pith. sign in

arxiv: 2506.17501 · v3 · submitted 2025-06-20 · 📡 eess.IV · cs.CV

DSA-NRP: No-Reflow Prediction from Angiographic Perfusion Dynamics in Stroke EVT

Shreeram Athreya , Carlos Olivares , Ameera Ismail , Kambiz Nael , William Speier , Corey Arnold This is my paper

Pith reviewed 2026-05-19 07:49 UTC · model grok-4.3

classification 📡 eess.IV cs.CV
keywords no-reflowendovascular thrombectomydigital subtraction angiographymachine learning predictionacute ischemic strokeperfusion dynamicsmicrovascular hypoperfusion
0
0 comments X

The pith

Perfusion features from DSA sequences predict no-reflow after successful EVT more accurately than clinical variables.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper develops a machine learning model to forecast no-reflow, which is ongoing poor blood flow in small vessels after a stroke clot has been removed from a large artery. It does this by analyzing perfusion patterns in digital subtraction angiography images taken during the procedure, along with some clinical details. This matters because doctors currently have to wait up to a day for an MRI to check for this issue, but the new method could flag problems right away. A reader would care if this means faster decisions to help brain tissue recover better. The results show the DSA-based model beats a baseline using only clinical info by a noticeable margin in accuracy and AUC scores.

Core claim

The authors claim that statistical and temporal perfusion features extracted from pre- and post-EVT DSA sequences in the target downstream territory allow ML classifiers to predict no-reflow, defined as persistent hypoperfusion on post-procedure MRI, with an AUC of 0.7703 compared to 0.5728 for clinical features alone, in patients who achieved good recanalization scores.

What carries the argument

Machine learning classifiers trained on statistical and temporal perfusion features from angiographic sequences of the target territory.

If this is right

  • Clinicians could identify high-risk patients immediately after EVT for proactive management.
  • Intra-procedural DSA data provides real-time insights into microvascular integrity.
  • This reduces dependence on delayed perfusion MRI for no-reflow detection.
  • The framework lays groundwork for integrating such predictions into acute stroke care workflows.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • If validated, this approach might be extended to predict other post-EVT complications like hemorrhage using similar DSA analysis.
  • It could influence the design of future EVT protocols to include automated perfusion feature extraction.
  • A testable extension would be applying the model to multi-center data to check generalizability.

Load-bearing premise

The features from DSA sequences capture microvascular integrity separately from whether the main artery was successfully opened.

What would settle it

A follow-up study in which the DSA model shows no better performance than the clinical baseline on new patient data would disprove the advantage.

Figures

Figures reproduced from arXiv: 2506.17501 by Ameera Ismail, Carlos Olivares, Corey Arnold, Kambiz Nael, Shreeram Athreya, William Speier.

Figure 1
Figure 1. Figure 1: Overview of clinical workflow and the proposed DSA-NRP pipeline for no-reflow prediction. The proposed approach enables immediate, intra-procedural prediction of no-reflow using features extracted from DSA sequences and machine learning, reducing reliance on delayed follow-up MRI and supporting earlier, more targeted intervention. hypoperfusion despite macrovascular recanalization, is well documented [7, 8… view at source ↗
Figure 2
Figure 2. Figure 2: Conversion of DSA video data from the TDT region into a 1D time-series signal. Mapping the mean pixel intensity within the TDT across all frames produces a compact time-series representation, enabling quantitative analysis of pre- and post-procedure perfusion dynamics for no-reflow prediction. were derived from pre- and post-procedure 1D intensity signals, xpre(t) and xpost(t), from AP and lateral DSA sequ… view at source ↗
Figure 3
Figure 3. Figure 3: Mean intensity time-series curves from the TDT region, averaged across patients for each outcome group. Distinct temporal profiles between reflow and no-reflow cases are observed in both lateral and anteroposterior views, highlighting key differences in post-procedural perfusion dynamics. 3.3.3 Temporal Flow Comparison (FLOW) These features quantify timing-related tracer perfusion dynamics, including onset… view at source ↗
Figure 4
Figure 4. Figure 4: Model performance and feature importance for no-reflow prediction using DSA-derived features. (a) ROC curves demonstrate strong discriminative ability of the proposed model. (b) Feature importance analysis highlights key contributors from the combined PEAK and SIPS set, underscoring the physiological relevance of intensity dynamics and statistical descriptors in predicting no-reflow. Difference in time to … view at source ↗
read the original abstract

