Building an Open-Vocabulary Video CLIP Model With Better Architectures, Optimization and Data.

Despite significant results achieved by Contrastive Language-Image Pretraining (CLIP) in zero-shot image recognition, limited effort has been made exploring its potential for zero-shot video recognition. This paper presents Open-VCLIP++, a simple yet effective framework that adapts CLIP to a strong zero-shot video classifier, capable of identifying novel actions and events during testing. Open-VCLIP++ minimally modifies CLIP to capture spatial-temporal relationships in videos, thereby creating a specialized video classifier while striving for generalization. We formally demonstrate that training Open-VCLIP++ is tantamount to continual learning with zero historical data. To address this problem, we introduce Interpolated Weight Optimization, a technique that leverages the advantages of weight interpolation during both training and testing. Furthermore, we build upon large language models to produce fine-grained video descriptions. These detailed descriptions are further aligned with video features, facilitating a better transfer of CLIP to the video domain. Our approach is evaluated on three widely used action recognition datasets, following a variety of zero-shot evaluation protocols. The results demonstrate that our method surpasses existing state-of-the-art techniques by significant margins. Specifically, we achieve zero-shot accuracy scores of 88.1%, 58.7%, and 81.2% on UCF, HMDB, and Kinetics-600 datasets respectively, outpacing the best-performing alternative methods by 8.5%, 8.2%, and 12.3%. We also evaluate our approach on the MSR-VTT video-text retrieval dataset, where it delivers competitive video-to-text and text-to-video retrieval performance, while utilizing substantially less fine-tuning data compared to other methods. Code is released at https://github.com/wengzejia1/Open-VCLIP.

Choroid plexus vascular reactivity in moyamoya: Implications for choroid plexus regulation in ischemic stress.

BACKGROUND AND PURPOSE: Choroid plexus (ChP) hyperemia has been observed in patients with intracranial vasculopathy and to reduce following successful surgical revascularization. This observation may be attributable to impaired vascular reserve of the ChP or other factors, such as the ChP responding to circulating markers of stress. We extend this work to test the hypothesis that vascular reserve of the ChP is unrelated to intracranial vasculopathy. METHODS: We performed hypercapnic reactivity (blood oxygenation level-dependent; echo time = 35 ms; spatial resolution = 3.5 × 3.5 × 3.5 mm, repetition time = 2000 ms) and catheter angiography assessments of ChP reserve capacity and vascular patency in moyamoya patients (n = 53) with and without prior surgical revascularization. Time regression analyses quantified maximum cerebrovascular reactivity and reactivity delay time in ChP and cortical flow territories of major intracranial vessels with steno-occlusion graded as <70%, 70%-99%, and occlusion using Warfarin-Aspirin-Symptomatic-Intracranial-Disease stenosis grading criteria. Analysis of variance (significance: two-sided Bonferroni-corrected p < .05) was applied to evaluate cortical and ChP reactivity, after accounting for end-tidal carbon dioxide change, for differing vasculopathy categories. RESULTS: In patients without prior revascularization, arterial vasculopathy was associated with reduced cortical reactivity and lengthened reactivity delay (p ≤ .01), as expected. Regardless of surgical history, the ChP reactivity metrics were not significantly related to the degree of proximal stenosis, consistent with ChP reactivity being largely preserved in this population. CONCLUSIONS: Findings are consistent with ChP reactivity in moyamoya not being dependent on observed vasculopathy. Future work may investigate the extent to which ChP hyperemia in chronic ischemia reflects circulating markers of glial or ischemic stress.

Towards Transferable Adversarial Attacks on Image and Video Transformers.

