Automatic segmentation of the thalamus using a massively trained 3D convolutional neural network: higher sensitivity for the detection of reduced thalamus volume by improved inter-scanner stability.

OBJECTIVES: To develop an automatic method for accurate and robust thalamus segmentation in T1w-MRI for widespread clinical use without the need for strict harmonization of acquisition protocols and/or scanner-specific normal databases. METHODS: A three-dimensional convolutional neural network (3D-CNN) was trained on 1975 T1w volumes from 170 MRI scanners using thalamus masks generated with FSL-FIRST as ground truth. Accuracy was evaluated with 18 manually labeled expert masks. Intra- and inter-scanner test-retest stability were assessed with 477 T1w volumes of a single healthy subject scanned on 123 MRI scanners. The sensitivity of 3D-CNN-based volume estimates for the detection of thalamus atrophy was tested with 127 multiple sclerosis (MS) patients and a normal database comprising 4872 T1w volumes from 160 scanners. The 3D-CNN was compared with a publicly available 2D-CNN (FastSurfer) and FSL. RESULTS: The Dice similarity coefficient of the automatic thalamus segmentation with manual expert delineation was similar for all tested methods (3D-CNN and FastSurfer 0.86 ± 0.02, FSL 0.87 ± 0.02). The standard deviation of the single healthy subject's thalamus volume estimates was lowest with 3D-CNN for repeat scans on the same MRI scanner (0.08 mL, FastSurfer 0.09 mL, FSL 0.15 mL) and for repeat scans on different scanners (0.28 mL, FastSurfer 0.62 mL, FSL 0.63 mL). The proportion of MS patients with significantly reduced thalamus volume was highest for 3D-CNN (24%, FastSurfer 16%, FSL 11%). CONCLUSION: The novel 3D-CNN allows accurate thalamus segmentation, similar to state-of-the-art methods, with considerably improved robustness with respect to scanner-related variability of image characteristics. This might result in higher sensitivity for the detection of disease-related thalamus atrophy. KEY POINTS: • A three-dimensional convolutional neural network was trained for automatic segmentation of the thalamus with a heterogeneous sample of T1w-MRI from 1975 patients scanned on 170 different scanners. • The network provided high accuracy for thalamus segmentation with manual segmentation by experts as ground truth. • Inter-scanner variability of thalamus volume estimates across different MRI scanners was reduced by more than 50%, resulting in increased sensitivity for the detection of thalamus atrophy.

Exploring in vivo lesion myelination dynamics: Longitudinal Myelin Water Imaging in early Multiple Sclerosis.

BACKGROUND: Multiple Sclerosis (MS) lesions are pathologically heterogeneous and the temporal behavior in terms of growth and myelination status of individual lesions is highly variable, especially in the early phase of the disease. Thus, monitoring the development of individual lesion myelination by using quantitative magnetic resonance myelin water imaging (MWI) could be valuable to capture the variability of disease pathology and get an individual insight into the subclinical disease activity. OBJECTIVE: The goal of this work was (1) to observe the variation and longitudinal change of in vivo lesion myelination by means of MWI and its parameter Myelin Water Fraction (MWF), and, (2) to identify individual lesion myelination patterns in early MS. METHODS: In this study n = 12 patients obtained conventional MRI and quantitative MWI derived from multi-component driven equilibrium single pulse observation of T1 and T2 (mcDESPOT) within four weeks after presenting a clinically isolated syndrome and remained within the study if clinically definitive MS was diagnosed within the 12 months study period. Four MRI sessions were acquired at baseline, 3, 6, and 12 months. The short-term and long-term variability of MWF maps was evaluated by scan-rescan measures and the coefficient of variation was determined in four healthy controls. Tracking of individual lesions was performed using the Automatic Follow-up of Individual Lesions (AFIL) algorithm. Lesion volume and MWF were evaluated for every individual lesion in all patients. Median lesion MWF change was used to define lesion categories as decreasing, varying, increasing and invariant for MWF variation. RESULTS: In total n = 386 T2 lesions were detected with a subset of n = 225 permanent lesions present at all four time-points. Among those, a heterogeneous lesion MWF reduction was found, with the majority of lesions bearing only mild MWF reduction, approximately a third with an intermediate MWF decrease and highest MWF reduction in acute-inflammatory active lesions. A moderate negative correlation was determined between individual lesion volumes and median MWF consistent across all time-points. Permanent lesions featured variable temporal dynamics with the majority of varying MWF (58 %), however decreasing (16 %), increasing (15 %) and invariant (11 %) subgroups could be identified resembling demyelinating activity and post-demyelinating inactivity known from histopathology studies. Inflammatory-active enhancing lesions showed a distinct pattern of MWF reduction followed by partial recovery after 3 months. This was similar in new enhancing lesions and those with a non-enhancing precursor lesion. CONCLUSION: This work provides in vivo evidence for an individual evolution of early demyelinated MS lesions measured by means of MWF imaging. Our results support the hypothesis, that MS lesions undergo multiple demyelination and remyelination episodes in the early acute phase. The in vivo MRI surrogate of myelin turnover bears capacity as a novel biomarker to select and potentially monitor personalized MS treatment.

