Deep-learning based detection of vessel occlusions on CT-angiography in patients with suspected acute ischemic stroke.

Swift diagnosis and treatment play a decisive role in the clinical outcome of patients with acute ischemic stroke (AIS), and computer-aided diagnosis (CAD) systems can accelerate the underlying diagnostic processes. Here, we developed an artificial neural network (ANN) which allows automated detection of abnormal vessel findings without any a-priori restrictions and in <2 minutes. Pseudo-prospective external validation was performed in consecutive patients with suspected AIS from 4 different hospitals during a 6-month timeframe and demonstrated high sensitivity (≥87%) and negative predictive value (≥93%). Benchmarking against two CE- and FDA-approved software solutions showed significantly higher performance for our ANN with improvements of 25-45% for sensitivity and 4-11% for NPV (p ≤ 0.003 each). We provide an imaging platform ( https://stroke.neuroAI-HD.org ) for online processing of medical imaging data with the developed ANN, including provisions for data crowdsourcing, which will allow continuous refinements and serve as a blueprint to build robust and generalizable AI algorithms.

Independent and additive contribution of white matter hyperintensities and Alzheimer's disease pathology to basal forebrain cholinergic system degeneration.

OBJECTIVES: Degeneration of the cholinergic basal forebrain nuclei (CBFN) system has been studied extensively in Alzheimer's disease (AD). White matter hyperintensities are a hallmark of aging as well as a common co-morbidity of AD, but their contribution to CBFN degeneration has remained unclear. Therefore, we explored the influence of white matter hyperintensities within cholinergic subcortical-cortical projection pathways on CBFN volumes and regional gray matter volumes in AD and age- and gender-matched controls. METHODS: We analyzed magnetic resonance images (MRI) from 42 patients with AD and 87 age- and gender-matched control subjects. We assessed the white matter hyperintensity burden within the cholinergic projection pathways using the Cholinergic Pathways Hyperintensities Scale (CHIPS), and applied probabilistic anatomical maps for the analysis of CBFN volumes, i.e. the Ch1-3 compartment and the Ch4 cell group (nucleus basalis of Meynert), by diffeomorphic anatomical registration using exponentiated lie algebra analysis of voxel-based morphometry. Using multiple linear regression analyses, we explored correlations between regional gray matter volumes and the extent of white matter hyperintensities or CBFN volumes in both groups. RESULTS: In AD, all CBFN volumes were significantly smaller than in controls, and white matter hyperintensity burden within the cholinergic projection pathways was not correlated with CBFN volume. In controls, white matter hyperintensity burden within the cholinergic projection pathways was inversely correlated with CBFN volume when corrected for sex and total intracranial volume, but this correlation was no longer significant after correction for age. Voxel-wise multiple linear regression analyses using threshold-free cluster enhancement revealed that in controls, cholinergic pathway hyperintensities correlated with gray matter loss in perisylvian areas, whereas the were no effects in AD. Moreover, we found that CBFN volumes correlated with distinct regional cortical atrophy patterns in both groups. CONCLUSION: Our results indicate that white matter hyperintensities and AD pathology contribute independently but additively to the degeneration of cholinergic basal forebrain structures. Whereas AD is primarily associated with CBFN volume loss, cholinergic degeneration associated with white matter hyperintensities appears to involve disruption of cholinergic cortical projection fibers with less pronounced effects on CBFN volumes.

Brain Volume Changes after COVID-19 Compared to Healthy Controls by Artificial Intelligence-Based MRI Volumetry.