Following successful large-vessel recanalization via endovascular thrombectomy (EVT) for acute ischemic stroke (AIS), some patients experience a complication known as no-reflow, defined by persistent microvascular hypoperfusion that undermines tissue recovery and worsens clinical outcomes. Although prompt identification is crucial, standard clinical practice relies on perfusion magnetic resonance imaging (MRI) within 24 hours post-procedure, delaying intervention. In this work, we introduce the first-ever machine learning (ML) framework to predict no-reflow immediately after EVT by leveraging previously unexplored intra-procedural digital subtraction angiography (DSA) sequences and clinical variables. Our retrospective analysis included AIS patients treated at UCLA Medical Center (2011-2024) who achieved favorable mTICI scores (2b-3) and underwent pre- and post-procedure MRI. No-reflow was defined as persistent hypoperfusion (Tmax > 6 s) on post-procedural imaging. From DSA sequences (AP and lateral views), we extracted statistical and temporal perfusion features from the target downstream territory to train ML classifiers for predicting no-reflow. Our novel method significantly outperformed a clinical-features baseline(AUC: 0.7703 $\pm$ 0.12 vs. 0.5728 $\pm$ 0.12; accuracy: 0.8125 $\pm$ 0.10 vs. 0.6331 $\pm$ 0.09), demonstrating that real-time DSA perfusion dynamics encode critical insights into microvascular integrity. This approach establishes a foundation for immediate, accurate no-reflow prediction, enabling clinicians to proactively manage high-risk patients without reliance on delayed imaging.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 1 minor

Summary. The paper introduces DSA-NRP, an ML framework that extracts statistical and temporal perfusion features from pre- and post-EVT DSA sequences (AP and lateral views) of the target downstream territory, combined with clinical variables, to predict no-reflow (Tmax > 6 s on post-procedural MRI) in mTICI 2b-3 patients. In a retrospective UCLA cohort (2011-2024), the model reports AUC 0.7703 ± 0.12 and accuracy 0.8125 ± 0.10, outperforming a clinical-features baseline (AUC 0.5728 ± 0.12, accuracy 0.6331 ± 0.09), claiming that DSA dynamics encode microvascular integrity for immediate prediction.

Significance. If the reported performance gain proves robust under pre-specified protocols and proper validation, the work could enable real-time intra-procedural no-reflow risk stratification, reducing reliance on delayed post-EVT MRI and supporting earlier intervention. The novelty of leveraging intra-procedural DSA perfusion dynamics is a clear strength, though the single-center retrospective design and limited methodological transparency temper immediate clinical impact.

major comments (2)
  1. [Abstract/Methods] Abstract/Methods: No pre-specified protocol is described for defining the target downstream territory, selecting AP vs. lateral views, or setting time-window boundaries for feature extraction. This leaves open the possibility that territory masks or windows were refined after inspecting MRI labels or outcomes, which would make the AUC improvement (0.7703 vs. 0.5728) potentially artifactual rather than evidence that DSA dynamics independently capture microvascular integrity after successful recanalization.
  2. [Results] Results: Only aggregate AUC and accuracy with standard deviations are presented; the manuscript provides no details on cross-validation folds, stratification, feature selection procedure, or class-imbalance handling. Without these, the statistical reliability of the headline comparison to the clinical baseline cannot be assessed and the central claim remains difficult to evaluate.
minor comments (1)
  1. [Abstract] Abstract: The total number of patients, no-reflow prevalence, and exact clinical features used in the baseline are not stated, which are needed to interpret the reported metrics and baseline performance.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the thoughtful and constructive review. The comments highlight important aspects of methodological transparency that we address point by point below. We have revised the manuscript to provide the requested details on protocol definition and validation procedures while preserving the integrity of the original analysis.

read point-by-point responses

  1. Referee: [Abstract/Methods] Abstract/Methods: No pre-specified protocol is described for defining the target downstream territory, selecting AP vs. lateral views, or setting time-window boundaries for feature extraction. This leaves open the possibility that territory masks or windows were refined after inspecting MRI labels or outcomes, which would make the AUC improvement (0.7703 vs. 0.5728) potentially artifactual rather than evidence that DSA dynamics independently capture microvascular integrity after successful recanalization.

    Authors: We agree that explicit documentation of the analysis protocol is essential to rule out post-hoc refinement. The target downstream territory was defined a priori as the vascular bed distal to the site of large-vessel occlusion, identified on the initial diagnostic DSA run using standard anatomical landmarks (e.g., MCA territory for M1 occlusions). AP and lateral projections were chosen because they constitute the routine biplane acquisitions performed during EVT at our center; no post-MRI selection occurred. Time windows were fixed to the full contrast transit phase (injection to venous clearance, typically 4–12 s) based on prior angiographic literature rather than outcome inspection. Nevertheless, to eliminate any ambiguity we have added a new subsection titled “Target Territory Definition and View Selection Protocol” in the Methods that codifies these rules with explicit exclusion criteria and timing parameters. We also include a supplementary figure illustrating example territory masks. revision: yes

  2. Referee: [Results] Results: Only aggregate AUC and accuracy with standard deviations are presented; the manuscript provides no details on cross-validation folds, stratification, feature selection procedure, or class-imbalance handling. Without these, the statistical reliability of the headline comparison to the clinical baseline cannot be assessed and the central claim remains difficult to evaluate.

    Authors: We acknowledge that the original Results section omitted the granular validation details required for independent assessment. The reported figures derive from 5-fold stratified cross-validation with folds balanced by no-reflow label (prevalence ~28 %). Stratification was performed at the patient level to avoid leakage across pre- and post-EVT sequences from the same individual. Feature selection used recursive feature elimination with cross-validation inside each outer fold; the final feature set was locked before evaluating the clinical baseline. Class imbalance was addressed by class-weighted logistic regression and random-forest variants; no synthetic oversampling was applied. We have expanded the Results section with a dedicated paragraph describing the full pipeline, added a supplementary table listing per-fold AUCs, and included a CONSORT-style flow diagram of the modeling steps. These additions allow direct evaluation of the 0.77 vs. 0.57 AUC difference. revision: yes

Circularity Check

0 steps flagged

No circularity: standard empirical ML prediction on held-out data

full rationale

The paper presents an empirical machine learning classifier that extracts statistical and temporal perfusion features from pre- and post-EVT DSA sequences in the target downstream territory and trains models to predict no-reflow (defined via post-procedural MRI Tmax > 6 s). Performance is reported with standard deviations consistent with cross-validation or held-out testing (AUC 0.7703 ± 0.12 vs baseline 0.5728 ± 0.12). No equations, derivations, or first-principles results are described that reduce the output to an input quantity by construction. No self-citations, uniqueness theorems, or ansatzes are invoked to justify the feature set or model; the central claim rests on direct comparison of classifier accuracy against a clinical-features baseline using standard ML practices.

Axiom & Free-Parameter Ledger

1 free parameters · 1 axioms · 0 invented entities

The central claim rests on the assumption that DSA-derived perfusion statistics are informative of microvascular status and that the retrospective cohort is representative; no new physical entities or ad-hoc constants are introduced beyond standard ML hyperparameters.

free parameters (1)
  • ML classifier hyperparameters
    Choice of model type, regularization, and feature selection thresholds fitted or tuned on the training split.
axioms (1)
  • domain assumption No-reflow is correctly defined by persistent Tmax > 6 s on post-procedural perfusion MRI.
    Definition used to label the target variable.

pith-pipeline@v0.9.0 · 5853 in / 1272 out tokens · 20802 ms · 2026-05-19T07:49:05.717034+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

43 extracted references · 43 canonical work pages

  1. [1]

    Global Epidemiology of Stroke and Access to Acute Ischemic Stroke Interventions Neurology

    Saini Vasu, Guada Luis, Yavagal Dileep R.. Global Epidemiology of Stroke and Access to Acute Ischemic Stroke Interventions Neurology. 2021;97

    work page 2021

  2. [2]

    Global stroke statistics 2023: Availability of reperfusion services around the world International Journal of Stroke

    Kim Joosup, Olaiya Muideen T, De Silva Deidre A, et al. Global stroke statistics 2023: Availability of reperfusion services around the world International Journal of Stroke. 2024;19:253–270

    work page 2023

  3. [3]

    Combined Diffusion and Perfusion MRI With Correlation to Single- Photon Emission CT in Acute Ischemic Stroke: Ischemic Penumbra Predicts Infarct Growth Stroke

    Karonen Jari O., Vanninen Ritva L., Liu Yawu, et al. Combined Diffusion and Perfusion MRI With Correlation to Single- Photon Emission CT in Acute Ischemic Stroke: Ischemic Penumbra Predicts Infarct Growth Stroke. 1999;30:1583–1590

    work page 1999

  4. [4]

    Endovascular Thrombectomy for Acute Ischemic Stroke: A Meta-analysis JAMA

    Badhiwala Jetan H., Nassiri Farshad, Alhazzani Waleed, et al. Endovascular Thrombectomy for Acute Ischemic Stroke: A Meta-analysis JAMA. 2015;314:1832

    work page 2015

  5. [5]

    Thrombectomy for Acute Ischemic Stroke: Recent Insights and Future Directions Current Neurology and Neuroscience Reports

    Ganesh Aravind, Goyal Mayank. Thrombectomy for Acute Ischemic Stroke: Recent Insights and Future Directions Current Neurology and Neuroscience Reports. 2018;18:59

    work page 2018

  6. [6]

    Hyperacute Management of Ischemic Strokes Journal of the American College of Cardiology

    Patel Pratit, Yavagal Dileep, Khandelwal Priyank. Hyperacute Management of Ischemic Strokes Journal of the American College of Cardiology. 2020;75:1844–1856

    work page 2020

  7. [7]

    No-reflow phenomenon in stroke patients: A systematic literature review and meta-analysis of clinical data International Journal of Stroke

    Mujanovic Adnan, Ng Felix, Meinel Thomas R, et al. No-reflow phenomenon in stroke patients: A systematic literature review and meta-analysis of clinical data International Journal of Stroke. 2024;19:58–67

    work page 2024

  8. [8]

    Recanalization Does Not Always Equate to Reperfusion: No-Reflow Phenomenon After Successful Thrombectomy Stroke

    Horie Nobutaka, Inoue Manabu, Morimoto Takeshi, et al. Recanalization Does Not Always Equate to Reperfusion: No-Reflow Phenomenon After Successful Thrombectomy Stroke. 2025;56:183–189

    work page 2025

  9. [9]

    Schiphorst Adrien Ter, Charron Sylvain, Hassen Wagih Ben, et al. Tissueno-reflow despite full recanalization following thrombectomy for anterior circulation stroke with proximal occlusion: A clinical study Journal of Cerebral Blood Flow & Metabolism. 2021;41:253–266

    work page 2021

  10. [10]

    Prevalence and Significance of Impaired Microvascular Tissue Reperfusion Despite Macrovascular Angiographic Reperfusion (No-Reflow) Neurology

    Ng Felix C., Churilov Leonid, Yassi Nawaf, et al. Prevalence and Significance of Impaired Microvascular Tissue Reperfusion Despite Macrovascular Angiographic Reperfusion (No-Reflow) Neurology. 2022;98. 10 DSA-NRP: No-Reflow Prediction from Angiographic Perfusion Dynamics in Stroke EVT

    work page 2022

  11. [11]

    eTICI reperfusion: defining success in endovascular stroke therapy Journal of NeuroInterventional Surgery

    Liebeskind David S, Bracard Serge, Guillemin Francis, et al. eTICI reperfusion: defining success in endovascular stroke therapy Journal of NeuroInterventional Surgery. 2019;11:433–438

    work page 2019

  12. [12]

    Association of Intravenous Thrombolysis with Delayed Reperfusion After Incomplete Mechanical Thrombectomy Clinical Neuroradiology

    Mujanovic Adnan, Kammer Christoph, Kurmann Christoph C., et al. Association of Intravenous Thrombolysis with Delayed Reperfusion After Incomplete Mechanical Thrombectomy Clinical Neuroradiology. 2023;33:87–98

    work page 2023

  13. [13]