Single-subject analysis of regional brain volumetric measures can be strongly influenced by the method for head size adjustment.

PURPOSE: Total intracranial volume (TIV) is often a nuisance covariate in MRI-based brain volumetry. This study compared two TIV adjustment methods with respect to their impact on z-scores in single subject analyses of regional brain volume estimates. METHODS: Brain parenchyma, hippocampus, thalamus, and TIV were segmented in a normal database comprising 5059 T1w images. Regional volume estimates were adjusted for TIV using the residual method or the proportion method. Age was taken into account by regression with both methods. TIV- and age-adjusted regional volumes were transformed to z-scores and then compared between the two adjustment methods. Their impact on the detection of thalamus atrophy was tested in 127 patients with multiple sclerosis. RESULTS: The residual method removed the association with TIV in all regions. The proportion method resulted in a switch of the direction without relevant change of the strength of the association. The reduction of physiological between-subject variability was larger with the residual method than with the proportion method. The difference between z-scores obtained with the residual method versus the proportion method was strongly correlated with TIV. It was larger than one z-score point in 5% of the subjects. The area under the ROC curve of the TIV- and age-adjusted thalamus volume for identification of multiple sclerosis patients was larger with the residual method than with the proportion method (0.84 versus 0.79). CONCLUSION: The residual method should be preferred for TIV and age adjustments of T1w-MRI-based brain volume estimates in single subject analyses.

Infratentorial lesions in multiple sclerosis patients: intra- and inter-rater variability in comparison to a fully automated segmentation using 3D convolutional neural networks.

OBJECTIVE: Automated quantification of infratentorial multiple sclerosis lesions on magnetic resonance imaging is clinically relevant but challenging. To overcome some of these problems, we propose a fully automated lesion segmentation algorithm using 3D convolutional neural networks (CNNs). METHODS: The CNN was trained on a FLAIR image alone or on FLAIR and T1-weighted images from 1809 patients acquired on 156 different scanners. An additional training using an extra class for infratentorial lesions was implemented. Three experienced raters manually annotated three datasets from 123 MS patients from different scanners. RESULTS: The inter-rater sensitivity (SEN) was 80% for supratentorial lesions but only 62% for infratentorial lesions. There was no statistically significant difference between the inter-rater SEN and the SEN of the CNN with respect to the raters. For supratentorial lesions, the CNN featured an intra-rater intra-scanner SEN of 0.97 (R1 = 0.90, R2 = 0.84) and for infratentorial lesion a SEN of 0.93 (R1 = 0.61, R2 = 0.73). CONCLUSION: The performance of the CNN improved significantly for infratentorial lesions when specifically trained on infratentorial lesions using a T1 image as an additional input and matches the detection performance of experienced raters. Furthermore, for infratentorial lesions the CNN was more robust against repeated scans than experienced raters. KEY POINTS: • A 3D convolutional neural network was trained on MRI data from 1809 patients (156 different scanners) for the quantification of supratentorial and infratentorial multiple sclerosis lesions. • Inter-rater variability was higher for infratentorial lesions than for supratentorial lesions. The performance of the 3D convolutional neural network (CNN) improved significantly for infratentorial lesions when specifically trained on infratentorial lesions using a T1 image as an additional input. • The detection performance of the CNN matches the detection performance of experienced raters.

