Yutong Chen
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Yutong Chen has a total of 40 co-authors. Here are the top co-authors by number of shared publications:
- Hugh Stephen Markus
From the Stroke Research Group, Department of Clinical Neurosciences, University of Cambridge, United Kingdom
4 shared publications - Christopher D. Anderson
Department of Neurology (C.D.A.), Brigham and Women’s Hospital, Harvard Medical School, Boston, MA
3 shared publications - Ernst Mayerhofer
Center for Genomic Medicine, Massachusetts General Hospital, Boston, MA, USA
3 shared publications - Andreas Harloff
Department of Neurology and Neurophysiology, University Medical Center Freiburg, Faculty of Medicine University of Freiburg Germany
2 shared publications - Cyprien A. Rivier
Yale University School of Medicine
2 shared publications Daniel J Tozer
Department of Clinical Neurosciences, University of Cambridge , Cambridge CB2 0QQ
2 shared publications
Latest Publications
- Open Access
- 14 August 2026
- Published by SAGE Publications in International Journal of Stroke
- Vol. 21 (9) , 1260-1273
The role of glymphatic dysfunction in cerebral small vessel disease and vascular cognitive impairment, and therapeutic implications
Background: The glymphatic system is a brain-wide perivascular clearance pathway that facilitates cerebrospinal fluid–interstitial fluid (CSF-ISF) exchange and the removal of metabolic waste from the brain parenchyma. Increasing evidence indicates that glymphatic transport is critically dependent on the integrity and function of cerebral small vessels, suggesting a close link with cerebral small vessel disease (cSVD). Structural and functional alterations of small vessels may disrupt glymphatic flow, while impaired glymphatic clearance may in turn exacerbate small vessel–related brain injury, indicating a bidirectional interaction between these two processes. Methods: In this review, we synthesize current evidence on the relationship between cSVD and glymphatic dysfunction, integrating findings from experimental models and in vivo human studies. We discuss shared pathophysiological mechanisms linking small vessel pathology to impaired perivascular fluid transport, summarize associations between glymphatic dysfunction and established cSVD imaging markers, and evaluate emerging therapeutic strategies aimed at restoring or enhancing glymphatic function. We highlight areas of convergence and inconsistency in the existing literature and identify key knowledge gaps that warrant further investigation. We also review potential therapeutic approaches, both behavioral and pharmacological to improve glymphatic function. Conclusion: Current evidence supports a close but incompletely understood bidirectional relationship between cSVD and glymphatic dysfunction. Targeting glymphatic function may represent a possible therapeutic strategy for cSVD and vascular cognitive impairment.
- 14 August 2026
- Published by SAGE Publications in International Journal of Stroke
- Vol. 21 (9) , 1260-1273
Open AccessThe role of glymphatic dysfunction in cerebral small vessel disease and vascular cognitive impairment, and therapeutic implications
Background: The glymphatic system is a brain-wide perivascular clearance pathway that facilitates cerebrospinal fluid–interstitial fluid (CSF-ISF) exchange and the removal of metabolic waste from the brain parenchyma. Increasing evidence indicates that glymphatic transport is critically dependent on the integrity and function of cerebral small vessels, suggesting a close link with cerebral small vessel disease (cSVD). Structural and functional alterations of small vessels may disrupt glymphatic flow, while impaired glymphatic clearance may in turn exacerbate small vessel–related brain injury, indicating a bidirectional interaction between these two processes. Methods: In this review, we synthesize current evidence on the relationship between cSVD and glymphatic dysfunction, integrating findings from experimental models and in vivo human studies. We discuss shared pathophysiological mechanisms linking small vessel pathology to impaired perivascular fluid transport, summarize associations between glymphatic dysfunction and established cSVD imaging markers, and evaluate emerging therapeutic strategies aimed at restoring or enhancing glymphatic function. We highlight areas of convergence and inconsistency in the existing literature and identify key knowledge gaps that warrant further investigation. We also review potential therapeutic approaches, both behavioral and pharmacological to improve glymphatic function. Conclusion: Current evidence supports a close but incompletely understood bidirectional relationship between cSVD and glymphatic dysfunction. Targeting glymphatic function may represent a possible therapeutic strategy for cSVD and vascular cognitive impairment.
