Hello everyone!
\nI wanted to start a thread about articles and expand it, and hopefully others will share as well, articles to gain knowledge in the field. I have no knowledge in this area, so I downloaded the articles that seemed relevant after reading their summaries.
\nPapers:
\nClassification of subtype of acute ischemic stroke. Definitions for use in a multicenter clinical trial. TOAST. Trial of Org 10172 in Acute Stroke Treatment. - The etiology of ischemic stroke affects prognosis, outcome, and management. Trials of therapies for patients with acute stroke should include measurements of responses as influenced by subtype of ischemic stroke. A system for categorization of subtypes of ischemic stroke mainly based on etiology has been developed for the Trial of Org 10172 in Acute Stroke Treatment (TOAST). A classification of subtypes was prepared using clinical features and the results of ancillary diagnostic studies. \"Possible\" and \"probable\" diagnoses can be made based on the physician's certainty of diagnosis. The usefulness and interrater agreement of the classification were tested by two neurologists who had not participated in the writing of the criteria. The neurologists independently used the TOAST classification system in their bedside evaluation of 20 patients, first based only on clinical features and then after reviewing the results of diagnostic tests. The TOAST classification denotes five subtypes of ischemic stroke: 1) large-artery atherosclerosis, 2) cardioembolism, 3) small-vessel occlusion, 4) stroke of other determined etiology, and 5) stroke of undetermined etiology. Using this rating system, interphysician agreement was very high. The two physicians disagreed in only one patient. They were both able to reach a specific etiologic diagnosis in 11 patients, whereas the cause of stroke was not determined in nine. The TOAST stroke subtype classification system is easy to use and has good interobserver agreement. This system should allow investigators to report responses to treatment among important subgroups of patients with ischemic stroke. Clinical trials testing treatments for acute ischemic stroke should include similar methods to diagnose subtypes of stroke.
Orbit image analysis machine learning software can be used for the histological quantification of acute ischemic stroke blood clots - Our aim was to assess the utility of a novel machine learning software (Orbit Image Analysis) in the histological quantification of acute ischemic stroke (AIS) clots. We analyzed 50 AIS blood clots retrieved using mechanical thrombectomy procedures. Following H&E staining, quantification of clot components was performed by two different methods: a pathologist using a reference standard method (Adobe Photoshop CC) and an experienced researcher using Orbit Image Analysis. Following quantification, the clots were categorized into 3 types: RBC dominant (≥60% RBCs), Mixed and Fibrin dominant (≥60% Fibrin). Correlations between clot composition and Hounsfield Units density on Computed Tomography (CT) were assessed. There was a significant correlation between the components of clots as quantified by the Orbit Image Analysis algorithm and the reference standard approach (ρ = 0.944*, p < 0.001, n = 150). A significant relationship was found between clot composition (RBC-Rich, Mixed, Fibrin-Rich) and the presence of a Hyperdense artery sign using the algorithmic method (X^2(2) = 6.712, p = 0.035) but not using the reference standard method (X^2(2) = 3.924, p = 0.141). Orbit Image Analysis machine learning software can be used for the histological quantification of AIS clots, reproducibly generating composition analyses similar to current reference standard methods.
A Machine Learning Approach for Classifying Ischemic Stroke Onset Time From Imaging - Current clinical practice relies on clinical history to determine the time since stroke (TSS) onset. Imaging-based determination of acute stroke onset time could provide critical information to clinicians in deciding stroke treatment options, such as thrombolysis. The patients with unknown or unwitnessed TSS are usually excluded from thrombolysis, even if their symptoms began within the therapeutic window. In this paper, we demonstrate a machine learning approach for TSS classification using routinely acquired imaging sequences. We develop imaging features from the magnetic resonance (MR) images and train machine learning models to classify the TSS. We also propose a deep-learning model to extract hidden representations for the MR perfusion-weighted images and demonstrate classification improvement by incorporating these additional deep features. The cross-validation results show that our best classifier achieved an area under the curve of 0.765, with a sensitivity of 0.788 and a negative predictive value of 0.609, outperforming existing methods. We show that the features generated by our deep-learning algorithm correlate with the MR imaging features, and validate the robustness of the model on imaging parameter variations (e.g., year of imaging). This paper advances magnetic resonance imaging analysis one-step-closer to an operational decision support tool for stroke treatment guidance.