The transferability of adversarial examples across different convolutional neural networks (CNNs) makes it feasible to perform black-box attacks, resulting in security threats for CNNs. However, fewer endeavors have been made to investigate transferable attacks for vision transformers (ViTs), which achieve superior performance on various computer vision tasks. Unlike CNNs, ViTs establish relationships between patches extracted from inputs by the self-attention module. Thus, adversarial examples crafted on CNNs might hardly attack ViTs. To assess the security of ViTs comprehensively, we investigate the transferability across different ViTs in both untargetd and targeted scenarios. More specifically, we propose a Pay No Attention (PNA) attack, which ignores attention gradients during backpropagation to improve the linearity of backpropagation. Additionally, we introduce a PatchOut/CubeOut attack for image/video ViTs. They optimize perturbations within a randomly selected subset of patches/cubes during each iteration, preventing over-fitting to the white-box surrogate ViT model. Furthermore, we maximize the L2 norm of perturbations, ensuring that the generated adversarial examples deviate significantly from the benign ones. These strategies are designed to be harmoniously compatible. Combining them can enhance transferability by jointly considering patch-based inputs and the self-attention of ViTs. Moreover, the proposed combined attack seamlessly integrates with existing transferable attacks, providing an additional boost to transferability. We conduct experiments on ImageNet and Kinetics-400 for image and video ViTs, respectively. Experimental results demonstrate the effectiveness of the proposed method.

Blood Pressure Management After Endovascular Therapy for Acute Ischemic Stroke: The BEST-II Randomized Clinical Trial.

Importance: The effects of moderate systolic blood pressure (SBP) lowering after successful recanalization with endovascular therapy for acute ischemic stroke are uncertain. Objective: To determine the futility of lower SBP targets after endovascular therapy (<140 mm Hg or 160 mm Hg) compared with a higher target (≤180 mm Hg). Design, Setting, and Participants: Randomized, open-label, blinded end point, phase 2, futility clinical trial that enrolled 120 patients with acute ischemic stroke who had undergone successful endovascular therapy at 3 US comprehensive stroke centers from January 2020 to March 2022 (final follow-up, June 2022). Intervention: After undergoing endovascular therapy, participants were randomized to 1 of 3 SBP targets: 40 to less than 140 mm Hg, 40 to less than 160 mm Hg, and 40 to 180 mm Hg or less (guideline recommended) group, initiated within 60 minutes of recanalization and maintained for 24 hours. Main Outcomes and Measures: Prespecified multiple primary outcomes for the primary futility analysis were follow-up infarct volume measured at 36 (±12) hours and utility-weighted modified Rankin Scale (mRS) score (range, 0 [worst] to 1 [best]) at 90 (±14) days. Linear regression models were used to test the harm-futility boundaries of a 10-mL increase (slope of 0.5) in the follow-up infarct volume or a 0.10 decrease (slope of -0.005) in the utility-weighted mRS score with each 20-mm Hg SBP target reduction after endovascular therapy (1-sided α = .05). Additional prespecified futility criterion was a less than 25% predicted probability of success for a future 2-group, superiority trial comparing SBP targets of the low- and mid-thresholds with the high-threshold (maximum sample size, 1500 with respect to the utility-weighted mRS score outcome). Results: Among 120 patients randomized (mean [SD] age, 69.6 [14.5] years; 69 females [58%]), 113 (94.2%) completed the trial. The mean follow-up infarct volume was 32.4 mL (95% CI, 18.0 to 46.7 mL) for the less than 140-mm Hg group, 50.7 mL (95% CI, 33.7 to 67.7 mL), for the less than 160-mm Hg group, and 46.4 mL (95% CI, 24.5 to 68.2 mL) for the 180-mm Hg or less group. The mean utility-weighted mRS score was 0.51 (95% CI, 0.38 to 0.63) for the less than 140-mm Hg group, 0.47 (95% CI, 0.35 to 0.60) for the less than 160-mm Hg group, and 0.58 (95% CI, 0.46 to 0.71) for the high-target group. The slope of the follow-up infarct volume for each mm Hg decrease in the SBP target, adjusted for the baseline Alberta Stroke Program Early CT score, was -0.29 (95% CI, -0.81 to ∞; futility P = .99). The slope of the utility-weighted mRS score for each mm Hg decrease in the SBP target after endovascular therapy, adjusted for baseline utility-weighted mRS score, was -0.0019 (95% CI, -∞ to 0.0017; futility P = .93). Comparing the high-target SBP group with the lower-target groups, the predicted probability of success for a future trial was 25% for the less than 140-mm Hg group and 14% for the 160-mm Hg group. Conclusions and Relevance: Among patients with acute ischemic stroke, lower SBP targets less than either 140 mm Hg or 160 mm Hg after successful endovascular therapy did not meet prespecified criteria for futility compared with an SBP target of 180 mm Hg or less. However, the findings suggested a low probability of benefit from lower SBP targets after endovascular therapy if tested in a future larger trial. Trial Registration: ClinicalTrials.gov Identifier: NCT04116112.