CT Angiography Manual Multiplanar Vessel Diameter Measurement vs. Semiautomated Perpendicular Area Minimal Caliber Computation of Internal Carotid Artery Stenosis.

Objective: To determine the diagnostic agreement of CT angiography (CTA) manual multiplanar reformatting (MPR) stenosis diameter measurement and semiautomated perpendicular stenosis area minimal caliber computation of extracranial internal carotid artery (ICA) stenosis. Methods: We analyzed acute cerebral ischemia CTA at our tertiary stroke center in a 12-month period. Prospective NASCET-type stenosis grading for each ICA was independently performed using (1) MPR to manually determine diameters and (2) perpendicular stenosis area with minimal caliber semiautomated computation to grade luminal constriction. Corresponding to clinically relevant NASCET strata, results were grouped into severity ranges: normal, 1-49%, 50-69%, and 70-99%, and occlusion. Results: We included 647 ICA pairs from 330 patients (median age of 74 [66-80, IQR]; 38-92 years; 58% men; median NIHSS 4 [1-9, IQR]). MPR diameter and semiautomated caliber measurements resulted in stenosis grades of 0-49% in 143 vs. 93, 50-69% in 29 vs. 27, 70-99% in 6 vs. 14, and occlusion in 34 vs. 34 ICAs (p = 0.003), respectively. We found excellent reliability between repeated manual CTA assessments of one expert reader (ICC = 0.997; 95% CI, 0.993-0.999) and assessments of two expert readers (ICC = 0.972; 95% CI, 0.936-0.988). For the semiautomated vessel analysis software, both intrarater reliability and interrater reliability were similarly strong (ICC = 0.981; 95% CI, 0.952-0.992 and ICC = 0.745; 95% CI, 0.486-0.883, respectively). However, Bland-Altman analysis revealed a mean difference of 1.6% between the methods within disease range with wide 95% limits of agreement (-16.7-19.8%). This interval even increased with exclusively considered vessel pairs of stenosis ≥1% (mean 5.3%; -24.1-34.7%) or symptomatic stenosis ≥50% (mean 0.1%; -25.7-26.0%). Conclusion: Our findings suggest that MPR-based diameter measurement and the semiautomated perpendicular area minimal caliber computation methods cannot be used interchangeably for the quantification of ICA steno-occlusive disease.

Video-Oculography-Assisted Head Impulse Test and Caloric Testing for Detecting Stroke in Acute Vertigo Patients via Modified HINTS Plus.

BACKGROUND: We assessed whether detection of stroke underlying acute vertigo using HINTS plus (head-impulse test, nystagmus type, test of skew, hearing loss) can be improved by video-oculography for automated head-impulse test (V-HIT) analysis. METHODS: We evaluated patients with acute vestibular syndrome (AVS) presenting to the emergency room using HINTS plus and V-HIT-assisted HINTS plus in a randomized sequence followed by cranial MRI and caloric testing. Image-confirmed posterior circulation stroke or vertebrobasilar TIA were the reference standards to calculate diagnostic accuracy. We repeated statistical analysis for a third protocol that was composed post hoc by replacing the head-impulse test with caloric testing in the HINTS plus protocol. RESULTS: We included 30 AVS patients (ages 55.4 ± 17.2 years, 14 females). Of these, 11 (36.7%) had posterior circulation stroke (n = 4) or TIA (n = 7). Acute V-HIT-assisted HINTS plus was feasible and displayed tendentially higher accuracy than conventional HINTS plus (sensitivity: 81.8%, 95% CI 48.2-97.7%; specificity 31.6%, 95% CI 12.6-56.6% vs. sensitivity 72.7%, 95% CI 39.0-94.0%; specificity 36.8%, 95% CI 16.3-61.6%). The new caloric-supported algorithm showed high accuracy (sensitivity 100%, 95% CI 66.4-100%; specificity 66.7%, 95% CI 41-86.7%). CONCLUSIONS: Our study provides pilot data on V-HIT-assisted HINTS plus for acute AVS assessment and indicates the diagnostic value of integrated acute caloric testing.

Acquired olfactory loss alters functional connectivity and morphology.