Machine Learning–Enabled Automated Determination of Acute Ischemic Core From Computed Tomography Angiography - The availability of and expertise to interpret advanced neuro-imaging recommended in the guideline-based endovascular stroke therapy (EST) evaluation are limited. Here, we develop and validate an automated machine learning-based method that evaluates for large vessel occlusion (LVO) and ischemic core volume in patients using a widely available modality, computed tomography angiogram (CTA). From our prospectively maintained stroke registry and electronic medical record, we identified patients with acute ischemic stroke and stroke mimics with contemporaneous CTA and computed tomography perfusion (CTP) with RAPID (IschemaView) post-processing as a part of the emergent stroke workup. A novel convolutional neural network named DeepSymNet was created and trained to identify LVO as well as infarct core from CTA source images, against CTP-RAPID definitions. Model performance was measured using 10-fold cross validation and receiver-operative curve area under the curve (AUC) statistics. Among the 297 included patients, 224 (75%) had acute ischemic stroke of which 179 (60%) had LVO. Mean CTP-RAPID ischemic core volume was 23±42 mL. LVO locations included internal carotid artery (13%), M1 (44%), and M2 (21%). The DeepSymNet algorithm autonomously learned to identify the intracerebral vasculature on CTA and detected LVO with AUC 0.88. The method was also able to determine infarct core as defined by CTP-RAPID from the CTA source images with AUC 0.88 and 0.90 (ischemic core ≤30 mL and ≤50 mL). These findings were maintained in patients presenting in early (0–6 hours) and late (6–24 hours) time windows (AUCs 0.90 and 0.91, ischemic core ≤50 mL). DeepSymNet probabilities from CTA images corresponded with CTP-RAPID ischemic core volumes as a continuous variable with r=0.7 (Pearson correlation, P<0.001). These results demonstrate that the information needed to perform the neuroimaging evaluation for endovascular therapy with comparable accuracy to advanced imaging modalities may be present in CTA, and the ability of machine learning to automate the analysis.
Structural analysis of ischemic stroke thrombi: histological indications for therapy resistance - Ischemic stroke is caused by a thromboembolic occlusion of cerebral arteries. Treatment is focused on fast and efficient removal of the occluding thrombus, either via intravenous thrombolysis or via endovascular thrombectomy. Recanalization, however, is not always successful and factors contributing to failure are not completely understood. Although the occluding thrombus is the primary target of acute treatment, little is known about its internal organization and composition. The aim of this study, therefore, was to better understand the internal organization of ischemic stroke thrombi on a molecular and cellular level. A total of 188 thrombi were collected from endovascularly treated ischemic stroke patients and analyzed histologically for fibrin, red blood cells (RBC), von Willebrand factor (vWF), platelets, leukocytes and DNA, using bright field and fluorescence microscopy. Our results show that stroke thrombi are composed of two main types of areas: RBC-rich areas and platelet-rich areas. RBC-rich areas have limited complexity as they consist of RBC that are entangled in a meshwork of thin fibrin. In contrast, platelet-rich areas are characterized by dense fibrin structures aligned with vWF and abundant amounts of leukocytes and DNA that accumulate around and in these platelet-rich areas. These findings are important to better understand why platelet-rich thrombi are resistant to thrombolysis and difficult to retrieve via thrombectomy, and can guide further improvements of acute ischemic stroke therapy.
Pages:
\n\nHave a good competition and don't hesitate to comment!
", "messages": [ { "id": 1846573, "postDate": "2022-07-07T07:03:24.017Z", "content": "Hello everyone!
\nI wanted to start a thread about articles and expand it, and hopefully others will share as well, articles to gain knowledge in the field. I have no knowledge in this area, so I downloaded the articles that seemed relevant after reading their summaries.
\nPapers:
\nClassification of subtype of acute ischemic stroke. Definitions for use in a multicenter clinical trial. TOAST. Trial of Org 10172 in Acute Stroke Treatment. - The etiology of ischemic stroke affects prognosis, outcome, and management. Trials of therapies for patients with acute stroke should include measurements of responses as influenced by subtype of ischemic stroke. A system for categorization of subtypes of ischemic stroke mainly based on etiology has been developed for the Trial of Org 10172 in Acute Stroke Treatment (TOAST). A classification of subtypes was prepared using clinical features and the results of ancillary diagnostic studies. \"Possible\" and \"probable\" diagnoses can be made based on the physician's certainty of diagnosis. The usefulness and interrater agreement of the classification were tested by two neurologists who had not participated in the writing of the criteria. The neurologists independently used the TOAST classification system in their bedside evaluation of 20 patients, first based only on clinical features and then after reviewing the results of diagnostic tests. The TOAST classification denotes five subtypes of ischemic stroke: 1) large-artery atherosclerosis, 2) cardioembolism, 3) small-vessel occlusion, 4) stroke of other determined etiology, and 5) stroke of undetermined etiology. Using this rating system, interphysician agreement was very high. The two physicians disagreed in only one patient. They were both able to reach a specific etiologic diagnosis in 11 patients, whereas the cause of stroke was not determined in nine. The TOAST stroke subtype classification system is easy to use and has good interobserver agreement. This system should allow investigators to report responses to treatment among important subgroups of patients with ischemic stroke. Clinical trials testing treatments for acute ischemic stroke should include similar methods to diagnose subtypes of stroke.