- Open Access
- 3 June 2026
- Published by SAGE Publications in International Journal of Stroke
- Vol. 21 (9) , 1348-1357
Two distinct patterns of progression of sporadic cerebral small vessel disease
Background: Sporadic cerebral small vessel disease (cSVD) is the most common cause of vascular dementia. cSVD features a wide range of neuroimaging markers, including white matter hyperintensity (WMH), lacunes, perivascular spaces, microbleeds, and diffusion magnetic resonance imaging (MRI)-based markers of white matter (WM) injury, and is accompanied by progressive cognitive decline. However, few studies have systematically investigated how and when various imaging and cognitive markers change along the complete cSVD progression trajectory. Aims: To establish changes in imaging and cognitive markers along the cSVD progression trajectory. Methods: In this cohort study, we selected 496 participants from the Radboud University Nijmegen Diffusion Tensor and Magnetic Resonance Imaging Cohort (RUNDMC) cohort as the discovery data set. The subtype and stage inference model was used to identify cSVD subtypes and their progression patterns. Five neuroimaging markers (WMH volume, median mean diffusivity (MD) in WM, WM volume, gray matter (GM) volume, and ventricle volume) were used in trajectory construction. The validation cohort was Cambridge cSVD (CamcSVD; n = 330). Results: We identified two subtypes in RUNDMC and replicated this finding in CamcSVD. One subtype displayed early WMH increase, followed by increased WM MD and tissue volume loss. The other subtype, which we termed “atrophy predominant,” showed early loss of GM and WM and ventricle enlargement, followed by WMH increase and increased WM MD. This subtype displayed an increase in markers of idiopathic normal pressure hydrocephalus (iNPH). Conclusion: We have delineated two distinct progression patterns in sporadic cSVD. Further studies are needed to investigate the relationship between the atrophy-predominant cSVD subtype and iNPH. Data access statement: Datasets may be shared upon request to the corresponding author.
- 3 June 2026
- Published by SAGE Publications in International Journal of Stroke
- Vol. 21 (9) , 1348-1357
Open AccessTwo distinct patterns of progression of sporadic cerebral small vessel disease
Background: Sporadic cerebral small vessel disease (cSVD) is the most common cause of vascular dementia. cSVD features a wide range of neuroimaging markers, including white matter hyperintensity (WMH), lacunes, perivascular spaces, microbleeds, and diffusion magnetic resonance imaging (MRI)-based markers of white matter (WM) injury, and is accompanied by progressive cognitive decline. However, few studies have systematically investigated how and when various imaging and cognitive markers change along the complete cSVD progression trajectory. Aims: To establish changes in imaging and cognitive markers along the cSVD progression trajectory. Methods: In this cohort study, we selected 496 participants from the Radboud University Nijmegen Diffusion Tensor and Magnetic Resonance Imaging Cohort (RUNDMC) cohort as the discovery data set. The subtype and stage inference model was used to identify cSVD subtypes and their progression patterns. Five neuroimaging markers (WMH volume, median mean diffusivity (MD) in WM, WM volume, gray matter (GM) volume, and ventricle volume) were used in trajectory construction. The validation cohort was Cambridge cSVD (CamcSVD; n = 330). Results: We identified two subtypes in RUNDMC and replicated this finding in CamcSVD. One subtype displayed early WMH increase, followed by increased WM MD and tissue volume loss. The other subtype, which we termed “atrophy predominant,” showed early loss of GM and WM and ventricle enlargement, followed by WMH increase and increased WM MD. This subtype displayed an increase in markers of idiopathic normal pressure hydrocephalus (iNPH). Conclusion: We have delineated two distinct progression patterns in sporadic cSVD. Further studies are needed to investigate the relationship between the atrophy-predominant cSVD subtype and iNPH. Data access statement: Datasets may be shared upon request to the corresponding author.
