breast lesions. Apparent diffusion coefficient (ADC) ADC (0, 800) and IVIM-based parameters D1' = ADC (50, 800), D2' = ADC (250, 800), f1' = f (0, 50, 800), f2' = f (0, 250, 800) and D*' = D* (0, 50, 250, 800) were voxel-wise calculated without fitting procedures. Regions of interest were analysed in vital tumour and perfusion hot spots. Beside the single parameters, the combined use of D1' with f1' and D2' with f2' was evaluated. Lesion differentiation was investigated for lesions (i) with hyperintensity on DWI with b = 800 s/mm2 (n = 191) and (ii) with suspicious contrast-enhancement (n = 135). RESULTS: All lesions with suspicious contrast-enhancement appeared also hyperintense on DWI with b = 800 s/mm2. For task (i), best discrimination was reached for the combination of D1' and f1' using perfusion hot spot regions-of-interest (accuracy 93.7%), which was higher than that of ADC (86.9%, p = 0.003) and single IVIM parameters D1' (88.0%) and f1' (87.4%). For task (ii), best discrimination was reached for single parameter D1' using perfusion hot spot regions-of-interest (92.6%), which were slightly but not significantly better than that of ADC (91.1%) and D2' (88.1%). Adding f1' to D1' did not improve discrimination. CONCLUSIONS: IVIM analysis yielded a higher accuracy than ADC. If stand-alone DWI is used, perfusion analysis is of special relevance.

Deep learning supports the differentiation of alcoholic and other-than-alcoholic cirrhosis based on MRI.

Although CT and MRI are standard procedures in cirrhosis diagnosis, differentiation of etiology based on imaging is not established. This proof-of-concept study explores the potential of deep learning (DL) to support imaging-based differentiation of the etiology of liver cirrhosis. This retrospective, monocentric study included 465 patients with confirmed diagnosis of (a) alcoholic (n = 221) and (b) other-than-alcoholic (n = 244) cirrhosis. Standard T2-weighted single-slice images at the caudate lobe level were randomly split for training with fivefold cross-validation (85%) and testing (15%), balanced for (a) and (b). After automated upstream liver segmentation, two different ImageNet pre-trained convolutional neural network (CNN) architectures (ResNet50, DenseNet121) were evaluated for classification of alcohol-related versus non-alcohol-related cirrhosis. The highest classification performance on test data was observed for ResNet50 with unfrozen pre-trained parameters, yielding an area under the receiver operating characteristic curve of 0.82 (95% confidence interval (CI) 0.71-0.91) and an accuracy of 0.75 (95% CI 0.64-0.85). An ensemble of both models did not lead to significant improvement in classification performance. This proof-of-principle study shows that deep-learning classifiers have the potential to aid in discriminating liver cirrhosis etiology based on standard MRI.

End-to-end automated body composition analyses with integrated quality control for opportunistic assessment of sarcopenia in CT.

OBJECTIVES: To develop a pipeline for automated body composition analysis and skeletal muscle assessment with integrated quality control for large-scale application in opportunistic imaging. METHODS: First, a convolutional neural network for extraction of a single slice at the L3/L4 lumbar level was developed on CT scans of 240 patients applying the nnU-Net framework. Second, a 2D competitive dense fully convolutional U-Net for segmentation of visceral and subcutaneous adipose tissue (VAT, SAT), skeletal muscle (SM), and subsequent determination of fatty muscle fraction (FMF) was developed on single CT slices of 1143 patients. For both steps, automated quality control was integrated by a logistic regression model classifying the presence of L3/L4 and a linear regression model predicting the segmentation quality in terms of Dice score. To evaluate the performance of the entire pipeline end-to-end, body composition metrics, and FMF were compared to manual analyses including 364 patients from two centers. RESULTS: Excellent results were observed for slice extraction (z-deviation = 2.46 ± 6.20 mm) and segmentation (Dice score for SM = 0.95 ± 0.04, VAT = 0.98 ± 0.02, SAT = 0.97 ± 0.04) on the dual-center test set excluding cases with artifacts due to metallic implants. No data were excluded for end-to-end performance analyses. With a restrictive setting of the integrated segmentation quality control, 39 of 364 patients were excluded containing 8 cases with metallic implants. This setting ensured a high agreement between manual and fully automated analyses with mean relative area deviations of ΔSM = 3.3 ± 4.1%, ΔVAT = 3.0 ± 4.7%, ΔSAT = 2.7 ± 4.3%, and ΔFMF = 4.3 ± 4.4%. CONCLUSIONS: This study presents an end-to-end automated deep learning pipeline for large-scale opportunistic assessment of body composition metrics and sarcopenia biomarkers in clinical routine. KEY POINTS: • Body composition metrics and skeletal muscle quality can be opportunistically determined from routine abdominal CT scans. • A pipeline consisting of two convolutional neural networks allows an end-to-end automated analysis. • Machine-learning-based quality control ensures high agreement between manual and automatic analysis.

Comparison of different ROI analysis methods for liver lesion characterization with simplified intravoxel incoherent motion (IVIM).

This study investigated the impact of different ROI placement and analysis methods on the diagnostic performance of simplified IVIM-DWI for differentiating liver lesions. 1.5/3.0-T DWI data from a respiratory-gated MRI sequence (b = 0, 50, 250, 800 s/mm2) were analyzed in patients with malignant (n = 74/54) and benign (n = 35/19) lesions. Apparent diffusion coefficient ADC = ADC(0,800) and IVIM parameters D1' = ADC(50,800), D2' = ADC(250,800), f1' = f(0,50,800), f2' = f(0,250,800), and D*' = D*(0,50,250,800) were calculated voxel-wise. For each lesion, a representative 2D-ROI, a 3D-ROI whole lesion, and a 3D-ROI from "good" slices were placed, including and excluding centrally deviating areas (CDA) if present, and analyzed with various histogram metrics. The diagnostic performance of 2D- and 3D-ROIs was not significantly different; e.g. AUC (ADC/D1'/f1') were 0.958/0.902/0.622 for 2D- and 0.942/0.892/0.712 for whole lesion 3D-ROIs excluding CDA at 1.5 T (p > 0.05). For 2D- and 3D-ROIs, AUC (ADC/D1'/D2') were significantly higher, when CDA were excluded. With CDA included, AUC (ADC/D1'/D2'/f1'/D*') improved when low percentiles were used instead of averages, and was then comparable to the results of average ROI analysis excluding CDA. For lesion differentiation the use of a representative 2D-ROI is sufficient. CDA should be excluded from ROIs by hand or automatically using low percentiles of diffusion coefficients.

Simplified intravoxel incoherent motion diffusion-weighted MRI of liver lesions: feasibility of combined two-colour index maps.

BACKGROUND: To evaluate the feasibility of two-colour index maps containing combined diffusion and perfusion information from simplified intravoxel incoherent motion (IVIM) for liver lesion malignancy assessment. METHODS: Diffusion-weighted data from a respiratory-gated 1.5-T magnetic resonance sequence were analysed in 109 patients with liver lesions. With three b values (0, 50, 800 s/mm2) estimated diffusion coefficient D', perfusion fraction f', and apparent diffusion coefficient (ADC) maps were calculated and analysed for regions of interest (ROIs). D' and f' cutoff values were determined by differentiating haemangiomas from other lesions and focal nodular hyperplasias from other lesions, respectively. Combined IDf index maps were generated with a voxel value set to 100, if both D' and f' voxel values were lower than their cutoff values (1,529.4 × 10-6 mm2/s and 114.4 × 10-3, respectively), otherwise to 0. Moreover, IADC index maps were generated from ADC cutoff value (1,338.5 × 10-6 mm2/s) obtained by differentiating benign from malignant lesions. Discriminatory power was assessed for both IDf and IADC. Index maps were displayed as two-colour overlays to b-800 images and visually assessed within the translucent hyperintense areas. RESULTS: For IDf, the same diagnostic accuracy was achieved as for the combined use of parameters D' and f' (93.6%). Compared to IADC, IDf showed a higher diagnostic accuracy. Visual judgment of IDf yielded an accuracy (95.4%) similar to that of quantitative analysis (93.6%). CONCLUSION: Voxel-wise combined two-colour index maps IDf provide similar diagnostic accuracy as ROI-based combination of estimated IVIM parameters D' and f' and are suitable for visual assessment of liver lesion malignancy.