Removing function from a developed and functional sensory system is known to alter both cerebral morphology and functional connections. To date, a majority of studies assessing sensory-dependent plasticity have focused on effects from either early onset or long-term sensory loss and little is known how the recent sensory loss affects the human brain. With the aim of determining how recent sensory loss affects cerebral morphology and functional connectivity, we assessed differences between individuals with acquired olfactory loss (duration 7-36 months) and matched healthy controls in their grey matter volume, using multivariate pattern analyses, and functional connectivity, using dynamic connectivity analyses, within and from the olfactory cortex. Our results demonstrate that acquired olfactory loss is associated with altered grey matter volume in, among others, posterior piriform cortex, a core olfactory processing area, as well as the inferior frontal gyrus and angular gyrus. In addition, compared to controls, individuals with acquired anosmia displayed significantly stronger dynamic functional connectivity from the posterior piriform cortex to, among others, the angular gyrus, a known multisensory integration area. When assessing differences in dynamic functional connectivity from the angular gyrus, individuals with acquired anosmia had stronger connectivity from the angular gyrus to areas primary responsible for basic visual processing. These results demonstrate that recently acquired sensory loss is associated with both changed cerebral morphology within core olfactory areas and increase dynamic functional connectivity from olfactory cortex to cerebral areas processing multisensory integration.

Open-access quantitative MRI data of the spinal cord and reproducibility across participants, sites and manufacturers.

In a companion paper by Cohen-Adad et al. we introduce the spine generic quantitative MRI protocol that provides valuable metrics for assessing spinal cord macrostructural and microstructural integrity. This protocol was used to acquire a single subject dataset across 19 centers and a multi-subject dataset across 42 centers (for a total of 260 participants), spanning the three main MRI manufacturers: GE, Philips and Siemens. Both datasets are publicly available via git-annex. Data were analysed using the Spinal Cord Toolbox to produce normative values as well as inter/intra-site and inter/intra-manufacturer statistics. Reproducibility for the spine generic protocol was high across sites and manufacturers, with an average inter-site coefficient of variation of less than 5% for all the metrics. Full documentation and results can be found at https://spine-generic.rtfd.io/ . The datasets and analysis pipeline will help pave the way towards accessible and reproducible quantitative MRI in the spinal cord.

Individual odor hedonic perception is coded in temporal joint network activity.

INTRO: The human sense of smell is highly individual and characterized by a strong variability in the perception and evaluation of olfactory stimuli, depending on cultural imprint and current physiological conditions. Since this individual perspective has often been neglected in fMRI studies on olfactory hedonic coding, this study focuses on the neuronal activity and connectivity patterns resulting from subject-specific olfactory stimulation. METHODS: Thirty-one normosmic participants took part in a fMRI block designed paradigm consisting of three olfactory stimulation sessions. The most pleasant and unpleasant odors were individually specified during a pre-test for each participant and validated in the main experiment. Mean activation and functional connectivity analysis focusing on the right and left piriform cortex were performed for the predefined olfactory regions-of-interest (ROIs) and compared between the three olfactory conditions. RESULTS: Individual unpleasant olfactory stimulation as compared to pleasant or neutral did not alter mean BOLD activation in the predefined olfactory ROIs but led to a change in connectivity pattern in the right piriform cortex. CONCLUSION: Our data suggests that the individual pleasantness of odors is not detectable by average BOLD magnitude changes in primary or secondary olfactory brain areas, but reflected in temporal patterns of joint activation that create a network between the right piriform cortex, the left insular cortex, the orbitofrontal cortex, and the precentral gyrus. This network may serve the evolutionary defense mechanism of olfaction by preparing goal-directed action.

Age-dependent cut-offs for pathological deep gray matter and thalamic volume loss using Jacobian integration.

INTRODUCTION: Several recent studies indicate that deep gray matter or thalamic volume loss (VL) might be promising surrogate markers of disease activity in multiple sclerosis (MS) patients. To allow applying these markers to individual MS patients in clinical routine, age-dependent cut-offs distinguishing physiological from pathological VL and an estimation of the measurement error, which provides the confidence of the result, are to be defined. METHODS: Longitudinal MRI scans of the following cohorts were analyzed in this study: 189 healthy controls (HC) (mean age 54 years, 22% female), 98 MS patients from Zurich university hospital (mean age 34 years, 62% female), 33 MS patients from Dresden university hospital (mean age 38 years, 60% female), and publicly available reliability data sets consisting of 162 short-term MRI scan-rescan pairs with scan intervals of days or few weeks. Percentage annualized whole brain volume loss (BVL), gray matter (GM) volume loss (GMVL), deep gray matter volume loss (deep GMVL), and thalamic volume loss (ThalaVL) were computed deploying the Jacobian integration (JI) method. BVL was additionally computed using Siena, an established method used in many Phase III drug trials. A linear mixed effect model was used to estimate the measurement error as the standard deviation (SD) of model residuals of all 162 scan-rescan pairs For estimation of age-dependent cut-offs, a quadratic regression function between age and the corresponding annualized VL values of the HC was computed. The 5th percentile was defined as the threshold for pathological VL per year since 95% of HC subjects exhibit a less pronounced VL for a given age. For the MS patients BVL, GMVL, deep GMVL, and ThalaVL were mutually compared and a paired t-test was used to test whether there are systematic differences in VL between these brain regions. RESULTS: Siena and JI showed a high agreement for BVL measures, with a median absolute difference of 0.1% and a correlation coefficient of r = 0.78. Siena and GMVL showed a similar standard deviation (SD) of the scan-rescan error of 0.28% and 0.29%, respectively. For deep GMVL, ThalaVL the SD of the scan-rescan error was slightly higher (0.43% and 0.5%, respectively). Among the HC the thalamus showed the highest mean VL (-0.16%, -0.39%, and -0.59% at ages 35, 55, and 75, respectively). Corresponding cut-offs for a pathological VL/year were -0.68%, -0.91%, and -1.11%. The MS cohorts did not differ in BVL and GMVL. However, both MS cohorts showed a significantly (p = 0.05) stronger deep GMVL than BVL per year. CONCLUSION: It might be methodologically feasible to assess deep GMVL using JI in individual MS patients. However, age and the measurement error need to be taken into account. Furthermore, deep GMVL may be used as a complementary marker to BVL since MS patients exhibit a significantly stronger deep GMVL than BVL.

Fully automated longitudinal segmentation of new or enlarged multiple sclerosis lesions using 3D convolutional neural networks.

The quantification of new or enlarged lesions from follow-up MRI scans is an important surrogate of clinical disease activity in patients with multiple sclerosis (MS). Not only is manual segmentation time consuming, but inter-rater variability is high. Currently, only a few fully automated methods are available. We address this gap in the field by employing a 3D convolutional neural network (CNN) with encoder-decoder architecture for fully automatic longitudinal lesion segmentation. Input data consist of two fluid attenuated inversion recovery (FLAIR) images (baseline and follow-up) per patient. Each image is entered into the encoder and the feature maps are concatenated and then fed into the decoder. The output is a 3D mask indicating new or enlarged lesions (compared to the baseline scan). The proposed method was trained on 1809 single point and 1444 longitudinal patient data sets and then validated on 185 independent longitudinal data sets from two different scanners. From the two validation data sets, manual segmentations were available from three experienced raters, respectively. The performance of the proposed method was compared to the open source Lesion Segmentation Toolbox (LST), which is a current state-of-art longitudinal lesion segmentation method. The mean lesion-wise inter-rater sensitivity was 62%, while the mean inter-rater number of false positive (FP) findings was 0.41 lesions per case. The two validated algorithms showed a mean sensitivity of 60% (CNN), 46% (LST) and a mean FP of 0.48 (CNN), 1.86 (LST) per case. Sensitivity and number of FP were not significantly different (p < 0.05) between the CNN and manual raters. New or enlarged lesions counted by the CNN algorithm appeared to be comparable with manual expert ratings. The proposed algorithm seems to outperform currently available approaches, particularly LST. The high inter-rater variability in case of manual segmentation indicates the complexity of identifying new or enlarged lesions. An automated CNN-based approach can quickly provide an independent and deterministic assessment of new or enlarged lesions from baseline to follow-up scans with acceptable reliability.

Parcellation of the Subthalamic Nucleus in Parkinson's Disease: A Retrospective Analysis of Atlas- and Diffusion-Based Methods.

