Dataset Information

Automated grading of enlarged perivascular spaces in clinical imaging data of an acute stroke cohort using an interpretable, 3D deep learning framework.

ABSTRACT: Enlarged perivascular spaces (EPVS), specifically in stroke patients, has been shown to strongly correlate with other measures of small vessel disease and cognitive impairment at 1 year follow-up. Typical grading of EPVS is often challenging and time consuming and is usually based on a subjective visual rating scale. The purpose of the current study was to develop an interpretable, 3D neural network for grading enlarged perivascular spaces (EPVS) severity at the level of the basal ganglia using clinical-grade imaging in a heterogenous acute stroke cohort, in the context of total cerebral small vessel disease (CSVD) burden. T2-weighted images from a retrospective cohort of 262 acute stroke patients, collected in 2015 from 5 regional medical centers, were used for analyses. Patients were given a label of 0 for none-to-mild EPVS (< 10) and 1 for moderate-to-severe EPVS (≥ 10). A three-dimensional residual network of 152 layers (3D-ResNet-152) was created to predict EPVS severity and 3D gradient class activation mapping (3DGradCAM) was used for visual interpretation of results. Our model achieved an accuracy 0.897 and area-under-the-curve of 0.879 on a hold-out test set of 15% of the total cohort (n = 39). 3DGradCAM showed areas of focus that were in physiologically valid locations, including other prevalent areas for EPVS. These maps also suggested that distribution of class activation values is indicative of the confidence in the model's decision. Potential clinical implications of our results include: (1) support for feasibility of automated of EPVS scoring using clinical-grade neuroimaging data, potentially alleviating rater subjectivity and improving confidence of visual rating scales, and (2) demonstration that explainable models are critical for clinical translation.

SUBMITTER: Williamson BJ

PROVIDER: S-EPMC8764081 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:Purpose To describe a fully automated segmentation method that yields object-based morphologic estimates of enlarged perivascular spaces (ePVSs) in clinical-field-strength (3.0-T) magnetic resonance (MR) imaging data. Materials and Methods In this HIPAA-compliant study, MR imaging data were obtained with a 3.0-T MR imager in research participants without dementia (mean age, 85.3 years; range, 70.4-101.2 years) who had given written informed consent. This method is built on (a) relative normalized white matter, ventricular and cortical signal intensities within T1-weighted, fluid-attenuated inversion recovery, T2-weighted, and proton density data and (b) morphologic (width, volume, linearity) characterization of each resultant cluster. Visual rating was performed by three raters, including one neuroradiologist, after established single-section guidelines. Correlations between visual counts and automated counts, as well session-to-session correlation of counts within each participant, were assessed with the Pearson correlation coefficient r. Results There was a significant correlation between counts by visual raters and automated detection of ePVSs in the same section (r = 0.65, P < .001; r = 0.69, P < .001; and r = 0.54, P < .01 for raters 1, 2, and 3, respectively). With regard to visual ratings and whole-brain count consistency, average visual rating scores were highly correlated with automated detection of total burden volume (r = 0.58, P < .01) and total ePVS number (r = 0.76, P < .01). Morphology of clusters across 28 data sets was consistent with published radiographic estimates of ePVS; mean width of clusters segmented was 3.12 mm (range, 1.7-13.5 mm). Conclusion This MR imaging-based method for multimodal autoidentification of perivascular spaces yields individual whole-brain morphologic characterization of ePVS in clinical MR imaging data and is an important tool in the detailed assessment of these features. © RSNA, 2017 Online supplemental material is available for this article.

Project description:Background and aimsEnlarged perivascular spaces (also known as Virchow-Robin spaces) on T2-weighted brain magnetic resonance imaging are common, but their etiology, and specificity to small vessel as opposed to general cerebrovascular disease or ageing, is unclear. We tested the association between enlarged perivascular spaces and ischemic stroke subtype, other markers of small vessel disease, and common vascular risk factors.MethodsWe prospectively recruited patients with acute stroke, diagnosed and subtyped by a stroke physician using clinical features and brain magnetic resonance imaging. A neuroradiologist rated basal ganglia and centrum semiovale enlarged perivascular spaces on a five-point scale, white matter lesions, recent and old infarcts, and cerebral atrophy. We assessed associations between basal ganglia-, centrum semiovale- and total (combined basal ganglia and centrum semiovale) enlarged perivascular spaces, stroke subtype, white matter lesions, atrophy, and vascular risk factors.ResultsAmong 298 patients (mean age 68 years), after adjusting for vascular risk factors and white matter lesions, basal ganglia-enlarged perivascular spaces were associated with increasing age (P = 0.001), centrum semiovale-enlarged perivascular spaces (P < 0.001), cerebral atrophy (P = 0.03), and lacunar stroke subtype (P = 0.04). Centrum semiovale-enlarged perivascular spaces were associated mainly with basal ganglia-enlarged perivascular spaces. Total enlarged perivascular spaces were associated with increasing age (P = 0.01), deep white matter lesions (P = 0.005), and previous stroke (P = 0.006).ConclusionsEnlarged perivascular spaces are associated with age, lacunar stroke subtype and white matter lesions and should be considered as another magnetic resonance imaging marker of cerebral small vessel disease. Further evaluation of enlarged perivascular spaces in studies of ageing, stroke, and dementia is needed to determine their pathophysiological importance.

Project description:Enlarged perivascular spaces (EPVS) are widely considered as a feature of cerebral small vessel diseases (SVD), but its underlying pathology is still under active investigation. The aim of this study was to explore the association between hemoglobin level and the severity of EPVS. Consecutive patients with acute ischemic stroke who underwent baseline MRI scan and hemoglobin testing were evaluated. EPVS in basal ganglia (BG) and central semiovale (CS) were rated with a validated 4-point semiquantitative scale (0 = none; 1 = 1-10; 2 = 11-20; 3 = 21-40; and 4 ≥ 40). Bivariate logistic regression models were used to identify the associations of hemoglobin with predefined high-degree (score > 1) CS-EPVS and BG-EPVS. Multinomial logistic regression models were used to analyze the associations between hemoglobin and CS-/BG-EPVS predominance patterns. A total of 401 patients were included in the final analysis, 94 patients (23.4%) had a high degree of CS-EPVS and 45 patients (11.2%) had a high degree of BG-EPVS. Compared with tertile 1 of hemoglobin, tertile 3 of hemoglobin was independently associated with high degree of CS-EPVS after adjusting for other features of SVD (odds ratio [OR] 2.399, 95% confidence interval [CI] 1.315-4.379, P = 0.004) and potential confounding factors (OR 2.611, 95% CI 1.346-5.066, P = 0.005). In multinomial logistic regression models, compared with tertile 1 of hemoglobin, tertile 2 (OR 2.463, 95% CI 1.195-5.075, P = 0.015) and tertile 3 (OR 2.625, 95% CI 1.102-6.251, P = 0.029) of hemoglobin were associated with higher odds of BG-EPVS = CS-EPVS pattern, and tertile 3 of hemoglobin (OR 2.576, 95% CI 1.004-6.608, P = 0.049) was associated with higher odds of BG-EPVS < CS-EPVS pattern. Elevated hemoglobin level was independently associated with high degree of CS-EPVS and higher odds of CS-EPVS predominance pattern, but not with BG-EPVS, which support that the topography of EPVS is characteristic. However, the pathogenesis linking hemoglobin and CS-EPVS is unclear and still needs further investigation.

Dataset Information

Automated grading of enlarged perivascular spaces in clinical imaging data of an acute stroke cohort using an interpretable, 3D deep learning framework.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets