Validation of a machine learning software tool for automated large vessel occlusion detection in patients with suspected acute stroke

Cimflová, Petra; Golan, Rotem; Ospel, Johanna M.; Sojoudi, Alireza; Duszynski, Chris; Elebute, Ibukun; El-Hariri, Houssam; Mousavi, Seyed Hossein; Neto, Luis A. Souto Maior; Pinky, Najratun Nayem; Beland, Benjamin; Bala, Fouzi; Kashani, Nima; Hu, William; Joshi, Manish; Qiu, Wu; Menon, Bijoy K

doi:10.1007/s00234-022-02978-x

Cited by 5 publications

(3 citation statements)

References 26 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…AI has been recently applied to create a machine-learning tool using CTA images for evaluating intracranial internal carotid artery (ICA) stenosis in patients with acute ischemic stroke. StrokeSENS LVO model has been created by analyzing retrospectively 400 studies (217 LVO, 183 other/no occlusion): the algorithm has shown a high accuracy (92.7%), sensitivity (85.7%), and specificity (87.4%) in detecting intracranial ICA occlusion, without differences in patient age, sex, or CTA acquisition characteristics [ 47 ]. Moreover, Buckler proposed an interesting deep-learning algorithm for the stratification of atherosclerotic lesions in different phenotypes based on plaque stability [ 48 ].…”

Section: Resultsmentioning

confidence: 99%

“…AI has also been applied to create a machine-learning tool using CTA images for evaluating intracranial internal carotid artery stenosis in patients with acute ischemic stroke. The machine-learning tool called the StrokeSENS LVO model was created using CTA images to evaluate intracranial internal carotid artery (ICA) stenosis in patients with acute ischemic stroke [ 47 ]. The algorithm showed high accuracy, sensitivity, and specificity in detecting intracranial ICA occlusion, without differences in patient age, sex, or CTA acquisition characteristics.…”

Section: Discussionmentioning

confidence: 99%

See 1 more Smart Citation

Artificial Intelligence in Acute Ischemic Stroke Subtypes According to Toast Classification: A Comprehensive Narrative Review

et al. 2023

View full text Add to dashboard Cite

The correct recognition of the etiology of ischemic stroke (IS) allows tempestive interventions in therapy with the aim of treating the cause and preventing a new cerebral ischemic event. Nevertheless, the identification of the cause is often challenging and is based on clinical features and data obtained by imaging techniques and other diagnostic exams. TOAST classification system describes the different etiologies of ischemic stroke and includes five subtypes: LAAS (large-artery atherosclerosis), CEI (cardio embolism), SVD (small vessel disease), ODE (stroke of other determined etiology), and UDE (stroke of undetermined etiology). AI models, providing computational methodologies for quantitative and objective evaluations, seem to increase the sensitivity of main IS causes, such as tomographic diagnosis of carotid stenosis, electrocardiographic recognition of atrial fibrillation, and identification of small vessel disease in magnetic resonance images. The aim of this review is to provide overall knowledge about the most effective AI models used in the differential diagnosis of ischemic stroke etiology according to the TOAST classification. According to our results, AI has proven to be a useful tool for identifying predictive factors capable of subtyping acute stroke patients in large heterogeneous populations and, in particular, clarifying the etiology of UDE IS especially detecting cardioembolic sources.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Discussionmentioning

confidence: 99%

Artificial Intelligence in Acute Ischemic Stroke Subtypes According to Toast Classification: A Comprehensive Narrative Review

et al. 2023

View full text Add to dashboard Cite

show abstract

“…DL has shown promise in developing models and software for diagnosing LVO ( Rava et al, 2021 ; Yahav-Dovrat et al, 2021 ; Cimflova et al, 2022 ; Czap et al, 2022 ; Seker et al, 2022 ), mostly based on monophasic CTA. However, 4D-CTA provides dynamic multi-phase scanning, offering temporal resolution for a comprehensive assessment of hemodynamic changes in AIS patients ( Frölich et al, 2014 ; Kortman et al, 2015 ).…”

Section: Discussionmentioning

confidence: 99%

Deep learning-assisted diagnosis of large vessel occlusion in acute ischemic stroke based on four-dimensional computed tomography angiography

Peng,

Liu,

Yao

et al. 2024

Front. Neurosci.

View full text Add to dashboard Cite

PurposeTo develop deep learning models based on four-dimensional computed tomography angiography (4D-CTA) images for automatic detection of large vessel occlusion (LVO) in the anterior circulation that cause acute ischemic stroke.MethodsThis retrospective study included 104 LVO patients and 105 non-LVO patients for deep learning models development. Another 30 LVO patients and 31 non-LVO patients formed the time-independent validation set. Four phases of 4D-CTA (arterial phase P1, arterial–venous phase P2, venous phase P3 and late venous phase P4) were arranged and combined and two input methods was used: combined input and superimposed input. Totally 26 models were constructed using a modified HRNet network. Assessment metrics included the areas under the curve (AUC), accuracy, sensitivity, specificity and F1 score. Kappa analysis was performed to assess inter-rater agreement between the best model and radiologists of different seniority.ResultsThe P1 + P2 model (combined input) had the best diagnostic performance. In the internal validation set, the AUC was 0.975 (95%CI: 0.878–0.999), accuracy was 0.911, sensitivity was 0.889, specificity was 0.944, and the F1 score was 0.909. In the time-independent validation set, the model demonstrated consistently high performance with an AUC of 0.942 (95%CI: 0.851–0.986), accuracy of 0.902, sensitivity of 0.867, specificity of 0.935, and an F1 score of 0.901. The best model showed strong consistency with the diagnostic efficacy of three radiologists of different seniority (k = 0.84, 0.80, 0.70, respectively).ConclusionThe deep learning model, using combined arterial and arterial–venous phase, was highly effective in detecting LVO, alerting radiologists to speed up the diagnosis.

show abstract

Machine learning applications for vascular stenosis detection in computed tomography angiography: a systematic review and meta-analysis

Anwer,

Karacan,

Enver

et al. 2024

Neural Comput & Applic

View full text Add to dashboard Cite

In an era in which cardiovascular disease has become the main cause of death all over the world, diagnostic accuracy in identifying blood vessels has become particularly important. Vascular stenosis causes serious health risks by affecting blood flow, leading to conditions like heart attacks and strokes. Traditional diagnostic methods face challenges in terms of timeliness and accuracy. Our systematic review aims to critically assess the role of machine learning (ML) techniques in enhancing computed tomography angiography’s (CTA) diagnostic capabilities for vasoconstriction. This review followed the predetermined inclusion and exclusion criteria and focused on research articles published between January 2013 and October 2023 collected from databases such as PubMed, IEEE, Web of Science, and Scopus. Studies focus on multiphase CTA or dynamic CTA; papers do not use the ML; and papers not in English are removed. The risk of bias of included studies was evaluated using the QUADAS2 tool. The results were analyzed in tabular form using metrics such as accuracy, sensitivity, and specificity and examine variations in stenosis detection by anatomical regions. In our review, a total of 63 studies were identified as relevant. These studies included a variety of ML applications for identifying anatomical stenosis of the arteries in different anatomical areas. The findings highlighted a trend of high sensitivity and specificity in broader anatomical assessments, with nuanced variations observed in detailed segmental analysis. The review acknowledges limitations within the existing studies, including the retrospective nature of most studies and their limited scope in terms of patient diversity and center variation. Nonetheless, the implications of integrating ML in vascular stenosis detection via CTA are profound, suggesting a pivotal shift toward more accurate, efficient, and patient-centric diagnostic practices in cardiovascular care.Registration: The protocol for this systematic review and meta-analysis was registered on PROSPERO, with the registration number CRD420234603.

show abstract

Validation of a machine learning software tool for automated large vessel occlusion detection in patients with suspected acute stroke

Cited by 5 publications

References 26 publications

Artificial Intelligence in Acute Ischemic Stroke Subtypes According to Toast Classification: A Comprehensive Narrative Review

Artificial Intelligence in Acute Ischemic Stroke Subtypes According to Toast Classification: A Comprehensive Narrative Review

Deep learning-assisted diagnosis of large vessel occlusion in acute ischemic stroke based on four-dimensional computed tomography angiography

Machine learning applications for vascular stenosis detection in computed tomography angiography: a systematic review and meta-analysis

Contact Info

Product

Resources

About