BACKGROUND: Deep brain stimulation (DBS) is an established method of treatment for Parkinson's disease (PD). A stimulation sweet spot at the interface between the motor and associative clusters of the subthalamic nucleus (STN) has recently been postulated. The aim of this study was to analyze the available clustering methods for the STN and their correlation to outcome. METHODS: This is a retrospective analysis of a group of 20 patients implanted with a DBS device for PD. Atlas-based and diffusion tractography-based parcellation of the STN was performed. The distances of the electrode to the obtained clusters were compared to each other and to outcome parameters, which included levodopa equivalent dose (LED) reduction, Unified Parkinson's Disease Rating Scale (UPDRS)-III scores, and reduction in scores for items 32 and 36 of the UPDRS-IV. RESULTS: The implanted electrodes were located nearest to the motor clusters of the STN. The following significant associations with postoperative LED reduction were found: (1) distance of the electrode to the motor cluster in the Accolla and DISTAL atlases (p < 0.01) and (2) distance of the electrode to the supplementary motor area cluster (p = 0.02). There was no association with either the UPDRS-III or the UPDRS-IV score. CONCLUSIONS: The results of this study suggest the possibility that atlas-based clustering, as well as diffusion tractography-based parcellation, can be useful in estimating the stimulation target ("sweet spot") for STN-DBS in PD patients. Atlas-based as well as diffusion-based clustering might become a useful tool in DBS trajectory planning.

Harnessing Real-World Data to Inform Decision-Making: Multiple Sclerosis Partners Advancing Technology and Health Solutions (MS PATHS).

Background: Multiple Sclerosis Partners Advancing Technology and Health Solutions (MS PATHS) is the first example of a learning health system in multiple sclerosis (MS). This paper describes the initial implementation of MS PATHS and initial patient characteristics. Methods: MS PATHS is an ongoing initiative conducted in 10 healthcare institutions in three countries, each contributing standardized information acquired during routine care. Institutional participation required the following: active MS patient census of ≥500, at least one Siemens 3T magnetic resonance imaging scanner, and willingness to standardize patient assessments, share standardized data for research, and offer universal enrolment to capture a representative sample. The eligible participants have diagnosis of MS, including clinically isolated syndrome, and consent for sharing pseudonymized data for research. MS PATHS incorporates a self-administered patient assessment tool, the Multiple Sclerosis Performance Test, to collect a structured history, patient-reported outcomes, and quantitative testing of cognition, vision, dexterity, and walking speed. Brain magnetic resonance imaging is acquired using standardized acquisition sequences on Siemens 3T scanners. Quantitative measures of brain volume and lesion load are obtained. Using a separate consent, the patients contribute DNA, RNA, and serum for future research. The clinicians retain complete autonomy in using MS PATHS data in patient care. A shared governance model ensures transparent data and sample access for research. Results: As of August 5, 2019, MS PATHS enrolment included participants (n = 16,568) with broad ranges of disease subtypes, duration, and severity. Overall, 14,643 (88.4%) participants contributed data at one or more time points. The average patient contributed 15.6 person-months of follow-up (95% CI: 15.5-15.8); overall, 166,158 person-months of follow-up have been accumulated. Those with relapsing-remitting MS demonstrated more demographic heterogeneity than the participants in six randomized phase 3 MS treatment trials. Across sites, a significant variation was observed in the follow-up frequency and the patterns of disease-modifying therapy use. Conclusions: Through digital health technology, it is feasible to collect standardized, quantitative, and interpretable data from each patient in busy MS practices, facilitating the merger of research and patient care. This approach holds promise for data-driven clinical decisions and accelerated systematic learning.

Multiple sclerosis lesion activity segmentation with attention-guided two-path CNNs.