R1ρ dispersion in white matter correlates with quantitative metrics of cognitive impairment.

Much previous neuroimaging research in Alzheimer's disease has focused on the roles of amyloid and tau proteins, but recent studies have implicated microvascular changes in white matter as early indicators of damage related to later dementia. We used MRI to derive novel, non-invasive measurements of R1ρ dispersion using different locking fields to characterize variations of microvascular structure and integrity in brain tissues. We developed a non-invasive 3D R1ρ dispersion imaging technique using different locking fields at 3T. We acquired MR images and cognitive assessments of participants with mild cognitive impairment (MCI) and compared them to age-matched healthy controls in a cross-sectional study. After providing informed consent, 40 adults aged 62 to 82 years (n = 17 MCI) were included in this study. White matter ΔR1ρ-fraction measured by R1ρ dispersion imaging showed a strong correlation with the cognitive status of older adults (ßstd = -0.4, p-value < 0.01) independent of age, in contrast to other conventional MRI markers such as T2, R1ρ, and white matter hyperintense lesion volume (WMHs) measured with T2-FLAIR. The correlation of WMHs with cognitive status was no longer significant after adjusting for age and sex in linear regression analysis, and the size of the regression coefficient was substantially decreased (53% lower). This work establishes a new non-invasive method that potentially characterizes impairment of the microvascular structure of white matter in MCI patients compared to healthy controls. The application of this method in longitudinal studies would improve our fundamental understanding of the pathophysiologic changes that accompany abnormal cognitive decline with aging and help identify potential targets for treatment of Alzheimer's disease.

Presurgical Magnetic Resonance Imaging Indicators of Revascularization Response in Adults With Moyamoya Vasculopathy.

