Brain Lesion Prediction Models for Recanalization Therapy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for estimating the penumbra zone in cerebral ischemia, relying on pre-treatment diffusion weighted image (DWI) and perfusion weighted image (PWI) data, have limitations as infarct progression can vary significantly between individuals, making it challenging to predict treatment outcomes for recanalization therapy.
Innovation Solution
A machine learning-based method that uses brain image data from previous patients to learn prediction models for successful and unsuccessful recanalization treatments, generating output image data on lesion distribution to inform treatment decisions, including the use of diffusion weighted image (DWI) and perfusion weighted image (PWI) data before and after treatment.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If pre-treatment DWI and PWI data are used to estimate the penumbra zone, then the estimation can be obtained, but the prediction accuracy is limited due to individual variations in infarct progression
Solution Approach 1:
The patent applies preliminary action by training prediction models in advance using post-treatment DWI data from multiple patients to learn the relationship between pre-treatment imaging features and actual treatment outcomes. These pre-trained models are then applied to individual patients before treatment to predict their specific infarct progression patterns, thereby improving prediction accuracy while accounting for individual variations.
Solution Approach 2:
The patent uses copying by creating virtual replicas of treatment outcomes through machine learning models. The models are trained on copied data from multiple patients (post-treatment DWI images showing actual infarct progression) and generate predicted outcome copies for individual patients, allowing accurate prediction of individual infarct progression without requiring actual post-treatment data from each patient.
2Measurement precision
If machine learning models trained on multiple patients' data are used, then prediction accuracy improves, but the system complexity increases
Solution Approach 1:
The patent applies segmentation by dividing the prediction system into two distinct models: a 'success' prediction model trained on patients who achieved successful recanalization, and a 'failure' prediction model trained on patients with unsuccessful recanalization. This segmentation allows the system to handle different treatment outcomes separately, improving prediction accuracy while maintaining manageable system complexity through modular architecture.
Data Source
AI summary
According to an embodiment of the present disclosure, there is provided a method of predicting a brain tissue lesion distribution, the method including: a model learning operation of learning a prediction model for predicting a brain tissue lesion distribution of a subject by using brain image data of a plurality of previous patients; an input obtaining operation of obtaining input data from brain image data of the subject; and an output operation of generating output image data including information on the lesion distribution after recanalization treatment for the subject, by using the prediction model. The prediction model includes a success prediction model that is learned by using data of patients in which recanalization treatment is successful among the plurality of previous patients, and a failure prediction model that is learned by using data of patients in which recanalization treatment fails among the plurality of previous patients.