- 1 March 2025
- Published by Oxford University Press (OUP) in European Stroke Journal
- Vol. 10 (1) , 225-235
Deep learning survival model predicts outcome after intracerebral hemorrhage from initial CT scan
Background: Predicting functional impairment after intracerebral hemorrhage (ICH) provides valuable information for planning of patient care and rehabilitation strategies. Current prognostic tools are limited in making long term predictions and require multiple expert-defined inputs and interpretation that make their clinical implementation challenging. This study aimed to predict long term functional impairment of ICH patients from admission non-contrast CT scans, leveraging deep learning models in a survival analysis framework. Methods: We used the admission non-contrast CT scans from 882 patients from the Massachusetts General Hospital ICH Study for training, hyperparameter optimization, and model selection, and 146 patients from the Yale New Haven ICH Study for external validation of a deep learning model predicting functional outcome. Disability (modified Rankin scale [mRS] > 2), severe disability (mRS > 4), and dependent living status were assessed via telephone interviews after 6, 12, and 24 months. The prediction methods were evaluated by the c-index and compared with ICH score and FUNC score. Results: Using non-contrast CT, our deep learning model achieved higher prediction accuracy of post-ICH dependent living, disability, and severe disability by 6, 12, and 24 months (c-index 0.742 [95% CI –0.700 to 0.778], 0.712 [95% CI –0.674 to 0.752], 0.779 [95% CI –0.733 to 0.832] respectively) compared with the ICH score (c-index 0.673 [95% CI –0.662 to 0.688], 0.647 [95% CI –0.637 to 0.661] and 0.697 [95% CI –0.675 to 0.717]) and FUNC score (c-index 0.701 [95% CI– 0.698 to 0.723], 0.668 [95% CI –0.657 to 0.680] and 0.727 [95% CI –0.708 to 0.753]). In the external independent Yale-ICH cohort, similar performance metrics were obtained for disability and severe disability (c-index 0.725 [95% CI –0.673 to 0.781] and 0.747 [95% CI –0.676 to 0.807], respectively). Similar AUC of predicting each outcome at 6 months, 1 and 2 years after ICH was achieved compared with ICH score and FUNC score. Conclusion: We developed a generalizable deep learning model to predict onset of dependent living and disability after ICH, which could help to guide treatment decisions, advise relatives in the acute setting, optimize rehabilitation strategies, and anticipate long-term care needs.
- 1 March 2025
- Published by Oxford University Press (OUP) in European Stroke Journal
- Vol. 10 (1) , 225-235
Deep learning survival model predicts outcome after intracerebral hemorrhage from initial CT scan
Background: Predicting functional impairment after intracerebral hemorrhage (ICH) provides valuable information for planning of patient care and rehabilitation strategies. Current prognostic tools are limited in making long term predictions and require multiple expert-defined inputs and interpretation that make their clinical implementation challenging. This study aimed to predict long term functional impairment of ICH patients from admission non-contrast CT scans, leveraging deep learning models in a survival analysis framework. Methods: We used the admission non-contrast CT scans from 882 patients from the Massachusetts General Hospital ICH Study for training, hyperparameter optimization, and model selection, and 146 patients from the Yale New Haven ICH Study for external validation of a deep learning model predicting functional outcome. Disability (modified Rankin scale [mRS] > 2), severe disability (mRS > 4), and dependent living status were assessed via telephone interviews after 6, 12, and 24 months. The prediction methods were evaluated by the c-index and compared with ICH score and FUNC score. Results: Using non-contrast CT, our deep learning model achieved higher prediction accuracy of post-ICH dependent living, disability, and severe disability by 6, 12, and 24 months (c-index 0.742 [95% CI –0.700 to 0.778], 0.712 [95% CI –0.674 to 0.752], 0.779 [95% CI –0.733 to 0.832] respectively) compared with the ICH score (c-index 0.673 [95% CI –0.662 to 0.688], 0.647 [95% CI –0.637 to 0.661] and 0.697 [95% CI –0.675 to 0.717]) and FUNC score (c-index 0.701 [95% CI– 0.698 to 0.723], 0.668 [95% CI –0.657 to 0.680] and 0.727 [95% CI –0.708 to 0.753]). In the external independent Yale-ICH cohort, similar performance metrics were obtained for disability and severe disability (c-index 0.725 [95% CI –0.673 to 0.781] and 0.747 [95% CI –0.676 to 0.807], respectively). Similar AUC of predicting each outcome at 6 months, 1 and 2 years after ICH was achieved compared with ICH score and FUNC score. Conclusion: We developed a generalizable deep learning model to predict onset of dependent living and disability after ICH, which could help to guide treatment decisions, advise relatives in the acute setting, optimize rehabilitation strategies, and anticipate long-term care needs.