BACKGROUND: Moyamoya is a progressive intracranial vasculopathy, primarily affecting distal segments of the internal carotid and middle cerebral arteries. Treatment may comprise angiogenesis-inducing surgical revascularization; however, lack of randomized trials often results in subjective treatment decisions. HYPOTHESIS: Compensatory presurgical posterior vertebrobasilar artery (VBA) flow-territory reactivity, including greater cerebrovascular reactivity (CVR) and reduced vascular delay time, portends greater neoangiogenic response verified on digital subtraction angiography (DSA) at 1-year follow-up. STUDY TYPE: Prospective intervention cohort. SUBJECTS: Thirty-one patients with moyamoya (26 females; age = 45 ± 13 years; 41 revascularized hemispheres). METHODS: Anatomical MRI, hypercapnic CVR MRI, and DSA acquired presurgically in adult moyamoya participants scheduled for clinically indicated surgical revascularization. One-year postsurgery, DSA was repeated to evaluate collateralization. FIELD STRENGTH: 3 T. SEQUENCE: Hypercapnic T 2 * -weighted gradient-echo blood-oxygenation-level-dependent, T2 -weighted turbo-spin-echo fluid-attenuated-inversion-recovery, T1 -weighted magnetization-prepared-rapid-gradient-echo, and T2 -weighted diffusion-weighted-imaging. ASSESSMENT: Presurgical maximum CVR and response times were evaluated in VBA flow-territories. Revascularization success was determined using an ordinal scoring system of neoangiogenic collateralization from postsurgical DSA by two cerebrovascular neurosurgeons (R.V.C. with 8 years of experience; M.R.F. with 9 years of experience) and one neuroradiologist (L.T.D. with 8 years of experience). Stroke risk factors (age, sex, race, vasculopathy, and diabetes) were recorded. STATISTICAL TESTS: Fisher's exact and Wilcoxon rank-sum tests were applied to compare presurgical variables between cohorts with angiographically confirmed good (>1/3 middle cerebral artery [MCA] territory revascularized) vs. poor (<1/3 MCA territory revascularized) outcomes. SIGNIFICANCE: two-sided P < 0.05. Normalized odds ratios (ORs) were calculated. RESULTS: Criteria for good collateralization were met in 25 of the 41 revascularized hemispheres. Presurgical normalized VBA flow-territory CVR was significantly higher in those with good (1.12 ± 0.13 unitless) vs. poor (1.04 ± 0.05 unitless) outcomes. Younger (OR = -0.60 ± 0.67) and White (OR = -1.81 ± 1.40) participants had highest revascularization success (good outcomes: age = 42 ± 14 years, race = 84% White; poor outcomes: age = 49 ± 11 years, race = 44% White). DATA CONCLUSION: Presurgical MRI-measures of VBA flow-territory CVR are highest in moyamoya participants with better angiographic responses to surgical revascularization. LEVEL OF EVIDENCE: 1 TECHNICAL EFFICACY STAGE: 4.

A Dynamic Frame Selection Framework for Fast Video Recognition.

We introduce AdaFrame, a conditional computation framework that adaptively selects relevant frames on a per-input basis for fast video recognition. AdaFrame, which contains a Long Short-Term Memory augmented with a global memory to provide context information, operates as an agent to interact with video sequences aiming to search over time which frames to use. Trained with policy search methods, at each time step, AdaFrame computes a prediction, decides where to observe next, and estimates a utility, i.e., expected future rewards, of viewing more frames in the future. Exploring predicted utilities at testing time, AdaFrame is able to achieve adaptive lookahead inference so as to minimize the overall computational cost without incurring a degradation in accuracy. We conduct extensive experiments on two large-scale video benchmarks, FCVID and ActivityNet. With a vanilla ResNet-101 model, AdaFrame achieves similar performance of using all frames while only requiring, on average, 8.21 and 8.65 frames on FCVID and ActivityNet, respectively. We also demonstrate AdaFrame is compatible with modern 2D and 3D networks for video recognition. Furthermore, we show, among other things, learned frame usage can reflect the difficulty of making prediction decisions both at instance-level within the same class and at class-level among different categories.

Scale Normalized Image Pyramids With AutoFocus for Object Detection.

We present an efficient foveal framework to perform object detection. A scale normalized image pyramid (SNIP) is generated that, like human vision, only attends to objects within a fixed size range at different scales. Such a restriction of objects' size during training affords better learning of object-sensitive filters, and therefore, results in better accuracy. However, the use of an image pyramid increases the computational cost. Hence, we propose an efficient spatial sub-sampling scheme which only operates on fixed-size sub-regions likely to contain objects (as object locations are known during training). The resulting approach, referred to as Scale Normalized Image Pyramid with Efficient Resampling or SNIPER, yields up to 3× speed-up during training. Unfortunately, as object locations are unknown during inference, the entire image pyramid still needs processing. To this end, we adopt a coarse-to-fine approach, and predict the locations and extent of object-like regions which will be processed in successive scales of the image pyramid. Intuitively, it's akin to our active human-vision that first skims over the field-of-view to spot interesting regions for further processing and only recognizes objects at the right resolution. The resulting algorithm is referred to as AutoFocus and results in a 2.5-5× speed-up during inference when used with SNIP. Code: https://github.com/mahyarnajibi/SNIPER.