    Powers William J., Derdeyn Colin P., Biller José, et al. 2015 American Heart Association/American Stroke Association Focused Update of the 2013 Guidelines for the Early Management of Patients With Acute Ischemic Stroke Regarding Endovascular Treatment: A Guideline for Healthcare Professionals From the American Heart Association/American Stroke Association...

    work page 2015

  14. [14]

    Cho Yong-Hwan, Choi Jae Hyung. Mechanical thrombectomy for acute ischemic stroke with occlusion of the M2 segment of the middle cerebral artery: A literature review Journal of Cerebrovascular and Endovascular Neurosurgery. 2021;23:193–200

    work page 2021

  15. [15]

    Putu Eka

    Liyis Bryan Gervais De, Surya Stevanus Christian, Tedyanto Eric Hartono, Pramana Nyoman Angga Krishna, Widyadharma I. Putu Eka. Mechanical thrombectomy in M1 and M2 segments of middle cerebral arteries: A meta-analysis of prospective cohort studies Clinical Neurology and Neurosurgery. 2023;231:107823

    work page 2023

  16. [16]

    Endovascular Treatment of Stroke Due to Medium-Vessel Occlusion New England Journal of Medicine

    Goyal Mayank, Ospel Johanna M., Ganesh Aravind, et al. Endovascular Treatment of Stroke Due to Medium-Vessel Occlusion New England Journal of Medicine. 2025;392:1385–1395

    work page 2025

  17. [17]

    No-reflow phenomenon following stroke recanalization therapy: Clinical assessment advances: A narrative review Brain Circulation

    Kan Yuan, Li Sijie, Zhang Bowei, Ding Yuchuan, Zhao Wenbo, Ji Xunming. No-reflow phenomenon following stroke recanalization therapy: Clinical assessment advances: A narrative review Brain Circulation. 2023;9:214–221

    work page 2023

  18. [18]

    Neuroimaging of Acute Ischemic Stroke: Multimodal Imaging Approach for Acute Endovascular Therapy Journal of Stroke

    Abdalkader Mohamad, Siegler James E., Lee Jin Soo, et al. Neuroimaging of Acute Ischemic Stroke: Multimodal Imaging Approach for Acute Endovascular Therapy Journal of Stroke. 2023;25:55–71

    work page 2023

  19. [19]

    Comparison of CT and three MR sequences for detecting and categorizing early (48 hours) hemorrhagic transformation in hyperacute ischemic stroke AJNR

    Arnould Marie-Cécile, Grandin Cécile B., Peeters André, Cosnard Guy, Duprez Thierry P.. Comparison of CT and three MR sequences for detecting and categorizing early (48 hours) hemorrhagic transformation in hyperacute ischemic stroke AJNR. American journal of neuroradiology. 2004;25:939–944

    work page 2004

  20. [20]

    Comparative Sensitivity of Computed Tomography vs

    Hwang David Y ., Silva Gisele S., Furie Karen L., Greer David M.. Comparative Sensitivity of Computed Tomography vs. Magnetic Resonance Imaging for Detecting Acute Posterior Fossa Infarct The Journal of Emergency Medicine. 2012;42:559–565

    work page 2012

  21. [21]

    Dynamic Cerebral Autoregulation Post Endovascular Thrombectomy in Acute Ischemic Stroke Brain Sciences

    Sheriff Faheem, Castro Pedro, Kozberg Mariel, et al. Dynamic Cerebral Autoregulation Post Endovascular Thrombectomy in Acute Ischemic Stroke Brain Sciences. 2020;10:641

    work page 2020

  22. [22]

    Current and Emerging Endovascular and Neurocritical Care Management Strategies in Large-Core Ischemic Stroke Journal of Clinical Medicine

    Migdady Ibrahim, Johnson-Black Phoebe H., Leslie-Mazwi Thabele, Malhotra Rishi. Current and Emerging Endovascular and Neurocritical Care Management Strategies in Large-Core Ischemic Stroke Journal of Clinical Medicine. 2023;12:6641

    work page 2023

  23. [23]

    Blood Pressure Management After Endovascular Thrombectomy Frontiers in Neurology

    Peng Teng J., Ortega-Gutiérrez Santiago, De Havenon Adam, Petersen Nils H.. Blood Pressure Management After Endovascular Thrombectomy Frontiers in Neurology. 2021;12:723461

    work page 2021

  24. [24]

    Blood pressure management after endovascular thrombectomy: Insights of recent randomized controlled trials CNS Neuroscience & Therapeutics

    Dong Xiao, Liu Yuanyuan, Chu Xuehong, et al. Blood pressure management after endovascular thrombectomy: Insights of recent randomized controlled trials CNS Neuroscience & Therapeutics. 2024;30:e14907

    work page 2024

  25. [25]

    Thrombus Aspiration during Primary Percutaneous Coronary Intervention New England Journal of Medicine

    Svilaas Tone, Vlaar Pieter J., Van Der Horst Iwan C., et al. Thrombus Aspiration during Primary Percutaneous Coronary Intervention New England Journal of Medicine. 2008;358:557–567

    work page 2008

  26. [26]

    Kaesmacher Johannes, Bellwald Sebastian, Dobrocky Tomas, et al. Safety and Efficacy of Intra-arterial Urokinase After Failed, Unsuccessful, or Incomplete Mechanical Thrombectomy in Anterior Circulation Large-Vessel Occlusion Stroke JAMA Neurology. 2020;77:318

    work page 2020

  27. [27]

    Renú Arturo, Millán Mónica, San Román Luis, et al. Effect of Intra-arterial Alteplase vs Placebo Following Successful Thrombectomy on Functional Outcomes in Patients With Large Vessel Occlusion Acute Ischemic Stroke: The CHOICE Randomized Clinical Trial JAMA. 2022;327:826

    work page 2022

  28. [28]

    P., Van Der Sluijs Matthijs, et al

    Su Ruisheng, Cornelissen Sandra A. P., Van Der Sluijs Matthijs, et al. autoTICI: Automatic Brain Tissue Reperfusion Scoring on 2D DSA Images of Acute Ischemic Stroke Patients IEEE Transactions on Medical Imaging. 2021;40:2380– 2391

    work page 2021

  29. [29]

    A Comprehensive Prediction Model for Futile Recanalization in AIS Patients Post-Endovascular Therapy: Integrating Clinical, Imaging, and No-Reflow Biomarkers.Aging and disease

    Huang Shuangfeng, Xu Jiali, Kang Haijuan, et al. A Comprehensive Prediction Model for Futile Recanalization in AIS Patients Post-Endovascular Therapy: Integrating Clinical, Imaging, and No-Reflow Biomarkers.Aging and disease. 2024;15:2852–2862. Place: United States

    work page 2024

  30. [30]

    No-reflow phenomenon in acute ischemic stroke: an angiographic evaluation Neurological Sciences

    Nicolini Ettore, Iacobucci Marta, De Michele Manuela, et al. No-reflow phenomenon in acute ischemic stroke: an angiographic evaluation Neurological Sciences. 2023;44:3939–3948

    work page 2023

  31. [31]

    qTICI: Quantitative assessment of brain tissue reperfusion on digital subtraction angiograms of acute ischemic stroke patients International Journal of Stroke

    Prasetya Haryadi, Ramos Lucas A, Epema Thabiso, et al. qTICI: Quantitative assessment of brain tissue reperfusion on digital subtraction angiograms of acute ischemic stroke patients International Journal of Stroke. 2021;16:207–216. 11 DSA-NRP: No-Reflow Prediction from Angiographic Perfusion Dynamics in Stroke EVT

    work page 2021

  32. [32]

    Predictors of futile recanalization after endovascular treatment in acute ischemic stroke: a multi-center study Frontiers in Neuroscience

    Sun Yu, Jou Eric, Nguyen Thanh N., et al. Predictors of futile recanalization after endovascular treatment in acute ischemic stroke: a multi-center study Frontiers in Neuroscience. 2023;17:1279366

    work page 2023

  33. [33]

    Image processing-based mTICI grading after endovascular treatment for acute ischemic stroke Current Directions in Biomedical Engineering

    Sabieleish Muhannad, Thormann Maximilian, Metzler Jonathan, et al. Image processing-based mTICI grading after endovascular treatment for acute ischemic stroke Current Directions in Biomedical Engineering. 2021;7:235–238

    work page 2021

  34. [34]

