Deep learning of tissue fate features in acute ischemic stroke

Abstract

In acute ischemic stroke treatment, prediction of tissue survival outcome plays a fundamental role in the clinical decision-making process, as it can be used to assess the balance of risk vs. possible benefit when considering endovascular clot-retrieval intervention. For the first time, we construct a deep learning model of tissue fate based on randomly sampled local patches from the hypoperfusion (Tmax) feature observed in MRI immediately after symptom onset. We evaluate the model with respect to the ground truth established by an expert neurologist four days after intervention. Experiments on 19 acute stroke patients evaluated the accuracy of the model in predicting tissue fate. Results show the superiority of the proposed regional learning framework versus a single-voxel-based regression model.

A Quantitative Method Using Head Noncontrast CT Scans to Detect Hyperacute Nonvisible Ischemic Changes in Patients With Stroke

“…Moreover, the localized cuboidal region on NCCT scans may serve as a basis for the application of computer vision and pattern recognition techniques, especially nonlinear regional models for accurate prediction of lesion size, location, and associated functional outcome . The majority of methods presented to date have been applied to MRI scans, such as artificial neural networks, multiclass support vector machines, or deep learning algorithms . Furthermore, the combination of image‐based features with clinical information may facilitate the accurate prediction of structural outcome in intracranial hemorrhage …”

“…26,27 The majority of methods presented to date have been applied to MRI scans, such as artificial neural networks, multiclass support vector machines, or deep learning algorithms. [28][29][30] Furthermore, the combination of image-based features with clinical information may facilitate the accurate prediction of structural outcome in intracranial hemorrhage. 31,32 We conclude that our computerized algorithm has the potential to support the decision to triage, and contribute to a reduction of time in stroke treatment.…”

Long Short-Term Memory Recurrent Neural Network for Stroke Prediction

“…In acute ischemic stroke treatment, the prediction of tissue survival outcome plays a fundamental role in the clinical decision-making process as it can be used to assess the balance of risk and possible benefit when endovascular c1otretrieval intervention is investigated. For the first time, Stier, Vincent, Liebeskind and Scalzo [11] constructed a deep learning model of tissue fate based on randomly sampled local patches from the hypoperfusion (Tmax) feature observed in MRI immediately after symptom onset. They evaluated the model with respect to the ground truth established by an expert neurologist four days after intervention.…”

Deep learning diffusion fingerprinting to detect brain tumour response to chemotherapy

“…Research into the use of deep learning with artificial neural networks (ANN) is being widely undertaken.A key application is in radiology, particularly in the classification and segmentation of biomedical images (1–4). ANNs have been trained and evaluated using x- ray computed tomography (CT) (5–7) and positron emission tomography (PET) imaging (8–10), and several magnetic resonance imaging (MRI) data types including structural T1- and T2-weighted image (11), perfusion images (12), MR spectroscopy (13) and diffusion MRI (14).…”