- Open Access
- 18 February 2025
- Published by Wolters Kluwer Health in Journal of the American Heart Association
- Vol. 14 (4) , e037930
Spatial Correlates of Dementia and Disability After Intracerebral Hemorrhage
Background: Dementia and disability are highly prevalent after spontaneous intracerebral hemorrhage (ICH). Previous studies categorizing ICH by large anatomic boundaries have demonstrated that lobar ICH is associated with dementia, while ICH in the basal ganglia is associated with disability. This study aims to refine our understanding of the association between ICH location and post‐ICH dementia and disability at a voxel level, which could improve the prognostic accuracy of these outcomes and provide mechanistic insights into post‐ICH functional outcomes. Methods and Results: In this cohort study, we segmented the ICH lesions from the noncontrast computed tomography scans from 882 patients from the MGH‐ICH (Massachusetts General Hospital ICH Study) as the discovery data set and from 146 patients from the Yale‐ICH cohort as the validation data set. Using electronic health records and follow‐up telephone interviews, incident dementia ( International Classification of Diseases , Ninth Revision [ ICD‐9 ] codes of dementia or modified telephone interview for cognitive status 2) were identified. The median follow‐up times of the MGH‐ICH and Yale‐ICH cohorts were 2.9 (interquartile range, 1.0–5.8) years and 1.0 (interquartile range, 0.6–1.0) years, respectively. Two techniques of lesion symptom mapping were applied on the ICH lesions: sparse canonical correlation analysis for neuroimaging and voxel‐based lesion symptom mappings. Dementia conversion after ICH was associated with ICH in the left temporo‐occipital region (mean hazard ratio [HR], 3.62 [95% CI, 2.71–4.63]) and left superior longitudinal fasciculus (mean HR, 2.91 [95% CI, 2.40–3.52]). Development of disability after ICH was linked to the right cerebral peduncle (mean HR, 3.10 [95% CI, 2.44–3.94]), right pallidum (mean HR, 2.96 [95% CI, 1.99–4.25]), and right posterior limb of the internal capsule (mean HR, 2.54 [95% CI, 1.88–3.96]). Conclusions: Specific distribution of ICH lesions is linked to development of dementia and disability after ICH. These insights have the potential to enhance clinical prognostic models for patients with ICH, facilitating more precise predictions of outcomes based on hemorrhage location.
- 18 February 2025
- Published by Wolters Kluwer Health in Journal of the American Heart Association
- Vol. 14 (4) , e037930
Open AccessSpatial Correlates of Dementia and Disability After Intracerebral Hemorrhage
Background: Dementia and disability are highly prevalent after spontaneous intracerebral hemorrhage (ICH). Previous studies categorizing ICH by large anatomic boundaries have demonstrated that lobar ICH is associated with dementia, while ICH in the basal ganglia is associated with disability. This study aims to refine our understanding of the association between ICH location and post‐ICH dementia and disability at a voxel level, which could improve the prognostic accuracy of these outcomes and provide mechanistic insights into post‐ICH functional outcomes. Methods and Results: In this cohort study, we segmented the ICH lesions from the noncontrast computed tomography scans from 882 patients from the MGH‐ICH (Massachusetts General Hospital ICH Study) as the discovery data set and from 146 patients from the Yale‐ICH cohort as the validation data set. Using electronic health records and follow‐up telephone interviews, incident dementia ( International Classification of Diseases , Ninth Revision [ ICD‐9 ] codes of dementia or modified telephone interview for cognitive status 2) were identified. The median follow‐up times of the MGH‐ICH and Yale‐ICH cohorts were 2.9 (interquartile range, 1.0–5.8) years and 1.0 (interquartile range, 0.6–1.0) years, respectively. Two techniques of lesion symptom mapping were applied on the ICH lesions: sparse canonical correlation analysis for neuroimaging and voxel‐based lesion symptom mappings. Dementia conversion after ICH was associated with ICH in the left temporo‐occipital region (mean hazard ratio [HR], 3.62 [95% CI, 2.71–4.63]) and left superior longitudinal fasciculus (mean HR, 2.91 [95% CI, 2.40–3.52]). Development of disability after ICH was linked to the right cerebral peduncle (mean HR, 3.10 [95% CI, 2.44–3.94]), right pallidum (mean HR, 2.96 [95% CI, 1.99–4.25]), and right posterior limb of the internal capsule (mean HR, 2.54 [95% CI, 1.88–3.96]). Conclusions: Specific distribution of ICH lesions is linked to development of dementia and disability after ICH. These insights have the potential to enhance clinical prognostic models for patients with ICH, facilitating more precise predictions of outcomes based on hemorrhage location.