Comparison of convolutional neural networks for detecting large vessel occlusion on computed tomography angiography.

PURPOSE: Artificial intelligence diagnosis and triage of large vessel occlusion may quicken clinical response for a subset of time-sensitive acute ischemic stroke patients, improving outcomes. Differences in architectural elements within data-driven convolutional neural network (CNN) models impact performance. Foreknowledge of effective model architectural elements for domain-specific problems can narrow the search for candidate models and inform strategic model design and adaptation to optimize performance on available data. Here, we study CNN architectures with a range of learnable parameters and which span the inclusion of architectural elements, such as parallel processing branches and residual connections with varying methods of recombining residual information. METHODS: We compare five CNNs: ResNet-50, DenseNet-121, EfficientNet-B0, PhiNet, and an Inception module-based network, on a computed tomography angiography large vessel occlusion detection task. The models were trained and preliminarily evaluated with 10-fold cross-validation on preprocessed scans (n = 240). An ablation study was performed on PhiNet due to superior cross-validated test performance across accuracy, precision, recall, specificity, and F1 score. The final evaluation of all models was performed on a withheld external validation set (n = 60) and these predictions were subsequently calibrated with sigmoid curves. RESULTS: Uncalibrated results on the withheld external validation set show that DenseNet-121 had the best average performance on accuracy, precision, recall, specificity, and F1 score. After calibration DenseNet-121 maintained superior performance on all metrics except recall. CONCLUSIONS: The number of learnable parameters in our five models and best-ablated PhiNet directly related to cross-validated test performance-the smaller the model the better. However, this pattern did not hold when looking at generalization on the withheld external validation set. DenseNet-121 generalized the best; we posit this was due to its heavy use of residual connections utilizing concatenation, which causes feature maps from earlier layers to be used deeper in the network, while aiding in gradient flow and regularization.

A Prospective, Longitudinal Magnetic Resonance Imaging Evaluation of Cerebrovascular Reactivity and Infarct Development in Patients With Intracranial Stenosis.

BACKGROUND: Patients with symptomatic atherosclerotic and non-atherosclerotic (i.e., moyamoya) intracranial steno-occlusive disease experience high 2-year infarct rates. PURPOSE: To investigate whether cerebral blood flow (CBF) and cerebrovascular reactivity (CVR) measures may provide biomarkers of 1-to-2-year infarct risk. STUDY TYPE: Prospective, longitudinal study. SUBJECTS: Adult participants (age = 18-85 years) with symptomatic intracranial atherosclerotic disease (N = 26) or non-atherosclerotic (i.e., moyamoya; N = 43) and stenosis ≥50% of a major intracranial artery were initially scanned within 45 days of stroke. Follow-up imaging (target  = 1.5 years) was acquired for new infarct assessment. FIELD STRENGTH/SEQUENCE: 3.0 Tesla with normocapnic arterial spin labeling (ASL) and blood oxygenation level-dependent (BOLD) imaging acquired during an interleaved hypercapnic (3 minutes) and normocapnic (3 minutes) respiratory stimulus. ASSESSMENT: CBF, maximum CVR, and time-to-maximum CVR (i.e., CVRDELAY ) were calculated. Laterality indices (difference between infarcted and contralesional hemispheres divided by sum of absolute values) of metrics at enrollment were contrasted between participants with vs. without new infarcts on follow-up. STATISTICAL TESTS: Laterality indices were compared using non-parametric Wilcoxon tests (significance: two-sided P < 0.05) and effect sizes as Cohen's d. Continuous variables are presented as mean ± SD. RESULTS: New infarcts were observed on follow-up in 15.0% of participants. The laterality index of the CVRDELAY was elevated (P = 0.01) in participants with atherosclerosis with new infarcts (index = 0.13) compared to participants without new infarcts (index = 0.05). DATA CONCLUSION: Elevated CVRDELAY may indicate brain parenchyma at increased risk for new infarcts in patients with symptomatic intracranial atherosclerotic disease treated with standard-of-care medical management. LEVEL OF EVIDENCE: 2 TECHNICAL EFFICACY STAGE: 3.