Detection of liver cirrhosis in standard T2-weighted MRI using deep transfer learning.

OBJECTIVES: To investigate the diagnostic performance of deep transfer learning (DTL) to detect liver cirrhosis from clinical MRI. METHODS: The dataset for this retrospective analysis consisted of 713 (343 female) patients who underwent liver MRI between 2017 and 2019. In total, 553 of these subjects had a confirmed diagnosis of liver cirrhosis, while the remainder had no history of liver disease. T2-weighted MRI slices at the level of the caudate lobe were manually exported for DTL analysis. Data were randomly split into training, validation, and test sets (70%/15%/15%). A ResNet50 convolutional neural network (CNN) pre-trained on the ImageNet archive was used for cirrhosis detection with and without upstream liver segmentation. Classification performance for detection of liver cirrhosis was compared to two radiologists with different levels of experience (4th-year resident, board-certified radiologist). Segmentation was performed using a U-Net architecture built on a pre-trained ResNet34 encoder. Differences in classification accuracy were assessed by the χ2-test. RESULTS: Dice coefficients for automatic segmentation were above 0.98 for both validation and test data. The classification accuracy of liver cirrhosis on validation (vACC) and test (tACC) data for the DTL pipeline with upstream liver segmentation (vACC = 0.99, tACC = 0.96) was significantly higher compared to the resident (vACC = 0.88, p < 0.01; tACC = 0.91, p = 0.01) and to the board-certified radiologist (vACC = 0.96, p < 0.01; tACC = 0.90, p < 0.01). CONCLUSION: This proof-of-principle study demonstrates the potential of DTL for detecting cirrhosis based on standard T2-weighted MRI. The presented method for image-based diagnosis of liver cirrhosis demonstrated expert-level classification accuracy. KEY POINTS: • A pipeline consisting of two convolutional neural networks (CNNs) pre-trained on an extensive natural image database (ImageNet archive) enables detection of liver cirrhosis on standard T2-weighted MRI. • High classification accuracy can be achieved even without altering the pre-trained parameters of the convolutional neural networks. • Other abdominal structures apart from the liver were relevant for detection when the network was trained on unsegmented images.

Feasibility of CT-derived myocardial strain measurement in patients with advanced cardiac valve disease.

To explore the feasibility of CT-derived myocardial strain measurement in patients with advanced cardiac valve disease and to compare it to strain measurements derived from transthoracic echocardiography (TTE). 43 consecutive patients with advanced cardiac valve disease and clinically indicated retrospectively gated cardiac CTs were retrospectively analyzed. The longitudinal, circumferential as well as radial systolic strain were determined in all patients utilizing a commercially available CT strain software. In 36/43 (84%) patients, CT-derived longitudinal strain was compared to speckle-tracking TTE. Pearson's correlation coefficients as well as Bland-Altman analysis were used to compare the CT-derived strain measurements to TTE. The intra- and inter-reader-reliability of the CT-derived strain measurements were assessed by intra-class correlation coefficients (ICCs). Strain measurements were feasible in all patients. CT-derived global longitudinal strain (GLS) correlated moderately with TTE-derived GLS (r = 0.6, p < 0.001). A moderate correlation between CT-derived GLS and CT-derived left ventricular ejection fraction was found (LVEF, r = - 0.66, p = 0.036). Bland-Altman analysis showed a systematic underestimation of myocardial strain by cardiac CT compared to TTE (mean difference: - 5.8%, 95% limit of agreement between - 13.3 and 1.8%). Strain measurements showed an excellent intra- and inter-reader-reliability with an intra-reader ICC of 1.0 and an inter-reader ICC of 0.99 for GLS measurements. CT-derived myocardial strain measurements are feasible in patients with advanced cardiac valve disease. They are highly reproducible and correlate with established parameters of strain measurements. Our results encourage the implementation of CT-derived strain measurement into clinical routine.