Multiple sclerosis is an inflammatory autoimmune demyelinating disease that is characterized by lesions in the central nervous system. Typically, magnetic resonance imaging (MRI) is used for tracking disease progression. Automatic image processing methods can be used to segment lesions and derive quantitative lesion parameters. So far, methods have focused on lesion segmentation for individual MRI scans. However, for monitoring disease progression, lesion activity in terms of new and enlarging lesions between two time points is a crucial biomarker. For this problem, several classic methods have been proposed, e.g., using difference volumes. Despite their success for single-volume lesion segmentation, deep learning approaches are still rare for lesion activity segmentation. In this work, convolutional neural networks (CNNs) are studied for lesion activity segmentation from two time points. For this task, CNNs are designed and evaluated that combine the information from two points in different ways. In particular, two-path architectures with attention-guided interactions are proposed that enable effective information exchange between the two time point's processing paths. It is demonstrated that deep learning-based methods outperform classic approaches and it is shown that attention-guided interactions significantly improve performance. Furthermore, the attention modules produce plausible attention maps that have a masking effect that suppresses old, irrelevant lesions. A lesion-wise false positive rate of 26.4% is achieved at a true positive rate of 74.2%, which is not significantly different from the interrater performance.

Association of Hypertensive Intracerebral Hemorrhage with Left Ventricular Hypertrophy on Transthoracic Echocardiography.

INTRODUCTION: Arterial hypertension is the most frequent cause for spontaneous intracerebral hemorrhage (sICH) and may also cause left ventricular hypertrophy (LVH). We sought to analyze whether hypertensive sICH etiology is associated with LVH. METHODS: We analyzed consecutive patients with sICH who were admitted to our tertiary stroke center during a four-year period and underwent transthoracic echocardiography (TTE) as part of the diagnostic work-up. We defined hypertensive sICH as typical localization of hemorrhage in patients with arterial hypertension and no other identified sICH etiology. We defined an increased end-diastolic interventricular septal wall thickness of ≥11 mm on TTE as a surrogate parameter for LVH. RESULTS: Among 395 patients with sICH, 260 patients (65.8%) received TTE as part of their diagnostic work-up. The median age was 71 years (interquartile range (IQR) 17), 160 patients (61.5%) were male, the median baseline National Institute of Health Stroke Scale (NIHSS) score was 8 (IQR 13). Of these, 159 (61.2%) patients had a hypertensive sICH and 156 patients (60%) had LVH. In univariable (113/159 (71.1%) vs. 43/101 (42.6%); odds ratio (OR) 3.31; 95% confidence interval (CI95%) 1.97-5.62); and multivariable (adjusted OR 2.95; CI95% 1.29-6.74) analysis, hypertensive sICH was associated with LVH. CONCLUSIONS: In patients with sICH, LVH is associated with hypertensive bleeding etiology. Performing TTE is meaningful for diagnosis of comorbidities and clarification of bleeding etiology in these patients. Future studies should include long-term outcome parameters and assess left ventricular mass as main indicator for LVH.

Aberrancies of Brain Network Structures in Patients with Anosmia.

Patients with anosmia exhibit structural and functional brain abnormalities. The present study explored changes in brain white matter (WM) in non-neurodegenerative anosmia using diffusion-tensor-based network analysis. Twenty patients with anosmia and sixteen healthy controls were recruited in the cross-sectional, case-control study. Participants underwent olfactory tests (orthonasal and retronasal), neuropsychological assessment (cognitive function and depressive symptoms) and diffusion tensor imaging measurement. Tract-Based Spatial Statistics, graph theoretical analysis and Network-Based Statistics were used to explore the white matter. There was no significant difference in fractional anisotropy (FA) between patients and controls. In global network topological properties comparisons, patients exhibited higher γ and λ levels than controls, and both groups satisfied the criteria of small-world (σ > 1). In local network topological properties, patients had reduced betweenness, degree and efficiency (global and local), as well as increased shortest path length and cluster coefficient in olfactory-related brain areas (anterior cingulum, lenticular nucleus, putamen, hippocampus, amygdala, caudate nucleus, orbito-frontal gyrus). Olfactory threshold scores and the retronasal score were negatively correlated with γ and λ, and the retronasal score was positively correlated with FA values in certain WM tracts, i.e. middle cerebellar peduncle, right inferior cerebellar peduncle, left inferior cerebellar peduncle, right cerebral peduncle, left cerebral peduncle, left cingulum (cingulate gyrus), right cingulum (hippocampus), superior fronto-occipital fasciculus, and, left tapetum. Patients with anosmia demonstrated relevant WM network dysfunction though their structural integrity remained intact. Their retronasal olfaction deficits revealed to be more strongly associated with WM alterations compared with orthonasal olfactory scores.