Orbit image analysis machine learning software can be used for the histological quantification of acute ischemic stroke blood clots - Our aim was to assess the utility of a novel machine learning software (Orbit Image Analysis) in the histological quantification of acute ischemic stroke (AIS) clots. We analyzed 50 AIS blood clots retrieved using mechanical thrombectomy procedures. Following H&E staining, quantification of clot components was performed by two different methods: a pathologist using a reference standard method (Adobe Photoshop CC) and an experienced researcher using Orbit Image Analysis. Following quantification, the clots were categorized into 3 types: RBC dominant (≥60% RBCs), Mixed and Fibrin dominant (≥60% Fibrin). Correlations between clot composition and Hounsfield Units density on Computed Tomography (CT) were assessed. There was a significant correlation between the components of clots as quantified by the Orbit Image Analysis algorithm and the reference standard approach (ρ = 0.944*, p < 0.001, n = 150). A significant relationship was found between clot composition (RBC-Rich, Mixed, Fibrin-Rich) and the presence of a Hyperdense artery sign using the algorithmic method (X^2(2) = 6.712, p = 0.035) but not using the reference standard method (X^2(2) = 3.924, p = 0.141). Orbit Image Analysis machine learning software can be used for the histological quantification of AIS clots, reproducibly generating composition analyses similar to current reference standard methods.
A Machine Learning Approach for Classifying Ischemic Stroke Onset Time From Imaging - Current clinical practice relies on clinical history to determine the time since stroke (TSS) onset. Imaging-based determination of acute stroke onset time could provide critical information to clinicians in deciding stroke treatment options, such as thrombolysis. The patients with unknown or unwitnessed TSS are usually excluded from thrombolysis, even if their symptoms began within the therapeutic window. In this paper, we demonstrate a machine learning approach for TSS classification using routinely acquired imaging sequences. We develop imaging features from the magnetic resonance (MR) images and train machine learning models to classify the TSS. We also propose a deep-learning model to extract hidden representations for the MR perfusion-weighted images and demonstrate classification improvement by incorporating these additional deep features. The cross-validation results show that our best classifier achieved an area under the curve of 0.765, with a sensitivity of 0.788 and a negative predictive value of 0.609, outperforming existing methods. We show that the features generated by our deep-learning algorithm correlate with the MR imaging features, and validate the robustness of the model on imaging parameter variations (e.g., year of imaging). This paper advances magnetic resonance imaging analysis one-step-closer to an operational decision support tool for stroke treatment guidance.
Machine Learning–Enabled Automated Determination of Acute Ischemic Core From Computed Tomography Angiography - The availability of and expertise to interpret advanced neuro-imaging recommended in the guideline-based endovascular stroke therapy (EST) evaluation are limited. Here, we develop and validate an automated machine learning-based method that evaluates for large vessel occlusion (LVO) and ischemic core volume in patients using a widely available modality, computed tomography angiogram (CTA). From our prospectively maintained stroke registry and electronic medical record, we identified patients with acute ischemic stroke and stroke mimics with contemporaneous CTA and computed tomography perfusion (CTP) with RAPID (IschemaView) post-processing as a part of the emergent stroke workup. A novel convolutional neural network named DeepSymNet was created and trained to identify LVO as well as infarct core from CTA source images, against CTP-RAPID definitions. Model performance was measured using 10-fold cross validation and receiver-operative curve area under the curve (AUC) statistics. Among the 297 included patients, 224 (75%) had acute ischemic stroke of which 179 (60%) had LVO. Mean CTP-RAPID ischemic core volume was 23±42 mL. LVO locations included internal carotid artery (13%), M1 (44%), and M2 (21%). The DeepSymNet algorithm autonomously learned to identify the intracerebral vasculature on CTA and detected LVO with AUC 0.88. The method was also able to determine infarct core as defined by CTP-RAPID from the CTA source images with AUC 0.88 and 0.90 (ischemic core ≤30 mL and ≤50 mL). These findings were maintained in patients presenting in early (0–6 hours) and late (6–24 hours) time windows (AUCs 0.90 and 0.91, ischemic core ≤50 mL). DeepSymNet probabilities from CTA images corresponded with CTP-RAPID ischemic core volumes as a continuous variable with r=0.7 (Pearson correlation, P<0.001). These results demonstrate that the information needed to perform the neuroimaging evaluation for endovascular therapy with comparable accuracy to advanced imaging modalities may be present in CTA, and the ability of machine learning to automate the analysis.
