Abstract:Purpose:Investigate the threshold setting and accuracy of CTP in predicting the infarct core volume of acute ischemic stroke.Material and method:43 acute ischemic stroke patients admitted to the Department of Neurology of the First Hospital of Jilin university underwent Intra-arterial therapy from 2017.1 to 2017.12, and were retrospectively analyzed. Multimodal CT perfusion were examined prior to therapy, and non-contrast CT(NCCT) or MRI were reexamined 2-14 days after onset. Based on the mTICI following DSA intervention, the subjects were further categorized into recanalization positive group(mTICI 2b-3) and recanalization negative group(mTICI<2b). The infarct core volume was predicted via the MTT/CBV dual-threshold by CTP(Threshold setting: MTT>145%; 7 thresholds for CBV with an interval of 0.5, ranging between 2.0-5.0). The final infarct volume(FIV) was determined by follow up NCCT or MRI, and the difference of FIV was analyzed between the recanalization positive group and the recanalization negative group.Results:The FIV and the predicted infarct core volumes were compared among all patients. The predicted volumes by threshold 3 (Z=1.717, p=0.086), threshold 4(Z=0.110, p=0.912) and threshold 5(Z=1.678, p=0.093) were close to the FIV. However, the difference compared to FIV did not differ significantly among threshold 3,4 and 5. The predicted volumes by threshold 2 (Z=1.882, p=0.060), threshold 3(Z=0.154, p=0.877) and threshold 4(Z=1.471, p=0.141) were close to the FIV in the recanalization-positive group.The difference analysis of threshold 2, 3 and 4 revealed closest prediction of threshold 2, with a median volume of 32.61(19.20 similar to 61.74). However, the difference compared to FIV did not differ significantly among threshold 2, 3 and 4. The predicted volumes by threshold threshold 5(Z=1.156, p=0.248),threshold 6(Z=0.800, p=0.424) and threshold 7(Z=0.267, p=0.790) were close to the FIV in the recanalization-negative group. The difference analysis of threshold 5, 6 and 7 revealed closest prediction of threshold 7, with a median volume of 20.92(11.88 similar to 110.63). However, the difference compared to FIV did not differ significantly among threshold 5, 6 and 7.Conclusions: CTP predicted infarct core volume in AIS is highly correlated to FIV. More precise predictions might be achieved via adopting dynamic parameter thresholds based on the state of recanalization.

Fully Automatic Identification of Post-Treatment Infarct Lesions after Endovascular Therapy Based on Non-Contrast Computed Tomography

DCTP-Net: Dual-Branch CLIP-Enhance Textual Prompt-Aware Network for Acute Ischemic Stroke Lesion Segmentation From CT Image

Segmentation of Infarct Lesions and Prognosis Prediction for Acute Ischemic Stroke Using Non-Contrast CT Scans

Segmenting Ischemic Penumbra and Infarct Core Simultaneously on Non-Contrast CT of Patients with Acute Ischemic Stroke Using Novel Convolutional Neural Network

CNN Based Segmentation of Infarcted Regions in Acute Cerebral Stroke Patients From Computed Tomography Perfusion Imaging

Combining convolutional attention mechanism and residual deformable Transformer for infarct segmentation from CT scans of acute ischemic stroke patients

Deep learning-based identification of acute ischemic core and deficit from non-contrast CT and CTA

Automatic identification of early ischemic lesions on non-contrast CT with deep learning approach

Automatic segmentation of intracerebral hemorrhage in CT images using encoder–decoder convolutional neural network

Hybrid CNN-Transformer Network with Circular Feature Interaction for Acute Ischemic Stroke Lesion Segmentation on Non-contrast CT Scans

U-net Models Based on Computed Tomography Perfusion Predict Tissue Outcome in Patients with Different Reperfusion Patterns

Target-based deep learning network surveillance of non-contrast computed tomography for small infarct core of acute ischemic stroke

Multi-scale Deep Convolutional Neural Network for Stroke Lesions Segmentation on CT Images

Automatic segmentation of hemorrhagic transformation on follow-up non-contrast CT after acute ischemic stroke

Identification Of Invisible Ischemic Stroke In Noncontrast Ct Based On Novel Two-Stage Convolutional Neural Network Model

Detecting the Early Infarct Core on Non-Contrast CT Images with a Deep Learning Residual Network

Automated Segmentation of Ischemic Stroke Lesions in Non-Contrast Computed Tomography Images for Enhanced Treatment and Prognosis

Deep learning derived automated ASPECTS on non‐contrast CT scans of acute ischemic stroke patients

Automatic ischemic stroke lesion segmentation from computed tomography perfusion images by image synthesis and attention-based deep neural networks

End to end stroke triage using cerebrovascular morphology and machine learning

Application Of Image Segmentation On Evaluating Infarct Core In Acute Ischemic Stroke Using Ct Perfusion