    Deep learning-based classification of DSA image sequences of patients with acute ischemic stroke International Journal of Computer Assisted Radiology and Surgery

    Mittmann Benjamin J., Braun Michael, Runck Frank, et al. Deep learning-based classification of DSA image sequences of patients with acute ischemic stroke International Journal of Computer Assisted Radiology and Surgery. 2022;17:1633– 1641

    work page 2022

  35. [35]

    Zhang Haoyue, Polson Jennifer, Nael Kambiz, et al. A Machine Learning Approach to Predict Acute Ischemic Stroke Thrombectomy Reperfusion using Discriminative MR Image Features in 2021 IEEE EMBS International Conference on Biomedical and Health Informatics (BHI)(Athens, Greece):1–4IEEE 2021

    work page 2021

  36. [36]

    A Deep Learning Approach to Predict Recanalization First-Pass Effect following Mechanical Thrombectomy in Patients with Acute Ischemic Stroke American Journal of Neuroradiology

    Zhang Haoyue, Polson Jennifer S., Wang Zichen, et al. A Deep Learning Approach to Predict Recanalization First-Pass Effect following Mechanical Thrombectomy in Patients with Acute Ischemic Stroke American Journal of Neuroradiology. 2024;45:1044–1052

    work page 2024

  37. [37]

    Ensemble machine learning to predict futile recanalization after mechanical thrombectomy based on non-contrast CT imaging Journal of Stroke and Cerebrovascular Diseases

    Da Ros Valerio, Cavallo Armando, Di Donna Carlo, et al. Ensemble machine learning to predict futile recanalization after mechanical thrombectomy based on non-contrast CT imaging Journal of Stroke and Cerebrovascular Diseases. 2024;33:107890

    work page 2024

  38. [38]

    DEEP MOVEMENT: Deep learning of movie files for management of endovascular thrombectomy European Radiology

    Kelly Brendan, Martinez Mesha, Do Huy, et al. DEEP MOVEMENT: Deep learning of movie files for management of endovascular thrombectomy European Radiology. 2023;33:5728–5739

    work page 2023

  39. [39]

    CVFSNet: A Cross View Fusion Scoring Network for end-to-end mTICI scoring Medical Image Analysis

    Xu Weijin, Tan Tao, Yang Huihua, et al. CVFSNet: A Cross View Fusion Scoring Network for end-to-end mTICI scoring Medical Image Analysis. 2025;102:103508

    work page 2025

  40. [40]

    Computed Tomography Perfusion After Thrombectomy: An Immediate Surrogate Marker of Outcome After Recanalization in Acute Stroke Stroke

    Rubiera Marta, Garcia-Tornel Alvaro, Olivé-Gadea Marta, et al. Computed Tomography Perfusion After Thrombectomy: An Immediate Surrogate Marker of Outcome After Recanalization in Acute Stroke Stroke. 2020;51:1736–1742

    work page 2020

  41. [41]

    Jann Kay, Hauf Martinus, Kellner Weldon Frauke, et al. Implication of cerebral circulation time in intracranial stenosis measured by digital subtraction angiography on cerebral blood flow estimation measured by arterial spin labelingDiagnostic and Interventional Radiology. 2016;22:481–488

    work page 2016

  42. [42]

    Automatic flow analysis of digital subtraction angiography using independent component analysis in patients with carotid stenosis PLOS ONE

    Lee Han-Jui, Hong Jia-Sheng, Lin Chung-Jung, et al. Automatic flow analysis of digital subtraction angiography using independent component analysis in patients with carotid stenosis PLOS ONE. 2017;12:e0185330

    work page 2017

  43. [43]

    Digital Subtraction Angiography Contrast Material Transport as a Direct Assessment for Blood Perfusion of Middle Cerebral Artery Stenosis Frontiers in Physiology

    Lu Yun-Hao, Cai Yan, Zhang Yi, Wang Rui, Li Zhi-Yong. Digital Subtraction Angiography Contrast Material Transport as a Direct Assessment for Blood Perfusion of Middle Cerebral Artery Stenosis Frontiers in Physiology. 2021;12:716173. 12

    work page 2021