Association of Age, Antipsychotic Medication, and Symptom Severity in Schizophrenia With Proton Magnetic Resonance Spectroscopy Brain Glutamate Level: A Mega-analysis of Individual Participant-Level Data.

Importance: Proton magnetic resonance spectroscopy (1H-MRS) studies indicate that altered brain glutamatergic function may be associated with the pathophysiology of schizophrenia and the response to antipsychotic treatment. However, the association of altered glutamatergic function with clinical and demographic factors is unclear. Objective: To assess the associations of age, symptom severity, level of functioning, and antipsychotic treatment with brain glutamatergic metabolites. Data Sources: The MEDLINE database was searched to identify journal articles published between January 1, 1980, and June 3, 2020, using the following search terms: MRS or magnetic resonance spectroscopy and (1) schizophrenia or (2) psychosis or (3) UHR or (4) ARMS or (5) ultra-high risk or (6) clinical high risk or (7) genetic high risk or (8) prodrome* or (9) schizoaffective. Authors of 114 1H-MRS studies measuring glutamate (Glu) levels in patients with schizophrenia were contacted between January 2014 and June 2020 and asked to provide individual participant data. Study Selection: In total, 45 1H-MRS studies contributed data. Data Extraction and Synthesis: Associations of Glu, Glu plus glutamine (Glx), or total creatine plus phosphocreatine levels with age, antipsychotic medication dose, symptom severity, and functioning were assessed using linear mixed models, with study as a random factor. Main Outcomes and Measures: Glu, Glx, and Cr values in the medial frontal cortex (MFC) and medial temporal lobe (MTL). Results: In total, 42 studies were included, with data for 1251 patients with schizophrenia (mean [SD] age, 30.3 [10.4] years) and 1197 healthy volunteers (mean [SD] age, 27.5 [8.8] years). The MFC Glu (F1,1211.9 = 4.311, P = .04) and Glx (F1,1079.2 = 5.287, P = .02) levels were lower in patients than in healthy volunteers, and although creatine levels appeared lower in patients, the difference was not significant (F1,1395.9 = 3.622, P = .06). In both patients and volunteers, the MFC Glu level was negatively associated with age (Glu to Cr ratio, F1,1522.4 = 47.533, P < .001; cerebrospinal fluid-corrected Glu, F1,1216.7 = 5.610, P = .02), showing a 0.2-unit reduction per decade. In patients, antipsychotic dose (in chlorpromazine equivalents) was negatively associated with MFC Glu (estimate, 0.10 reduction per 100 mg; SE, 0.03) and MFC Glx (estimate, -0.11; SE, 0.04) levels. The MFC Glu to Cr ratio was positively associated with total symptom severity (estimate, 0.01 per 10 points; SE, 0.005) and positive symptom severity (estimate, 0.04; SE, 0.02) and was negatively associated with level of global functioning (estimate, 0.04; SE, 0.01). In the MTL, the Glx to Cr ratio was positively associated with total symptom severity (estimate, 0.06; SE, 0.03), negative symptoms (estimate, 0.2; SE, 0.07), and worse Clinical Global Impression score (estimate, 0.2 per point; SE, 0.06). The MFC creatine level increased with age (estimate, 0.2; SE, 0.05) but was not associated with either symptom severity or antipsychotic medication dose. Conclusions and Relevance: Findings from this mega-analysis suggest that lower brain Glu levels in patients with schizophrenia may be associated with antipsychotic medication exposure rather than with greater age-related decline. Higher brain Glu levels may act as a biomarker of illness severity in schizophrenia.

A pilot study of magnetic resonance fingerprinting in Parkinson's disease.

Parkinson's disease (PD) affects more than six million people, but reliable MRI biomarkers with which to diagnose patients have not been established. Magnetic resonance fingerprinting (MRF) is a recent quantitative technique that can provide relaxometric maps from a single sequence. The purpose of this study is to assess the potential of MRF to identify PD in patients and their disease severity, as well as to evaluate comfort during MRF. Twenty-five PD patients and 25 matching controls underwent 3 T MRI, including an axial 2D spoiled gradient echo MRF sequence. T1 and T2 maps were generated by voxel-wise matching the measured MRF signal to a precomputed dictionary. All participants also received standard inversion recovery T1 and multi-echo T2 mapping. An ROI-based analysis of relaxation times was performed. Differences between patients and controls as well as techniques were determined by logistic regression, Spearman correlation and t-test. Patients were asked to estimate the subjective comfort of the MRF sequence. Both MRF-based T1 and T2 mapping discriminated patients from controls: T1 relaxation times differed most in cortical grey matter (PD 1337 ± 38 vs. control 1386 ± 37 ms; mean ± SD; P = .0001) and, in combination with normal-appearing white matter, enabled correct discrimination in 85.7% of cases (sensitivity 83.3%; specificity 88.0%; receiver-operating characteristic [ROC]) area under the curve [AUC] 0.87), while for T2 mapping the left putamen was the strongest classifier (40.54 ± 6.28 vs. 34.17 ± 4.96 ms; P = .0001), enabling differentiation of groups in 84.0% of all cases (sensitivity 80.0%; specificity 88.0%; ROC AUC 0.87). Relaxation time differences were not associated with disease severity. Standard mapping techniques generated significantly different relaxation time values and identified other structures as different between groups other than MRF. Twenty-three out of 25 PD patients preferred the MRF examination instead of a standard MRI. MRF-based mapping can identify PD patients with good comfort but needs further assessment regarding disease severity identification and its potential for comparability with standard mapping technique results.

Body composition analysis using CT and MRI: intra-individual intermodal comparison of muscle mass and myosteatosis.

Computed tomography (CT) and magnetic resonance imaging (MRI) can quantify muscle mass and quality. However, it is still unclear if CT and MRI derived measurements can be used interchangeable. In this prospective study, fifty consecutive participants of a cancer screening program underwent same day low-dose chest CT and MRI. Cross-sectional areas (CSA) of the paraspinal skeletal muscles were obtained. CT and MRI muscle fat infiltration (MFI) were assessed by mean radiodensity in Hounsfield units (HU) and proton density fat fraction (MRIPDFF), respectively. CSA and MFI were highly correlated between CT and MRI (CSA: r = 0.93, P < 0.001; MFI: r = - 0.90, P < 0.001). Mean CSA was higher in CT compared to MRI (46.6cm2 versus 43.0cm2; P = 0.05) without significance. Based on MRIPDFF, a linear regression model was established to directly estimate skeletal muscle fat content from CT. Bland-Altman plots showed a difference between measurements of - 0.5 cm2 to 7.6 cm2 and - 4.2% to 2.4% regarding measurements of CSA and MFI, respectively. In conclusion, the provided results indicate interchangeability of CT and MRI derived imaging biomarkers of skeletal muscle quantity and quality. Comparable to MRIPDFF, skeletal muscle fat content can be quantified from CT, which might have an impact of analyses in larger cohort studies, particularly in sarcopenia patients.

