Automated intracranial vessel segmentation of 4D flow MRI data in patients with atherosclerotic stenosis using a convolutional neural network.

Introduction: Intracranial 4D flow MRI enables quantitative assessment of hemodynamics in patients with intracranial atherosclerotic disease (ICAD). However, quantitative assessments are still challenging due to the time-consuming vessel segmentation, especially in the presence of stenoses, which can often result in user variability. To improve the reproducibility and robustness as well as to accelerate data analysis, we developed an accurate, fully automated segmentation for stenosed intracranial vessels using deep learning. Methods: 154 dual-VENC 4D flow MRI scans (68 ICAD patients with stenosis, 86 healthy controls) were retrospectively selected. Manual segmentations were used as ground truth for training. For automated segmentation, deep learning was performed using a 3D U-Net. 20 randomly selected cases (10 controls, 10 patients) were separated and solely used for testing. Cross-sectional areas and flow parameters were determined in the Circle of Willis (CoW) and the sinuses. Furthermore, the flow conservation error was calculated. For statistical comparisons, Dice scores (DS), Hausdorff distance (HD), average symmetrical surface distance (ASSD), Bland-Altman analyses, and interclass correlations were computed using the manual segmentations from two independent observers as reference. Finally, three stenosis cases were analyzed in more detail by comparing the 4D flow-based segmentations with segmentations from black blood vessel wall imaging (VWI). Results: Training of the network took approximately 10 h and the average automated segmentation time was 2.2 ± 1.0 s. No significant differences in segmentation performance relative to two independent observers were observed. For the controls, mean DS was 0.85 ± 0.03 for the CoW and 0.86 ± 0.06 for the sinuses. Mean HD was 7.2 ± 1.5 mm (CoW) and 6.6 ± 3.7 mm (sinuses). Mean ASSD was 0.15 ± 0.04 mm (CoW) and 0.22 ± 0.17 mm (sinuses). For the patients, the mean DS was 0.85 ± 0.04 (CoW) and 0.82 ± 0.07 (sinuses), the HD was 8.4 ± 3.1 mm (CoW) and 5.7 ± 1.9 mm (sinuses) and the mean ASSD was 0.22 ± 0.10 mm (CoW) and 0.22 ± 0.11 mm (sinuses). Small bias and limits of agreement were observed in both cohorts for the flow parameters. The assessment of the cross-sectional lumen areas in stenosed vessels revealed very good agreement (ICC: 0.93) with the VWI segmentation but a consistent overestimation (bias ± LOA: 28.1 ± 13.9%). Discussion: Deep learning was successfully applied for fully automated segmentation of stenosed intracranial vasculatures using 4D flow MRI data. The statistical analysis of segmentation and flow metrics demonstrated very good agreement between the CNN and manual segmentation and good performance in stenosed vessels. To further improve the performance and generalization, more ICAD segmentations as well as other intracranial vascular pathologies will be considered in the future.

Detection and Hemodynamic Evaluation of Flap Fenestrations in Type B Aortic Dissection with 4D Flow MRI: Comparison with Conventional MRI and CTA.

PURPOSE: The purpose of this study was to compare dissection flap fenestration visualization between 4D flow MRI, clinical MRI/MRA, and clinical CTA studies and describe the presence of hemodynamically active fenestration flow using 4D flow. MATERIALS AND METHODS: Nineteen patients with type B dissection (age: 57±5 years) who had undergone standard of-care MRI/MRA of the chest including 4D flow MRI were retrospectively identified. Fourteen of the 19 patients also had CTA performed within 2 years of the MRI/MRA study with no interval surgery. Image review was performed independently by two radiologists. The number of fenestrations (including entry and exit tears), location, and flow directionality were recorded. Differences in the rate of detection between techniques was assessed using a Wilcoxon signed rank test. RESULTS: 4D flow detected more fenestrations relative to MRI/MRA [rev 1: +3 (10%), rev 2: +5 (20%)]. There were similar numbers of fenestrations detected by 4D flow relative to CTA [rev 1: +1 (4%), rev 2: -3 (-12%)]. MRI/MRA detected fewer fenestration relative to CTA in this cohort [rev 1: -6 (-24%), rev 2: -5 (-19%)]. No differences were significant. Combining 4D flow and MRI/MRA resulted in additional fenestration detection. Most fenestrations demonstrated biphasic flow over the cardiac cycle (flow entering false lumen in systole and exiting during diastole, rev 1:18/33, rev 2: 16/30). CONCLUSIONS: 4D flow MRI can detect small flap fenestration in type B dissection patients while providing additional information about flow through fenestrations throughout the cardiac cycle relative to CTA and conventional MRI.

Thoracic aorta 3D hemodynamics in pediatric and young adult patients with bicuspid aortic valve.

BACKGROUND: To evaluate the 3D hemodynamics in the thoracic aorta of pediatric and young adult bicuspid aortic valve (BAV) patients. METHODS: 4D flow MRI was performed in 30 pediatric and young adult BAV patients (age: 13.9 ± 4.4 (range: [3.4, 20.7]) years old, M:F = 17:13) as part of this Institutional Review Board-approved study. Nomogram-based aortic root Z-scores were calculated to assess aortic dilatation and degree of aortic stenosis (AS) severity was assessed on MRI. Data analysis included calculation of time-averaged systolic 3D wall shear stress (WSSsys ) along the entire aorta wall, and regional quantification of maximum and mean WSSsys and peak systolic velocity (velsys ) in the ascending aorta (AAo), arch, and descending aorta (DAo). The 4D flow MRI AAo velsys was also compared with echocardiography peak velocity measurements. RESULTS: There was a positive correlation with both mean and max AAo WSSsys and peak AAo velsys (mean: r = 0.84, P < 0.001, max: r = 0.94, P < 0.001) and AS (mean: rS = 0.43, P = 0.02, max: rS = 0.70, P < 0.001). AAo peak velocity was significantly higher when measured with echo compared with 4D flow MRI (2.1 ± 0.98 m/s versus 1.27 ± 0.49 m/s, P < 0.001). CONCLUSION: In pediatric and young adult patients with BAV, AS and peak ascending aorta velocity are associated with increased AAo WSS, while aortic dilation, age, and body surface area do not significantly impact AAo hemodynamics. Prospective studies are required to establish the role of WSS as a risk-stratification tool in these patients.

Neurofunctional effects of quetiapine in patients with bipolar mania.

OBJECTIVES: Several lines of evidence suggest that abnormalities within portions of the extended limbic network involved in affective regulation and expression contribute to the neuropathophysiology of bipolar disorder. In particular, portions of the prefrontal cortex have been implicated in the appearance of manic symptomatology. The effect of atypical antipsychotics on activation of these regions, however, remains poorly understood. METHODS: Twenty-two patients diagnosed with bipolar mania and 26 healthy subjects participated in a baseline functional magnetic resonance imaging scan during which they performed a continuous performance task with neutral and emotional distractors. Nineteen patients with bipolar disorder were treated for eight weeks with quetiapine monotherapy and then rescanned. Regional activity in response to emotional stimuli was compared between healthy and manic subjects at baseline; and in the subjects with bipolar disorder between baseline and eight-week scans. RESULTS: At baseline, functional activity did not differ between subjects with bipolar disorder and healthy subjects in any region examined. After eight weeks of treatment, subjects with bipolar disorder showed a significant decrease in ratings on the Young Mania Rating Scale (YMRS) (p < 0.001), and increased activation in the right orbitofrontal cortex (OFC) (p = 0.002); there was a significant association between increased right OFC activity and YMRS improvement (p = 0.003). CONCLUSIONS: These findings are consistent with suggestions that mania involves a loss of emotional modulatory activity in the prefrontal cortex--restoration of the relatively greater elevation in prefrontal activity widely observed in euthymic patients is associated with clinical improvement. It is not clear, however, whether changes are related to quetiapine treatment or represent a non-specific marker of affective change.

4D flow imaging with MRI.

Magnetic resonance imaging (MRI) has become an important tool for the clinical evaluation of patients with cardiovascular disease. Since its introduction in the late 1980s, 2-dimensional phase contrast MRI (2D PC-MRI) has become a routine part of standard-of-care cardiac MRI for the assessment of regional blood flow in the heart and great vessels. More recently, time-resolved PC-MRI with velocity encoding along all three flow directions and three-dimensional (3D) anatomic coverage (also termed '4D flow MRI') has been developed and applied for the evaluation of cardiovascular hemodynamics in multiple regions of the human body. 4D flow MRI allows for the comprehensive evaluation of complex blood flow patterns by 3D blood flow visualization and flexible retrospective quantification of flow parameters. Recent technical developments, including the utilization of advanced parallel imaging techniques such as k-t GRAPPA, have resulted in reasonable overall scan times, e.g., 8-12 minutes for 4D flow MRI of the aorta and 10-20 minutes for whole heart coverage. As a result, the application of 4D flow MRI in a clinical setting has become more feasible, as documented by an increased number of recent reports on the utility of the technique for the assessment of cardiac and vascular hemodynamics in patient studies. A number of studies have demonstrated the potential of 4D flow MRI to provide an improved assessment of hemodynamics which might aid in the diagnosis and therapeutic management of cardiovascular diseases. The purpose of this review is to describe the methods used for 4D flow MRI acquisition, post-processing and data analysis. In addition, the article provides an overview of the clinical applications of 4D flow MRI and includes a review of applications in the heart, thoracic aorta and hepatic system.