Structural analysis of ischemic stroke thrombi: histological indications for therapy resistance - Ischemic stroke is caused by a thromboembolic occlusion of cerebral arteries. Treatment is focused on fast and efficient removal of the occluding thrombus, either via intravenous thrombolysis or via endovascular thrombectomy. Recanalization, however, is not always successful and factors contributing to failure are not completely understood. Although the occluding thrombus is the primary target of acute treatment, little is known about its internal organization and composition. The aim of this study, therefore, was to better understand the internal organization of ischemic stroke thrombi on a molecular and cellular level. A total of 188 thrombi were collected from endovascularly treated ischemic stroke patients and analyzed histologically for fibrin, red blood cells (RBC), von Willebrand factor (vWF), platelets, leukocytes and DNA, using bright field and fluorescence microscopy. Our results show that stroke thrombi are composed of two main types of areas: RBC-rich areas and platelet-rich areas. RBC-rich areas have limited complexity as they consist of RBC that are entangled in a meshwork of thin fibrin. In contrast, platelet-rich areas are characterized by dense fibrin structures aligned with vWF and abundant amounts of leukocytes and DNA that accumulate around and in these platelet-rich areas. These findings are important to better understand why platelet-rich thrombi are resistant to thrombolysis and difficult to retrieve via thrombectomy, and can guide further improvements of acute ischemic stroke therapy.
Pages:
\n\nHave a good competition and don't hesitate to comment!
", "rawMarkdown": "Hello everyone!\n\nI wanted to start a thread about articles and expand it, and hopefully others will share as well, articles to gain knowledge in the field. I have no knowledge in this area, so I downloaded the articles that seemed relevant after reading their summaries.\n\n**Papers:**\n\n- [Classification of subtype of acute ischemic stroke. Definitions for use in a multicenter clinical trial. TOAST. Trial of Org 10172 in Acute Stroke Treatment.](https://www.ahajournals.org/doi/10.1161/01.str.24.1.35) - The etiology of ischemic stroke affects prognosis, outcome, and management. Trials of therapies for patients with acute stroke should include measurements of responses as influenced by subtype of ischemic stroke. A system for categorization of subtypes of ischemic stroke mainly based on etiology has been developed for the Trial of Org 10172 in Acute Stroke Treatment (TOAST). A classification of subtypes was prepared using clinical features and the results of ancillary diagnostic studies. \"Possible\" and \"probable\" diagnoses can be made based on the physician's certainty of diagnosis. The usefulness and interrater agreement of the classification were tested by two neurologists who had not participated in the writing of the criteria. The neurologists independently used the TOAST classification system in their bedside evaluation of 20 patients, first based only on clinical features and then after reviewing the results of diagnostic tests. The TOAST classification denotes five subtypes of ischemic stroke: 1) large-artery atherosclerosis, 2) cardioembolism, 3) small-vessel occlusion, 4) stroke of other determined etiology, and 5) stroke of undetermined etiology. Using this rating system, interphysician agreement was very high. The two physicians disagreed in only one patient. They were both able to reach a specific etiologic diagnosis in 11 patients, whereas the cause of stroke was not determined in nine. The TOAST stroke subtype classification system is easy to use and has good interobserver agreement. This system should allow investigators to report responses to treatment among important subgroups of patients with ischemic stroke. Clinical trials testing treatments for acute ischemic stroke should include similar methods to diagnose subtypes of stroke.