Fully Automated Segmentation of Connective Tissue Compartments for CT-Based Body Composition Analysis: A Deep Learning Approach.

OBJECTIVE: Body composition comprises prognostic information in patients with various malignancies and can be opportunistically determined from routine computed tomography (CT) scans. However, accurate assessment of patients with alterations, for example, due to ascites or anasarca, and accurate identification of intermuscular fat remain challenging. In this study, we aimed to develop a fully automated and highly accurate segmentation tool for connective tissue compartments from abdominal CT scans using the open-source Convolutional Neural Network (CNN) DeepMedic. MATERIALS AND METHODS: In this retrospective study, a CNN was developed using data of 1143 consecutive patients undergoing either preinterventional CT for transcatheter aortic valve implantation (TAVI) (82%) or diagnostic CT for liver cirrhosis with portosystemic shunting (PTSS) (18%). All analyses were performed on single-slice images at the L3/L4 level. The data were subdivided into subsets of training (70%), validation (15%), and test data (15%), balanced for TAVI and PTSS patients. To demonstrate the generalizability of the applied method with respect to nonspecific clinical routine data, the model with the highest performance in TAVI and PTSS patients was further tested on 100 randomly selected patients who underwent CT for routine diagnostic purposes at a hospital of maximum care, including critically ill patients. The applicability of the method to native CT examinations was additionally tested on 50 patients. RESULTS: Compared with the ground truth of the test data, the presented method achieved highly accurate segmentation results (subcutaneous adipose tissue [SAT], Dice score [DSC]: 0.98 ± 0.01; visceral adipose tissue [VAT], DSC: 0.96 ± 0.04; skeletal muscles [SM], DSC: 0.95 ± 0.02) and showed excellent generalizability on the routine CT diagnostic patients (SAT, DSC: 0.97 ± 0.04; VAT, DSC: 0.95 ± 0.05; SM, DSC: 0.95 ± 0.04) and also on native CT scans (SAT, DSC: 0.99 ± 0.01; VAT, DSC: 0.97 ± 0.03; SM, DSC: 0.97 ± 0.02). CONCLUSIONS: Fully automated determination of body composition based on CT can be performed with excellent results using the open-source CNN DeepMedic. The trained model is made usable for research by a deployable and sharable application.

Yttrium-90 radioembolization for hepatocellular carcinoma: Outcome prediction with MRI derived fat-free muscle area.

PURPOSE: Sarcopenia is associated with adverse outcomes in several gastrointestinal malignancies and liver cirrhosis. We aimed to study the utility of magnetic resonance imaging (MRI) derived fat-free muscle area (FFMA) to predict clinical outcome in patients receiving yttrium-90 radioembolization (RE) for treatment of hepatocellular carcinoma (HCC). METHODS: Fifty-eight patients with unresectable HCC and pre-interventional liver MRI undergoing salvage RE were retrospectively evaluated. Using axial T2-weighted turbo spin echo sequences, FFMA was calculated by subtraction of the intramuscular adipose tissue area from the total cross-sectional area of paraspinal skeletal muscles at the superior mesenteric artery level. FFMA values lower than 3582 mm2 in male and 2301 mm2 in female patients were defined as low FFMA. Main outcomes were progression-free survival (PFS) and overall survival (OS). For outcome analysis, the Kaplan-Meier method with log rank test and multivariate cox regression analysis were used. RESULTS: Mean time from pre-interventional MRI to RE was 27 ± 20 days. Median OS and PFS after RE were 250 (range: 21-1230 days) and 156 days (range: 21-674 days), respectively. Patients with low FFMA showed significantly reduced OS (197 vs. 294 days, P = 0.024) and tended to have shortened PFS (109 vs. 185 days, P = 0.068). Low FFMA (HR 2.675; P = 0.011), estimated liver tumor burden (HR 4.058; P = 0.001), and Eastern Cooperative Oncology Group (ECOG) performance status (1.763; P = 0.009) were independent predictors of OS on multivariate analysis. CONCLUSIONS: FFMA as a measure of sarcopenia predicts OS and might represent a promising new biomarker for survival prognosis in patients undergoing RE for treatment of unresectable HCC.