Magnetic resonance spectroscopy imaging of lactate in patients with bipolar disorder.

Although brain lactate levels are typically low and difficult to measure, a few previous investigators have reported that brain lactate levels are elevated in patients with bipolar disorder. The present study investigated the distribution of lactate in bipolar and healthy brains using 2D proton magnetic resonance spectroscopic imaging on a 4-Tesla magnetic resonance imaging system. Ratios of the concentration of lactate to N-acetylaspartate, and of lactate to total creatine, were significantly higher in bipolar than in healthy subjects. Lactate signals were primarily localized to the bipolar subjects' caudate and anterior cingulate cortices, components of the frontal-subcortical circuit, suggesting that affective dysregulation may be related to metabolic abnormalities in this network.

Neurochemical effects of quetiapine in patients with bipolar mania: a proton magnetic resonance spectroscopy study.

Although the neurophysiology underlying pharmacotherapy for bipolar disorder remains poorly understood, recent studies suggest that therapeutic mechanisms may be reflected in changes in concentrations of N-acetylaspartate (NAA), a putative measure of neuronal integrity and metabolism. In this study, we used magnetic resonance spectroscopy (MRS) to examine prefrontal NAA in patients receiving quetiapine for bipolar mania. On the basis of previous findings, we hypothesized that remission would be associated with increased NAA concentrations in the prefrontal cortex. Thirty-one manic bipolar patients and 13 healthy subjects were recruited to participate in this prospective study. All subjects participated in MRS at baseline and after 8 weeks of treatment. Bipolar subjects received open-label quetiapine monotherapy (mean dose [SD], 584 [191] mg). Fourteen patients remitted (Young Mania Rating Scale ≤ 12) ("remitters"), 11 patients did not ("nonremitters"), and 6 patients were lost to follow-up. Bipolar and healthy subjects did not significantly differ in baseline NAA or degree of change during the 8 weeks. Remitters showed greater mean baseline NAA concentrations in the right ventrolateral prefrontal cortex compared with nonremitters (P < 0.05). In the anterior cingulate, remitters showed near significantly decreased baseline NAA concentrations at baseline (P < 0.06), and significant differences in NAA change during the 8 weeks of treatment (P < 0.03). Manic patients who remitted with quetiapine treatment in the course of this study exhibited distinct patterns of baseline prefrontal NAA concentration, coupled with decreased NAA in the anterior cingulate with treatment; the latter possibly reflecting disparate effects of quetiapine on neuronal metabolism. These data support suggestions that therapeutic effects of quetiapine involve metabolic effects on specific prefrontal regions.

Progressive neurostructural changes in adolescent and adult patients with bipolar disorder.

OBJECTIVES: Several lines of evidence suggest that bipolar disorder is associated with progressive changes in gray matter volume (GMV), particularly in brain structures involved in emotional regulation and expression. The majority of these studies however, have been cross-sectional in nature. In this study we compared baseline and follow-up scans in groups of bipolar disorder and healthy subjects. We hypothesized bipolar disorder subjects would demonstrate significant GMV changes over time. METHODS: A total of 58 bipolar disorder and 48 healthy subjects participated in structural magnetic resonance imaging (MRI). Subjects were rescanned 3-34 months after their baseline MRI. MRI images were segmented, normalized to standard stereotactic space, and compared voxel-by-voxel using statistical parametrical mapping software (SPM2). A model was developed to investigate differences in GMV at baseline, and associated with time and episodes, as well as in comparison to healthy subjects. RESULTS: We observed increases in GMV in bipolar disorder subjects across several brain regions at baseline and over time, including portions of the prefrontal cortex as well as limbic and subcortical structures. Time-related changes differed to some degree between adolescent and adult bipolar disorder subjects. The interval between scans positively correlated with GMV increases in bipolar disorder subjects in portions of the prefrontal cortex, and both illness duration and number of depressive episodes were associated with increased GMV in subcortical and limbic structures. CONCLUSIONS: Our findings support suggestions that widely observed progressive neurofunctional changes in bipolar disorder patients may be related to structural brain abnormalities in anterior limbic structures. Abnormalities largely involve regions previously noted to be integral to emotional expression and regulation, and appear to vary by age.