- Top Cited PapersOpen Access
- 7 November 2024
- Published by Oxford University Press (OUP) in Brain
- Vol. 148 (5) , 1540-1550
Perivascular space dysfunction in cerebral small vessel disease is related to neuroinflammation
Enlarged perivascular spaces are a feature of cerebral small vessel disease, and it has been hypothesized that they may reflect impaired glymphatic drainage. The mechanisms underlying perivascular spaces enlargement are not fully understood, but both increased inflammation and blood brain barrier permeability have been hypothesized to play a role. We investigated the relationship between perivascular spaces and both CNS and peripheral inflammation, as well as blood brain barrier permeability, in cerebral small vessel disease. Fifty-four symptomatic sporadic cerebral small vessel disease patients were studied. perivascular spaces were quantified using both a visual rating scale, and by measurement of perivascular spaces volume, in both the white matter and basal ganglia. PET-MRI was used to simultaneously measure microglial activation using the radioligand 11C-PK11195, and blood brain barrier permeability was acquired using dynamic contrast enhanced MRI. We determined 11C-PK11195 binding and blood brain barrier permeability in the local vicinity of individual perivascular spaces in concentric shells surrounding perivascular spaces. In addition, both mean 11C-PK11195 binding and blood brain barrier permeability in both the white matter and basal ganglia were determined. To assess systemic inflammation a panel of 93 blood biomarkers relating to cardiovascular disease, inflammation and endothelial activation were measured. Within the white matter tissue in closest proximity to perivascular spaces displayed greater 11C-PK11195 binding (p < 0.001), in the vicinity of perivascular spaces. Higher white matter perivascular spaces burden on the visual rating scale was associated with higher white matter 11C-PK11195 binding (rho = 0.469, FDR-p =0.009); values for perivascular spaces volume showed a similar trend. In contrast there were no associations between basal ganglia perivascular spaces burden and 11C-PK11195 binding. No perivascular spaces marker correlated with blood brain barrier permeability. There was no association between perivascular spaces markers and systemic inflammation blood biomarkers. Our findings demonstrate that white matter perivascular spaces are associated with increased 11C-PK11195 binding, consistent with neuroinflammation playing a role in white matter perivascular spaces enlargement. Further longitudinal and intervention studies and required to determine whether the relationship between neuroinflammation with enlarged perivascular spaces is causal.
- 7 November 2024
- Published by Oxford University Press (OUP) in Brain
- Vol. 148 (5) , 1540-1550
Top Cited PapersOpen AccessPerivascular space dysfunction in cerebral small vessel disease is related to neuroinflammation
Enlarged perivascular spaces are a feature of cerebral small vessel disease, and it has been hypothesized that they may reflect impaired glymphatic drainage. The mechanisms underlying perivascular spaces enlargement are not fully understood, but both increased inflammation and blood brain barrier permeability have been hypothesized to play a role. We investigated the relationship between perivascular spaces and both CNS and peripheral inflammation, as well as blood brain barrier permeability, in cerebral small vessel disease. Fifty-four symptomatic sporadic cerebral small vessel disease patients were studied. perivascular spaces were quantified using both a visual rating scale, and by measurement of perivascular spaces volume, in both the white matter and basal ganglia. PET-MRI was used to simultaneously measure microglial activation using the radioligand 11C-PK11195, and blood brain barrier permeability was acquired using dynamic contrast enhanced MRI. We determined 11C-PK11195 binding and blood brain barrier permeability in the local vicinity of individual perivascular spaces in concentric shells surrounding perivascular spaces. In addition, both mean 11C-PK11195 binding and blood brain barrier permeability in both the white matter and basal ganglia were determined. To assess systemic inflammation a panel of 93 blood biomarkers relating to cardiovascular disease, inflammation and endothelial activation were measured. Within the white matter tissue in closest proximity to perivascular spaces displayed greater 11C-PK11195 binding (p < 0.001), in the vicinity of perivascular spaces. Higher white matter perivascular spaces burden on the visual rating scale was associated with higher white matter 11C-PK11195 binding (rho = 0.469, FDR-p =0.009); values for perivascular spaces volume showed a similar trend. In contrast there were no associations between basal ganglia perivascular spaces burden and 11C-PK11195 binding. No perivascular spaces marker correlated with blood brain barrier permeability. There was no association between perivascular spaces markers and systemic inflammation blood biomarkers. Our findings demonstrate that white matter perivascular spaces are associated with increased 11C-PK11195 binding, consistent with neuroinflammation playing a role in white matter perivascular spaces enlargement. Further longitudinal and intervention studies and required to determine whether the relationship between neuroinflammation with enlarged perivascular spaces is causal.