Reduced oxygen extraction efficiency in sickle cell anemia patients with evidence of cerebral capillary shunting.

Arterial spin labeling (ASL) magnetic resonance imaging (MRI) utilizes arterial blood water as an endogenous contrast agent to provide a quantitative measure of cerebral blood flow (CBF). Recently, hyperintense signal within dural venous sinuses in ASL images of sickle cell anemia (SCA) patients has been shown to be consistent with elevated flow velocities and may indicate capillary shunting and reduced oxygen extraction. Here, we performed oxygen extraction fraction (OEF) and CBF measurements in adults (cumulative n = 114) with (n = 69) and without (n = 45) SCA to test the hypothesis that hyperintense venous ASL signal is associated with reduced OEF. Higher categorical scores of shunting on ASL MRI were associated with lower OEF in participants with silent cerebral infarcts or white matter hyperintensities (p = 0.003), but not in those without lesions (p = 0.551). These findings indicate that venous hyperintense signal in ASL images in SCA patients may represent a marker of capillary-level disturbances in oxygen exchange efficiency and small vessel pathology.

A cross-sectional, case-control study of intracranial arterial wall thickness and complete blood count measures in sickle cell disease.

In sickle cell disease (SCD), cerebral oxygen delivery is dependent on the cerebral vasculature's ability to increase blood flow and volume through relaxation of the smooth muscle that lines intracranial arteries. We hypothesised that anaemia extent and/or circulating markers of inflammation lead to concentric macrovascular arterial wall thickening, visible on intracranial vessel wall magnetic resonance imaging (VW-MRI). Adult and pediatric SCD (n = 69; age = 19.9 ± 8.6 years) participants and age- and sex-matched control participants (n = 38; age = 22.2 ± 8.9 years) underwent 3-Tesla VW-MRI; two raters measured basilar and bilateral supraclinoid internal carotid artery (ICA) wall thickness independently. Mean wall thickness was compared with demographic, cerebrovascular and haematological variables. Mean vessel wall thickness was elevated (P < 0·001) in SCD (1·07 ± 0·19 mm) compared to controls (0·97 ± 0·07 mm) after controlling for age and sex. Vessel wall thickness was higher in participants on chronic transfusions (P = 0·013). No significant relationship between vessel wall thickness and flow velocity, haematocrit, white blood cell count or platelet count was observed; however, trends (P < 0·10) for wall thickness increasing with decreasing haematocrit and increasing white blood cell count were noted. Findings are discussed in the context of how anaemia and circulating inflammatory markers may impact arterial wall morphology.

Glioblastoma Distance From the Subventricular Neural Stem Cell Niche Does Not Correlate With Survival.

OBJECTIVE: To determine the relationship between survival and glioblastoma distance from the ventricular-subventricular neural stem cell niche (VSVZ). METHODS: 502 pre-operative gadolinium-enhanced, T1-weighted MRIs with glioblastoma retrieved from an institutional dataset (n = 252) and The Cancer Imaging Atlas (n=250) were independently reviewed. The shortest distance from the tumor contrast enhancement to the nearest lateral ventricular wall, the location of the VSVZ, was measured (GBM-VSVZDist). The relationship of GBM-VSVZDist with the proportion of glioblastomas at each distance point and overall survival was explored with a Pearson's correlation and Cox regression model, respectively, adjusting for the well-established glioblastoma prognosticators. RESULTS: 244/502 glioblastomas had VSVZ contact. The proportion of non-VSVZ-contacting glioblastomas correlated inversely with GBM-VSVZDist (partial Pearson's correlation adjusted for tumor volume R=-0.79, p=7.11x10-7). A fit of the Cox regression model adjusted for age at diagnosis, Karnofsky performance status score, post-operative treatment with temozolomide and/or radiotherapy, IDH1/2 mutation status, MGMT promoter methylation status, tumor volume, and extent of resection demonstrated a significantly decreased overall survival only when glioblastoma contacted the VSVZ. Overall survival did not correlate with GBM-VSVZDist. CONCLUSIONS: In the two independent cohorts analyzed, glioblastomas at diagnosis were found in close proximity or in contact with the VSVZ with a proportion that decreased linearly with GBM-VSVZDist. Patient survival was only influenced by the presence or absence of a gadolinium-enhanced glioblastoma contact with the VSVZ. These results may guide analyses to test differential effectiveness of VSVZ radiation in VSVZ-contacting and non-contacting glioblastomas and/or inform patient selection criteria in clinical trials of glioblastoma radiation.

A Generic Improvement to Deep Residual Networks Based on Gradient Flow.

Preactivation ResNets consistently outperforms the original postactivation ResNets on the CIFAR10/100 classification benchmark. However, these results surprisingly do not carry over to the standard ImageNet benchmark. First, we theoretically analyze this incongruity in terms of how the two variants differ in handling the propagation of gradients. Although identity shortcuts are critical in both variants for improving optimization and performance, we show that postactivation variants enable early layers to receive a diverse dynamic composition of gradients from effectively deeper paths in comparison to preactivation variants, enabling the network to make maximal use of its representational capacity. Second, we show that downsampling projections (while only a few in number) have a significantly detrimental effect on performance. We show that by simply replacing downsampling projections with identitylike dense-reshape shortcuts, the classification results of standard residual architectures such as ResNets, ResNeXts, and SE-Nets improve by up to 1.2% on ImageNet, without any increase in computational complexity (FLOPs).

Classifying intracranial stenosis disease severity from functional MRI data using machine learning.

Translation of many non-invasive hemodynamic MRI methods to cerebrovascular disease patients has been hampered by well-known artifacts associated with delayed blood arrival times and reduced microvascular compliance. Using machine learning and support vector machine (SVM) algorithms, we investigated whether arrival time-related artifacts in these methods could be exploited as novel contrast sources to discriminate angiographically confirmed stenotic flow territories. Intracranial steno-occlusive moyamoya patients (n = 53; age = 45 ± 14.2 years; sex = 43 F) underwent (i) catheter angiography, (ii) anatomical MRI, (iii) cerebral blood flow (CBF)-weighted arterial spin labeling, and (iv) cerebrovascular reactivity (CVR)-weighted hypercapnic blood-oxygenation-level-dependent MRI. Mean, standard deviation (std), and 99th percentile of CBF, CVR, CVRDelay, and CVRMax were calculated in major anterior and posterior flow territories perfused by vessels with vs. without stenosis (≥70%) confirmed by catheter angiography. These and demographic variables were input into SVMs to evaluate discriminatory capacity for stenotic flow territories using k-fold cross-validation and receiver-operating-characteristic-area-under-the-curve to quantify variable combination relevance. Anterior circulation CBF-std, attributable to heterogeneous endovascular signal and prolonged arterial transit times, was the best performing single variable and CVRDelay-mean and CBF-std, both reflective of delayed vascular compliance, were a high-performing two-variable combination (specificity = 0.67; sensitivity = 0.75). Findings highlight the relevance of hemodynamic imaging and machine learning for identifying cerebrovascular impairment.

Current Volumetric Models Overestimate Vestibular Schwannoma Size Following Stereotactic Radiosurgery.