\n\n- [Orbit image analysis machine learning software can be used for the histological quantification of acute ischemic stroke blood clots](https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0225841) - Our aim was to assess the utility of a novel machine learning software (Orbit Image Analysis) in the histological quantification of acute ischemic stroke (AIS) clots. We analyzed 50 AIS blood clots retrieved using mechanical thrombectomy procedures. Following H&E staining, quantification of clot components was performed by two different methods: a pathologist using a reference standard method (Adobe Photoshop CC) and an experienced researcher using Orbit Image Analysis. Following quantification, the clots were categorized into 3 types: RBC dominant (≥60% RBCs), Mixed and Fibrin dominant (≥60% Fibrin). Correlations between clot composition and Hounsfield Units density on Computed Tomography (CT) were assessed. There was a significant correlation between the components of clots as quantified by the Orbit Image Analysis algorithm and the reference standard approach (ρ = 0.944**, p < 0.001, n = 150). A significant relationship was found between clot composition (RBC-Rich, Mixed, Fibrin-Rich) and the presence of a Hyperdense artery sign using the algorithmic method (X^2(2) = 6.712, p = 0.035*) but not using the reference standard method (X^2(2) = 3.924, p = 0.141). Orbit Image Analysis machine learning software can be used for the histological quantification of AIS clots, reproducibly generating composition analyses similar to current reference standard methods.\n\n- [A Machine Learning Approach for Classifying Ischemic Stroke Onset Time From Imaging](https://ieeexplore-ieee-org.ezproxy.uphf.fr/document/8651325) - Current clinical practice relies on clinical history to determine the time since stroke (TSS) onset. Imaging-based determination of acute stroke onset time could provide critical information to clinicians in deciding stroke treatment options, such as thrombolysis. The patients with unknown or unwitnessed TSS are usually excluded from thrombolysis, even if their symptoms began within the therapeutic window. In this paper, we demonstrate a machine learning approach for TSS classification using routinely acquired imaging sequences. We develop imaging features from the magnetic resonance (MR) images and train machine learning models to classify the TSS. We also propose a deep-learning model to extract hidden representations for the MR perfusion-weighted images and demonstrate classification improvement by incorporating these additional deep features. The cross-validation results show that our best classifier achieved an area under the curve of 0.765, with a sensitivity of 0.788 and a negative predictive value of 0.609, outperforming existing methods. We show that the features generated by our deep-learning algorithm correlate with the MR imaging features, and validate the robustness of the model on imaging parameter variations (e.g., year of imaging). This paper advances magnetic resonance imaging analysis one-step-closer to an operational decision support tool for stroke treatment guidance.\n\n- [Machine Learning–Enabled Automated Determination of Acute Ischemic Core From Computed Tomography Angiography](https://www.ahajournals.org/doi/10.1161/STROKEAHA.119.026189) - The availability of and expertise to interpret advanced neuro-imaging recommended in the guideline-based endovascular stroke therapy (EST) evaluation are limited. Here, we develop and validate an automated machine learning-based method that evaluates for large vessel occlusion (LVO) and ischemic core volume in patients using a widely available modality, computed tomography angiogram (CTA). From our prospectively maintained stroke registry and electronic medical record, we identified patients with acute ischemic stroke and stroke mimics with contemporaneous CTA and computed tomography perfusion (CTP) with RAPID (IschemaView) post-processing as a part of the emergent stroke workup. A novel convolutional neural network named DeepSymNet was created and trained to identify LVO as well as infarct core from CTA source images, against CTP-RAPID definitions. Model performance was measured using 10-fold cross validation and receiver-operative curve area under the curve (AUC) statistics. Among the 297 included patients, 224 (75%) had acute ischemic stroke of which 179 (60%) had LVO. Mean CTP-RAPID ischemic core volume was 23±42 mL. LVO locations included internal carotid artery (13%), M1 (44%), and M2 (21%). The DeepSymNet algorithm autonomously learned to identify the intracerebral vasculature on CTA and detected LVO with AUC 0.88. The method was also able to determine infarct core as defined by CTP-RAPID from the CTA source images with AUC 0.88 and 0.90 (ischemic core ≤30 mL and ≤50 mL). These findings were maintained in patients presenting in early (0–6 hours) and late (6–24 hours) time windows (AUCs 0.90 and 0.91, ischemic core ≤50 mL). DeepSymNet probabilities from CTA images corresponded with CTP-RAPID ischemic core volumes as a continuous variable with r=0.7 (Pearson correlation, P<0.001). These results demonstrate that the information needed to perform the neuroimaging evaluation for endovascular therapy with comparable accuracy to advanced imaging modalities may be present in CTA, and the ability of machine learning to automate the analysis.\n\n- [Structural analysis of ischemic stroke thrombi: histological indications for therapy resistance](https://www.haematologica.org/article/view/9534) - Ischemic stroke is caused by a thromboembolic occlusion of cerebral arteries. Treatment is focused on fast and efficient removal of the occluding thrombus, either via intravenous thrombolysis or via endovascular thrombectomy. Recanalization, however, is not always successful and factors contributing to failure are not completely understood. Although the occluding thrombus is the primary target of acute treatment, little is known about its internal organization and composition. The aim of this study, therefore, was to better understand the internal organization of ischemic stroke thrombi on a molecular and cellular level. A total of 188 thrombi were collected from endovascularly treated ischemic stroke patients and analyzed histologically for fibrin, red blood cells (RBC), von Willebrand factor (vWF), platelets, leukocytes and DNA, using bright field and fluorescence microscopy. Our results show that stroke thrombi are composed of two main types of areas: RBC-rich areas and platelet-rich areas. RBC-rich areas have limited complexity as they consist of RBC that are entangled in a meshwork of thin fibrin. In contrast, platelet-rich areas are characterized by dense fibrin structures aligned with vWF and abundant amounts of leukocytes and DNA that accumulate around and in these platelet-rich areas. These findings are important to better understand why platelet-rich thrombi are resistant to thrombolysis and difficult to retrieve via thrombectomy, and can guide further improvements of acute ischemic stroke therapy.\n\n**Pages:**\n- [Ischemic Stroke (Clots)](https://www.stroke.org/en/about-stroke/types-of-stroke/ischemic-stroke-clots)\n\n**Have a good competition and don't hesitate to comment!**", "votes": 42 }, { "id": 1861817, "postDate": "2022-07-19T09:05:15.940Z", "content": "Great resource! Thanks for sharing. Do you have something on federated learning? Thanks!
", "rawMarkdown": "Great resource! Thanks for sharing. Do you have something on federated learning? Thanks!", "votes": 1 }, { "id": 1847097, "postDate": "2022-07-07T16:22:04.093Z", "content": "Thanks for sharing! its great and helpful
", "rawMarkdown": "Thanks for sharing! its great and helpful", "votes": 2 }, { "id": 1846762, "postDate": "2022-07-07T10:36:59.750Z", "content": "Great compilation! Thanks for sharing it with the community!
", "rawMarkdown": "Great compilation! Thanks for sharing it with the community!", "votes": 2 }, { "id": 1861425, "postDate": "2022-07-19T02:36:39.863Z", "content": "Thanks for sharing
", "rawMarkdown": "Thanks for sharing ", "votes": 1 }, { "id": 1849421, "postDate": "2022-07-09T13:50:47.437Z", "content": "Great! Thanks for sharing 👍
", "rawMarkdown": "Great! Thanks for sharing 👍", "votes": 1 }, { "id": 1846712, "postDate": "2022-07-07T09:48:27.353Z", "content": "Good resources thank you.
", "rawMarkdown": "Good resources thank you.", "votes": 1 }, { "id": 1914299, "postDate": "2022-08-26T01:15:13.610Z", "content": "Thanks for sharing, it's helpful.
", "rawMarkdown": "Thanks for sharing, it's helpful.", "votes": 2 }, { "id": 1846748, "postDate": "2022-07-07T10:21:53.333Z", "content": "@datascientistfp , thanks for sharing👍
", "rawMarkdown": "@datascientistfp , thanks for sharing👍", "votes": 2 } ], "comments": [ { "id": 1861817, "author_name": "Felipe Dutra", "author_url": "", "post_date": "2022-07-19T09:05:15.940000", "content": "Great resource! Thanks for sharing. Do you have something on federated learning? Thanks!
", "votes": 1, "replies": [] }, { "id": 1847097, "author_name": "VK", "author_url": "", "post_date": "2022-07-07T16:22:04.093000", "content": "Thanks for sharing! its great and helpful
", "votes": 2, "replies": [] }, { "id": 1846762, "author_name": "Subhajit", "author_url": "", "post_date": "2022-07-07T10:36:59.750000", "content": "Great compilation! Thanks for sharing it with the community!
", "votes": 2, "replies": [] }, { "id": 1861425, "author_name": "ProverbW", "author_url": "", "post_date": "2022-07-19T02:36:39.863000", "content": "Thanks for sharing
", "votes": 1, "replies": [] }, { "id": 1849421, "author_name": "Manikanth", "author_url": "", "post_date": "2022-07-09T13:50:47.437000", "content": "Great! Thanks for sharing 👍
", "votes": 1, "replies": [] }, { "id": 1846712, "author_name": "Med Ali Bouchhioua", "author_url": "", "post_date": "2022-07-07T09:48:27.353000", "content": "Good resources thank you.
", "votes": 1, "replies": [] }, { "id": 1914299, "author_name": "guansuo", "author_url": "", "post_date": "2022-08-26T01:15:13.610000", "content": "Thanks for sharing, it's helpful.