- Open Access
- 1 September 2024
- Published by Wolters Kluwer Health in Stroke
- Vol. 55 (9) , 2254-2263
Improved Dementia Prediction in Cerebral Small Vessel Disease Using Deep Learning–Derived Diffusion Scalar Maps From T1
BACKGROUND: Cerebral small vessel disease is the most common pathology underlying vascular dementia. In small vessel disease, diffusion tensor imaging is more sensitive to white matter damage and better predicts dementia risk than conventional magnetic resonance imaging sequences, such as T1 and fluid attenuation inversion recovery, but diffusion tensor imaging takes longer to acquire and is not routinely available in clinical practice. As diffusion tensor imaging–derived scalar maps—fractional anisotropy (FA) and mean diffusivity (MD)—are frequently used in clinical settings, one solution is to synthesize FA/MD from T1 images. METHODS: We developed a deep learning model to synthesize FA/MD from T1. The training data set consisted of 4998 participants with the highest white matter hyperintensity volumes in the UK Biobank. Four external validations data sets with small vessel disease were included: SCANS (St George’s Cognition and Neuroimaging in Stroke; n=120), RUN DMC (Radboud University Nijmegen Diffusion Tensor and Magnetic Resonance Imaging Cohort; n=502), PRESERVE (Blood Pressure in Established Cerebral Small Vessel Disease; n=105), and NETWORKS (n=26), along with 1000 normal controls from the UK Biobank. RESULTS: The synthetic maps resembled ground-truth maps (structural similarity index >0.89 for MD maps and >0.80 for FA maps across all external validation data sets except for SCANS). The prediction accuracy of dementia using whole-brain median MD from the synthetic maps is comparable to the ground truth (SCANS ground-truth c-index, 0.822 and synthetic, 0.821; RUN DMC ground truth, 0.816 and synthetic, 0.812) and better than white matter hyperintensity volume (SCANS, 0.534; RUN DMC, 0.710). CONCLUSIONS: We have developed a fast and generalizable method to synthesize FA/MD maps from T1 to improve the prediction accuracy of dementia in small vessel disease when diffusion tensor imaging data have not been acquired.
- 1 September 2024
- Published by Wolters Kluwer Health in Stroke
- Vol. 55 (9) , 2254-2263
Open AccessImproved Dementia Prediction in Cerebral Small Vessel Disease Using Deep Learning–Derived Diffusion Scalar Maps From T1
BACKGROUND: Cerebral small vessel disease is the most common pathology underlying vascular dementia. In small vessel disease, diffusion tensor imaging is more sensitive to white matter damage and better predicts dementia risk than conventional magnetic resonance imaging sequences, such as T1 and fluid attenuation inversion recovery, but diffusion tensor imaging takes longer to acquire and is not routinely available in clinical practice. As diffusion tensor imaging–derived scalar maps—fractional anisotropy (FA) and mean diffusivity (MD)—are frequently used in clinical settings, one solution is to synthesize FA/MD from T1 images. METHODS: We developed a deep learning model to synthesize FA/MD from T1. The training data set consisted of 4998 participants with the highest white matter hyperintensity volumes in the UK Biobank. Four external validations data sets with small vessel disease were included: SCANS (St George’s Cognition and Neuroimaging in Stroke; n=120), RUN DMC (Radboud University Nijmegen Diffusion Tensor and Magnetic Resonance Imaging Cohort; n=502), PRESERVE (Blood Pressure in Established Cerebral Small Vessel Disease; n=105), and NETWORKS (n=26), along with 1000 normal controls from the UK Biobank. RESULTS: The synthetic maps resembled ground-truth maps (structural similarity index >0.89 for MD maps and >0.80 for FA maps across all external validation data sets except for SCANS). The prediction accuracy of dementia using whole-brain median MD from the synthetic maps is comparable to the ground truth (SCANS ground-truth c-index, 0.822 and synthetic, 0.821; RUN DMC ground truth, 0.816 and synthetic, 0.812) and better than white matter hyperintensity volume (SCANS, 0.534; RUN DMC, 0.710). CONCLUSIONS: We have developed a fast and generalizable method to synthesize FA/MD maps from T1 to improve the prediction accuracy of dementia in small vessel disease when diffusion tensor imaging data have not been acquired.