OBJECTIVE: Accurate volume assessment is essential for the management of vestibular schwannoma after stereotactic radiosurgery (SRS). A cuboidal approximation for volume is the standard surveillance method; however, this may overestimate tumor volume. We sought to evaluate several volumetric models and their suitability for post-SRS surveillance. STUDY DESIGN: Retrospective cohort study. SETTING: Tertiary referral center. PATIENTS: We evaluated 54 patients with vestibular schwannoma before and after SRS. INTERVENTION(S): Gold-standard volumes were obtained by a radiation oncologist using contouring software. Volume was also calculated by cuboidal, ellipsoidal, and spherical formulae using tumor diameters obtained by a neuroradiologist. MAIN OUTCOME MEASURE(S): Percent error (PE) and absolute percent error (APE) were calculated. Paired t test evaluated bias, and the Bland-Altman method evaluated reproducibility. Linear regression evaluated predictors of model error. RESULTS: All models overestimated volume compared with the gold standard. The cuboidal model was not reproducible before SRS (p < 0.001), and no model was reproducible after SRS (cuboidal p < 0.001; ellipsoidal p = 0.02; spherical p = 0.02). Significant bias was present before SRS for the cuboidal model (p < 0.001), and post-SRS for all models [cuboidal (p < 0.001), ellipsoidal (p < 0.02), and spherical (p = 0.005)]. Model error was negatively associated with pretreatment volume for the cuboidal (PE p = 0.03; APE p = 0.03), ellipsoidal (PE p = 0.03; APE p = 0.04), and spherical (PE p = 0.02; APE p = 0.03) methods and lost linearity post-SRS. CONCLUSIONS: The standard cuboidal practice for following vestibular schwannoma tumor volume after SRS overestimates size. Ellipsoidal and spherical estimations have improved performance but also overestimate volume and lack reliability post-SRS. The development of other volumetric models or application of contouring software should be investigated.

The CAFA challenge reports improved protein function prediction and new functional annotations for hundreds of genes through experimental screens.

BACKGROUND: The Critical Assessment of Functional Annotation (CAFA) is an ongoing, global, community-driven effort to evaluate and improve the computational annotation of protein function. RESULTS: Here, we report on the results of the third CAFA challenge, CAFA3, that featured an expanded analysis over the previous CAFA rounds, both in terms of volume of data analyzed and the types of analysis performed. In a novel and major new development, computational predictions and assessment goals drove some of the experimental assays, resulting in new functional annotations for more than 1000 genes. Specifically, we performed experimental whole-genome mutation screening in Candida albicans and Pseudomonas aureginosa genomes, which provided us with genome-wide experimental data for genes associated with biofilm formation and motility. We further performed targeted assays on selected genes in Drosophila melanogaster, which we suspected of being involved in long-term memory. CONCLUSION: We conclude that while predictions of the molecular function and biological process annotations have slightly improved over time, those of the cellular component have not. Term-centric prediction of experimental annotations remains equally challenging; although the performance of the top methods is significantly better than the expectations set by baseline methods in C. albicans and D. melanogaster, it leaves considerable room and need for improvement. Finally, we report that the CAFA community now involves a broad range of participants with expertise in bioinformatics, biological experimentation, biocuration, and bio-ontologies, working together to improve functional annotation, computational function prediction, and our ability to manage big data in the era of large experimental screens.

Angiocentric glioma mimicking encephalomalacia.

Angiocentric glioma is a rare low-grade neoplasm of the central nervous system which typically presents with medication-refractory seizures in children and young adults. On magnetic resonance imaging, angiocentric glioma is classically T1 hypointense and T2/FLAIR hyperintense. We present the case of a 40-year-old male who had been followed by our institution for 17 years for management of epilepsy. Initial and repeat brain imaging showed an apparent region of cystic encephalomalacia in the right frontal lobe. In an attempt to control his seizures, the lesion was resected. Grossly, the cut surface of the specimen was characterized by multiple small cystic spaces. Microscopically, the lesion was composed of an infiltrative population of glial cells variably arranged in perivascular "pseudorosettes," nodules, and subpial "palisades." The final diagnosis was angiocentric glioma. This is the second reported case of an angiocentric glioma mistaken for encephalomalacia.