MR fingerprinting as a diagnostic tool in patients with frontotemporal lobe degeneration: A pilot study.

Several very rare forms of dementia are associated with characteristic focal atrophy predominantly of the frontal and/or temporal lobes and currently lack imaging solutions to monitor disease. Magnetic resonance fingerprinting (MRF) is a recently developed technique providing quantitative relaxivity maps and images with various tissue contrasts out of a single sequence acquisition. This pilot study explores the utility of MRF-based T1 and T2 mapping to discover focal differences in relaxation times between patients with frontotemporal lobe degenerative dementia and healthy controls. 8 patients and 30 healthy controls underwent a 3 T MRI including an axial 2D spoiled gradient echo MRF sequence. T1 and T2 relaxation maps were generated based on an extended phase graphs algorithm-founded dictionary involving inner product pattern matching. A region of interest (ROI)-based analysis of T1 and T2 relaxation times was performed with FSL and ITK-SNAP. Depending on the brain region analyzed, T1 relaxation times were up to 10.28% longer in patients than in controls reaching significant differences in cortical gray matter (P = .047) and global white matter (P = .023) as well as in both hippocampi (P = .001 left; P = .027 right). T2 relaxation times were similarly longer in the hippocampus by up to 19.18% in patients compared with controls. The clinically most affected patient had the most control-deviant relaxation times. There was a strong correlation of T1 relaxation time in the amygdala with duration of the clinically manifest disease (Spearman Rho = .94; P = .001) and of T1 relaxation times in the left hippocampus with disease severity (Rho = .90, P = .002). In conclusion, MRF-based relaxometry is a promising and time-saving new MRI tool to study focal cerebral alterations and identify patients with frontotemporal lobe degeneration. To validate the results of this pilot study, MRF is worth further exploration as a diagnostic tool in neurodegenerative diseases.

Quantification of liver fibrosis: extracellular volume fraction using an MRI bolus-only technique in a rat animal model.

BACKGROUND: To determine the utility of single-contrast-bolus hepatic extracellular volume (ECV) fraction measurement at different time points to detect and quantify hepatic fibrosis. METHODS: Different grades of liver fibrosis were induced in 23 male Sprague-Dawley rats by carbon-tetrachloride (CCl4) intoxication. In ten control rats, no fibrosis was induced. Native T1 values and ECV fraction were assessed by using quantitative magnetic resonance imaging (MRI) mapping; only one contrast bolus was applied (gadobutrol 0.1 mmol/kg). ECV values were determined 5, 15, and 25 min after injection. Hepatic fibrosis was quantified histologically by Sirius red staining. RESULTS: For the 8-week-CCl4 group, the ECV fraction values obtained 5 (23.5 ± 4.8%, mean ± standard deviation), 15 (23.6 ± 4.8%), and 25 min (23.7 ± 4.7%) after injection were constant over time (p = 0.998); constant data 5-25 min after injection were also observed for the 16-week-CCl4 group and controls. Liver ECV after 15 min significantly increased with the severity of fibrosis: 18.0 ± 3.0% (controls) versus 23.6 ± 4.8% (8-week-CCl4) versus 30.5 ± 3.3% (16-week-CCl4) (p <  0.001). ECV values after 5, 15, and 25 min significantly correlated with Sirius red staining (p <  0.001 for all parameters). CONCLUSIONS: Hepatic ECV obtained using a single-contrast-bolus technique can be measured 5, 15, and 25 min after injection, obtaining constant values over time, each of them being suitable to detect diffuse hepatic fibrosis. In clinical practice, post-contrast T1 relaxation times for liver ECV fraction determination might be obtained at only one time point.