- Open Access
- 1 September 2023
- Published by Oxford University Press (OUP) in European Stroke Journal
- Vol. 8 (3) , 629-637
Prediction of response to thrombolysis in acute stroke using neural network analysis of CT perfusion imaging
Background: In ischaemic stroke patients undergoing reperfusion therapy, the amount of salvageable tissue, that is, extent of the ischaemic penumbra, predicts the clinical outcomes. CT perfusion (CTP) enables quantification of penumbral tissues to guide decision making, and current programmes have automated its analysis. More advanced machine learning techniques utilising the CTP maps may improve prediction beyond the ischaemic volume measures. Method: We determined whether applying convolutional neural networks (CNN), a key machine learning technique in modelling image-label relationships, to post-processed CTP maps improved prediction of outcome, assessed by 3 months modified Rankin scale (mRS). Patients who underwent thrombolysis but not thrombectomy were included. CTP maps of a retrospective cohort of 230 patients with middle cerebral artery stroke were used to develop the model, which was validated in an independent cohort of 129 patients. Results: We constructed a CNN model that predicted a favourable post-thrombolysis outcome (mRS 0–2 at 3 months) with an area under receiver-operator characteristics curve (AUC) of 0.792 (95% CI, 0.707–0.877). This model outperformed a currently clinically used MISTAR software using previously validated thresholds (AUC = 0.583, 95% CI, 0.480–0.686) and a model modified using thresholds from the derivation cohort (AUC = 0.670, 95% CI, 0.571–0.769). By combining CNN-derived features and baseline demographic features, the prediction AUC was improved to 0.865 (95% CI, 0.794–0.936). Conclusion: CNN improved prediction of post-thrombolysis outcome, and may be useful in selecting which patients benefit from thrombolysis.
- 1 September 2023
- Published by Oxford University Press (OUP) in European Stroke Journal
- Vol. 8 (3) , 629-637
Open AccessPrediction of response to thrombolysis in acute stroke using neural network analysis of CT perfusion imaging
Background: In ischaemic stroke patients undergoing reperfusion therapy, the amount of salvageable tissue, that is, extent of the ischaemic penumbra, predicts the clinical outcomes. CT perfusion (CTP) enables quantification of penumbral tissues to guide decision making, and current programmes have automated its analysis. More advanced machine learning techniques utilising the CTP maps may improve prediction beyond the ischaemic volume measures. Method: We determined whether applying convolutional neural networks (CNN), a key machine learning technique in modelling image-label relationships, to post-processed CTP maps improved prediction of outcome, assessed by 3 months modified Rankin scale (mRS). Patients who underwent thrombolysis but not thrombectomy were included. CTP maps of a retrospective cohort of 230 patients with middle cerebral artery stroke were used to develop the model, which was validated in an independent cohort of 129 patients. Results: We constructed a CNN model that predicted a favourable post-thrombolysis outcome (mRS 0–2 at 3 months) with an area under receiver-operator characteristics curve (AUC) of 0.792 (95% CI, 0.707–0.877). This model outperformed a currently clinically used MISTAR software using previously validated thresholds (AUC = 0.583, 95% CI, 0.480–0.686) and a model modified using thresholds from the derivation cohort (AUC = 0.670, 95% CI, 0.571–0.769). By combining CNN-derived features and baseline demographic features, the prediction AUC was improved to 0.865 (95% CI, 0.794–0.936). Conclusion: CNN improved prediction of post-thrombolysis outcome, and may be useful in selecting which patients benefit from thrombolysis.
- 1 January 2023
- Published by Wolters Kluwer Health in Stroke
- Vol. 54 (1) , E7-E8
Beyond the Limits of Visual Learning: Prediction of Time From Onset From Noncontrast CT of Acute Ischemic Stroke
- 1 January 2023
- Published by Wolters Kluwer Health in Stroke
- Vol. 54 (1) , E7-E8
Beyond the Limits of Visual Learning: Prediction of Time From Onset From Noncontrast CT of Acute Ischemic Stroke