", "votes": 2, "replies": [] }, { "id": 1846748, "author_name": "HASIB AL MUZDADID", "author_url": "", "post_date": "2022-07-07T10:21:53.333000", "content": "@datascientistfp , thanks for sharing👍
", "votes": 2, "replies": [] } ], "raw_markdown_by_id": { "1846573": "Hello everyone!\n\nI wanted to start a thread about articles and expand it, and hopefully others will share as well, articles to gain knowledge in the field. I have no knowledge in this area, so I downloaded the articles that seemed relevant after reading their summaries.\n\n**Papers:**\n\n- [Classification of subtype of acute ischemic stroke. Definitions for use in a multicenter clinical trial. TOAST. Trial of Org 10172 in Acute Stroke Treatment.](https://www.ahajournals.org/doi/10.1161/01.str.24.1.35) - The etiology of ischemic stroke affects prognosis, outcome, and management. Trials of therapies for patients with acute stroke should include measurements of responses as influenced by subtype of ischemic stroke. A system for categorization of subtypes of ischemic stroke mainly based on etiology has been developed for the Trial of Org 10172 in Acute Stroke Treatment (TOAST). A classification of subtypes was prepared using clinical features and the results of ancillary diagnostic studies. \"Possible\" and \"probable\" diagnoses can be made based on the physician's certainty of diagnosis. The usefulness and interrater agreement of the classification were tested by two neurologists who had not participated in the writing of the criteria. The neurologists independently used the TOAST classification system in their bedside evaluation of 20 patients, first based only on clinical features and then after reviewing the results of diagnostic tests. The TOAST classification denotes five subtypes of ischemic stroke: 1) large-artery atherosclerosis, 2) cardioembolism, 3) small-vessel occlusion, 4) stroke of other determined etiology, and 5) stroke of undetermined etiology. Using this rating system, interphysician agreement was very high. The two physicians disagreed in only one patient. They were both able to reach a specific etiologic diagnosis in 11 patients, whereas the cause of stroke was not determined in nine. The TOAST stroke subtype classification system is easy to use and has good interobserver agreement. This system should allow investigators to report responses to treatment among important subgroups of patients with ischemic stroke. Clinical trials testing treatments for acute ischemic stroke should include similar methods to diagnose subtypes of stroke.\n\n- [Orbit image analysis machine learning software can be used for the histological quantification of acute ischemic stroke blood clots](https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0225841) - Our aim was to assess the utility of a novel machine learning software (Orbit Image Analysis) in the histological quantification of acute ischemic stroke (AIS) clots. We analyzed 50 AIS blood clots retrieved using mechanical thrombectomy procedures. Following H&E staining, quantification of clot components was performed by two different methods: a pathologist using a reference standard method (Adobe Photoshop CC) and an experienced researcher using Orbit Image Analysis. Following quantification, the clots were categorized into 3 types: RBC dominant (≥60% RBCs), Mixed and Fibrin dominant (≥60% Fibrin). Correlations between clot composition and Hounsfield Units density on Computed Tomography (CT) were assessed. There was a significant correlation between the components of clots as quantified by the Orbit Image Analysis algorithm and the reference standard approach (ρ = 0.944**, p < 0.001, n = 150). A significant relationship was found between clot composition (RBC-Rich, Mixed, Fibrin-Rich) and the presence of a Hyperdense artery sign using the algorithmic method (X^2(2) = 6.712, p = 0.035*) but not using the reference standard method (X^2(2) = 3.924, p = 0.141). Orbit Image Analysis machine learning software can be used for the histological quantification of AIS clots, reproducibly generating composition analyses similar to current reference standard methods.\n\n- [A Machine Learning Approach for Classifying Ischemic Stroke Onset Time From Imaging](https://ieeexplore-ieee-org.ezproxy.uphf.fr/document/8651325) - Current clinical practice relies on clinical history to determine the time since stroke (TSS) onset. Imaging-based determination of acute stroke onset time could provide critical information to clinicians in deciding stroke treatment options, such as thrombolysis. The patients with unknown or unwitnessed TSS are usually excluded from thrombolysis, even if their symptoms began within the therapeutic window. In this paper, we demonstrate a machine learning approach for TSS classification using routinely acquired imaging sequences. We develop imaging features from the magnetic resonance (MR) images and train machine learning models to classify the TSS. We also propose a deep-learning model to extract hidden representations for the MR perfusion-weighted images and demonstrate classification improvement by incorporating these additional deep features. The cross-validation results show that our best classifier achieved an area under the curve of 0.765, with a sensitivity of 0.788 and a negative predictive value of 0.609, outperforming existing methods. We show that the features generated by our deep-learning algorithm correlate with the MR imaging features, and validate the robustness of the model on imaging parameter variations (e.g., year of imaging). This paper advances magnetic resonance imaging analysis one-step-closer to an operational decision support tool for stroke treatment guidance.