A dual-task mutual learning framework for predicting post-thrombectomy cerebral hemorrhage

Caiwen Jiang,Tianyu Wang,Xiaodan Xing,Mianxin Liu,Guang Yang,Zhongxiang Ding,Dinggang Shen

2024-08-02

Abstract:Ischemic stroke is a severe condition caused by the blockage of brain blood vessels, and can lead to the death of brain tissue due to oxygen deprivation. Thrombectomy has become a common treatment choice for ischemic stroke due to its immediate effectiveness. But, it carries the risk of postoperative cerebral hemorrhage. Clinically, multiple CT scans within 0-72 hours post-surgery are used to monitor for hemorrhage. However, this approach exposes radiation dose to patients, and may delay the detection of cerebral hemorrhage. To address this dilemma, we propose a novel prediction framework for measuring postoperative cerebral hemorrhage using only the patient's initial CT scan. Specifically, we introduce a dual-task mutual learning framework to takes the initial CT scan as input and simultaneously estimates both the follow-up CT scan and prognostic label to predict the occurrence of postoperative cerebral hemorrhage. Our proposed framework incorporates two attention mechanisms, i.e., self-attention and interactive attention. Specifically, the self-attention mechanism allows the model to focus more on high-density areas in the image, which are critical for diagnosis (i.e., potential hemorrhage areas). The interactive attention mechanism further models the dependencies between the interrelated generation and classification tasks, enabling both tasks to perform better than the case when conducted individually. Validated on clinical data, our method can generate follow-up CT scans better than state-of-the-art methods, and achieves an accuracy of 86.37% in predicting follow-up prognostic labels. Thus, our work thus contributes to the timely screening of post-thrombectomy cerebral hemorrhage, and could significantly reform the clinical process of thrombectomy and other similar operations related to stroke.

Image and Video Processing,Computer Vision and Pattern Recognition

What problem does this paper attempt to address?

The paper aims to address the issue of postoperative cerebral hemorrhage that may occur in patients with acute ischemic stroke after undergoing endovascular thrombectomy. Specifically, the goal of the study is to predict whether cerebral hemorrhage will occur within 0 to 72 hours post-surgery using only the patient's initial CT scan. This approach aims to avoid the radiation risks associated with multiple CT scans and to timely screen for hemorrhage. To achieve this goal, the researchers proposed a dual-task mutual learning framework that can simultaneously predict subsequent CT scan images and prognosis labels from the patient's initial CT scan. This system combines self-attention mechanisms and interactive attention mechanisms, enabling the model to focus more on high-density areas in the images (which are crucial for diagnosis) and to establish a connection between the two related tasks of generating subsequent CT images and predicting prognosis labels, thereby improving overall prediction performance. The main contributions of the paper include: 1. The first attempt to achieve early prediction of cerebral hemorrhage post-thrombectomy by estimating subsequent CT scans and prognosis labels. 2. The development of a novel dual-task mutual learning framework to perform this task. 3. Extensive experiments validating the effectiveness of the proposed method, showing that it outperforms existing techniques on the collected dataset.

A dual-task mutual learning framework for predicting post-thrombectomy cerebral hemorrhage

Prediction of large vessel occlusion for ischaemic stroke by using the machine learning model random forests

Deep learning for prediction of post-thrombectomy outcomes based on admission CT angiography in large vessel occlusion stroke

A New Deep Learning Framework for Accurate Intracranial Brain Hemorrhage Detection and Classification Using Real-Time Collected NCCT Images

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

Intracerebral Haemorrhage Growth Prediction Based on Displacement Vector Field and Clinical Metadata

A Deep Learning-Based Framework for Predicting Intracerebral Hemorrhage Hematoma Expansion Using Head Non-contrast CT Scan

Pairwise hemorrhage-brain region interaction-driven hemorrhagic stroke assessment in CT

Prediction of Hemorrhagic Complication after Thrombolytic Therapy Based on Multimodal Data from Multiple Centers: An Approach to Machine Learning and System Implementation.

Automated Hematoma Detection and Outcome Prediction in Patients With Traumatic Brain Injury

Predicting Thrombectomy Recanalization from CT Imaging Using Deep Learning Models

Prediction of Thrombectomy Functional Outcomes using Multimodal Data

Deep Learning for Hemorrhagic Lesion Detection and Segmentation on Brain CT Images

Deep-Learning Tool for Early Identifying Non-Traumatic Intracranial Hemorrhage Etiology based on CT Scan

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

A machine learning model based on results of a comprehensive radiological evaluation can predict the prognosis of basal ganglia cerebral hemorrhage treated with neuroendoscopy

A deep learning-based model for prediction of hemorrhagic transformation after stroke

Multi-task Learning Approach for Intracranial Hemorrhage Prognosis

Personalized Risk Prediction of Symptomatic Intracerebral Hemorrhage after Stroke Thrombolysis Using a Machine-Learning Model

Predicting Prognosis of Primary Pontine Hemorrhage Using CT Image and Deep Learning.

End to end stroke triage using cerebrovascular morphology and machine learning