Cohort studies that quantify volumetric brain data among individuals with different levels of COVID-19 severity are presently limited. It is still uncertain whether there exists a potential correlation between disease severity and the effects of COVID-19 on brain integrity. Our objective was to assess the potential impact of COVID-19 on measured brain volume in patients with asymptomatic/mild and severe disease after recovery from infection, compared with healthy controls, using artificial intelligence (AI)-based MRI volumetry. A total of 155 participants were prospectively enrolled in this IRB-approved analysis of three cohorts with a mild course of COVID-19 (n = 51, MILD), a severe hospitalised course (n = 48, SEV), and healthy controls (n = 56, CTL) all undergoing a standardised MRI protocol of the brain. Automated AI-based determination of various brain volumes in mL and calculation of normalised percentiles of brain volume was performed with mdbrain software, using a 3D T1-weighted magnetisation-prepared rapid gradient echo (MPRAGE) sequence. The automatically measured brain volumes and percentiles were analysed for differences between groups. The estimated influence of COVID-19 and demographic/clinical variables on brain volume was determined using multivariate analysis. There were statistically significant differences in measured brain volumes and percentiles of various brain regions among groups, even after the exclusion of patients undergoing intensive care, with significant volume reductions in COVID-19 patients, which increased with disease severity (SEV > MILD > CTL) and mainly affected the supratentorial grey matter, frontal and parietal lobes, and right thalamus. Severe COVID-19 infection, in addition to established demographic parameters such as age and sex, was a significant predictor of brain volume loss upon multivariate analysis. In conclusion, neocortical brain degeneration was detected in patients who had recovered from SARS-CoV-2 infection compared to healthy controls, worsening with greater initial COVID-19 severity and mainly affecting the fronto-parietal brain and right thalamus, regardless of ICU treatment. This suggests a direct link between COVID-19 infection and subsequent brain atrophy, which may have major implications for clinical management and future cognitive rehabilitation strategies.

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.

From research to clinical practice: a European neuroradiological survey on quantitative advanced MRI implementation.

OBJECTIVE: Quantitative MRI (qMRI) methods provide versatile neuroradiological applications and are a hot topic in research. The degree of their clinical implementation is however barely known. This survey was created to illuminate which and how qMRI techniques are currently applied across Europe. METHODS: In total, 4753 neuroradiologists from 27 countries received an online questionnaire. Demographic and professional data, experience with qMRI techniques in the brain and head and neck, usage, reasons for/against application, and knowledge of the QIBA and EIBALL initiatives were assessed. RESULTS: Two hundred seventy-two responders in 23 countries used the following techniques clinically (mean values in %): DWI (82.0%, n = 223), DSC (67.3%, n = 183), MRS (64.3%, n = 175), DCE (43.4%, n = 118), BOLD-fMRI (42.6%, n = 116), ASL (37.5%, n = 102), fat quantification (25.0%, n = 68), T2 mapping (16.9%, n = 46), T1 mapping (15.1%, n = 41), PET-MRI (11.8%, n = 32), IVIM (5.5%, n = 15), APT-CEST (4.8%, n = 13), and DKI (3.3%, n = 9). The most frequent usage indications for any qMRI technique were tissue differentiation (82.4%, n = 224) and oncological monitoring (72.8%, n = 198). Usage differed between countries, e.g. ASL: Germany (n = 13/63; 20.6%) vs. France (n = 31/40; 77.5%). Neuroradiologists endorsed the use of qMRI because of an improved diagnostic accuracy (89.3%, n = 243), but 50.0% (n = 136) are in need of better technology, 34.9% (n = 95) wish for more communication, and 31.3% need help with result interpretation/generation (n = 85). QIBA and EIBALL were not well known (12.5%, n = 34, and 11.0%, n = 30). CONCLUSIONS: The clinical implementation of qMRI methods is highly variable. Beyond the aspect of readiness for clinical use, better availability of support and a wider dissemination of guidelines could catalyse a broader implementation. KEY POINTS: • Neuroradiologists endorse the use of qMRI techniques as they subjectively improve diagnostic accuracy. • Clinical implementation is highly variable between countries, techniques, and indications. • The use of advanced imaging could be promoted through an increase in technical support and training of both doctors and technicians.

The monocyte-dependent immune response to bacteria is suppressed in smoking-induced COPD.

COPD patients have an increased susceptibility to bacterial airway infections that can induce exacerbations. In response to infections, circulating monocytes become recruited to the infected tissue and secrete cytokines. We hypothesized that this cytokine response is reduced in COPD. Cultured peripheral blood monocytes of never smokers (NS) and smokers without (S) and with COPD (3 study populations, n = 36-37) were stimulated with extracts of Haemophilus influenzae, Staphylococcus aureus, or Streptococcus pneumoniae or with four different pathogen-associated molecular patterns (PAMPs). Four cytokines and 9 PAMP-related signaling molecules were measured and compared between the groups. Granulocyte-macrophage-colony-stimulating-factor responses to all stimulants were reduced in S and COPD compared to NS. Tumor-necrosis-factor-α responses to all bacterial extracts, peptidoglycan, and lipopolysaccharide were reduced in S and/or COPD. Interleukin-10 responses to S. aureus and lipoteichoic acid were increased in COPD. Correlations to pack-years and lung function were found. The peptidoglycan-receptor NOD2 and the mRNA of the lipopolysaccharide-receptor TLR4 were reduced in S and COPD. Cytokine responses of monocytes to bacteria are suppressed by smoking and in COPD possibly due to NOD2 and TLR4 reduction and/or interleukin-10 increase. This might help to explain the increased susceptibility to bacterial infections. These systemic molecular pathologies might be targets for therapeutic strategies to prevent infection-induced exacerbations. KEY MESSAGES: COPD subjects have an increased susceptibility to bacterial infections. This implies defects in the immune response to bacteria and is critical for disease progression. The cytokine response of monocytes to bacteria is reduced in COPD. This might be due to a reduced NOD2 and TLR4 and an increased IL-10 expression. This can explain the increased susceptibility to infections and help to identify drug targets.

Simvastatin requires activation in accessory cells to modulate T-cell responses in asthma and COPD.

T-cell-dependent airway and systemic inflammation triggers the progression of chronic obstructive pulmonary disease (COPD) and asthma. Retrospective studies suggest that simvastatin has anti-inflammatory effects in both diseases but it is unclear, which cell types are targeted. We hypothesized that simvastatin modulates T-cell activity. Circulating CD4+ and CD8+ T-cells, either pure, co-cultured with monocytes or alveolar macrophages (AM) or in peripheral blood mononuclear cells (PBMCs), were ex vivo activated towards Th1/Tc1 or Th2/Tc2 and incubated with simvastatin. Markers for Th1/Tc1 (IFNγ) and Th2/Tc2 (IL-5, IL-13) were measured by ELISA; with PBMCs this was done comparative between 11 healthy never-smokers, 11 current smokers without airflow limitation, 14 smokers with COPD and 11 never-smokers with atopic asthma. T-cell activation induced IFNγ, IL-5 and IL-13 in the presence and absence of accessory cells. Simvastatin did not modulate cytokine expression in pure T-cell fractions. ß-hydroxy-simvastatin acid (activated simvastatin) suppressed IL-5 and IL-13 in pure Th2- and Tc2-cells. Simvastatin suppressed IL-5 and IL-13 in Th2-cells co-cultivated with monocytes or AM, which was partially reversed by the carboxylesterase inhibitor benzil. Simvastatin suppressed IL-5 production of Th2/Tc2-cells in PBMCs without differences between cohorts and IL-13 stronger in never-smokers and asthma compared to COPD. Simvastatin induced IFNγ in Th1/Tc1-cells in PBMCs of all cohorts except asthmatics. Simvastatin requires activation in accessory cells likely by carboxylesterase to suppress IL-5 and IL-13 in Th2/Tc2-cells. The effects on Il-13 are partially reduced in COPD. Asthma pathogenesis prevents simvastatin-induced IFNγ up-regulation. Simvastatin has anti-inflammatory effects that could be of interest for asthma therapy.