\n\n- [Machine Learning–Enabled Automated Determination of Acute Ischemic Core From Computed Tomography Angiography](https://www.ahajournals.org/doi/10.1161/STROKEAHA.119.026189) - The availability of and expertise to interpret advanced neuro-imaging recommended in the guideline-based endovascular stroke therapy (EST) evaluation are limited. Here, we develop and validate an automated machine learning-based method that evaluates for large vessel occlusion (LVO) and ischemic core volume in patients using a widely available modality, computed tomography angiogram (CTA). From our prospectively maintained stroke registry and electronic medical record, we identified patients with acute ischemic stroke and stroke mimics with contemporaneous CTA and computed tomography perfusion (CTP) with RAPID (IschemaView) post-processing as a part of the emergent stroke workup. A novel convolutional neural network named DeepSymNet was created and trained to identify LVO as well as infarct core from CTA source images, against CTP-RAPID definitions. Model performance was measured using 10-fold cross validation and receiver-operative curve area under the curve (AUC) statistics. Among the 297 included patients, 224 (75%) had acute ischemic stroke of which 179 (60%) had LVO. Mean CTP-RAPID ischemic core volume was 23±42 mL. LVO locations included internal carotid artery (13%), M1 (44%), and M2 (21%). The DeepSymNet algorithm autonomously learned to identify the intracerebral vasculature on CTA and detected LVO with AUC 0.88. The method was also able to determine infarct core as defined by CTP-RAPID from the CTA source images with AUC 0.88 and 0.90 (ischemic core ≤30 mL and ≤50 mL). These findings were maintained in patients presenting in early (0–6 hours) and late (6–24 hours) time windows (AUCs 0.90 and 0.91, ischemic core ≤50 mL). DeepSymNet probabilities from CTA images corresponded with CTP-RAPID ischemic core volumes as a continuous variable with r=0.7 (Pearson correlation, P<0.001). These results demonstrate that the information needed to perform the neuroimaging evaluation for endovascular therapy with comparable accuracy to advanced imaging modalities may be present in CTA, and the ability of machine learning to automate the analysis.\n\n- [Structural analysis of ischemic stroke thrombi: histological indications for therapy resistance](https://www.haematologica.org/article/view/9534) - Ischemic stroke is caused by a thromboembolic occlusion of cerebral arteries. Treatment is focused on fast and efficient removal of the occluding thrombus, either via intravenous thrombolysis or via endovascular thrombectomy. Recanalization, however, is not always successful and factors contributing to failure are not completely understood. Although the occluding thrombus is the primary target of acute treatment, little is known about its internal organization and composition. The aim of this study, therefore, was to better understand the internal organization of ischemic stroke thrombi on a molecular and cellular level. A total of 188 thrombi were collected from endovascularly treated ischemic stroke patients and analyzed histologically for fibrin, red blood cells (RBC), von Willebrand factor (vWF), platelets, leukocytes and DNA, using bright field and fluorescence microscopy. Our results show that stroke thrombi are composed of two main types of areas: RBC-rich areas and platelet-rich areas. RBC-rich areas have limited complexity as they consist of RBC that are entangled in a meshwork of thin fibrin. In contrast, platelet-rich areas are characterized by dense fibrin structures aligned with vWF and abundant amounts of leukocytes and DNA that accumulate around and in these platelet-rich areas. These findings are important to better understand why platelet-rich thrombi are resistant to thrombolysis and difficult to retrieve via thrombectomy, and can guide further improvements of acute ischemic stroke therapy.\n\n**Pages:**\n- [Ischemic Stroke (Clots)](https://www.stroke.org/en/about-stroke/types-of-stroke/ischemic-stroke-clots)\n\n**Have a good competition and don't hesitate to comment!**", "1861817": "Great resource! Thanks for sharing. Do you have something on federated learning? Thanks!", "1847097": "Thanks for sharing! its great and helpful", "1846762": "Great compilation! Thanks for sharing it with the community!", "1861425": "Thanks for sharing ", "1849421": "Great! Thanks for sharing 👍", "1846712": "Good resources thank you.", "1914299": "Thanks for sharing, it's helpful.", "1846748": "@datascientistfp , thanks for sharing👍" } }