Medical Image Biomarker Prediction for Representative Lesion Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current clinical practices for selecting cancer lesions for tissue collection are subjective and inefficient, leading to unnecessary procedures that delay treatment and pose health risks to patients.
Innovation Solution
A method and system using machine learning models to predict biomarker expression from medical images, enabling accurate identification of optimal lesions for tissue collection by extracting and analyzing image regions, and outputting indices for priority determination.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If multiple tissue collections are performed to confirm drug effectiveness, then treatment accuracy is improved, but treatment time is extended and patient health is compromised
Solution Approach 1:
The system performs preliminary analysis of multiple lesions using machine learning models to predict biomarker expression and determine the most representative lesion before tissue collection. This preliminary action identifies the optimal lesion in advance, eliminating the need for multiple repeated tissue collections to confirm drug effectiveness, thereby reducing treatment time while maintaining treatment accuracy.
2Reliability
If multiple tissue collections are performed to confirm drug effectiveness, then treatment accuracy is improved, but patient health risk increases
Solution Approach 1:
The system performs preliminary analysis of multiple lesions using machine learning models to predict biomarker expression and determine the most representative lesion before tissue collection. This preliminary action identifies the optimal lesion in advance, eliminating the need for multiple repeated tissue collections to confirm drug effectiveness, thereby reducing patient health risk while maintaining treatment accuracy.
3Ease of operation
If clinicians select lesions based on subjective criteria such as size and location, then lesion selection is simplified, but representation accuracy of all patient lesions decreases
Solution Approach 1:
The system replaces the subjective mechanical selection process with an automated machine learning-based analysis system. The machine learning models objectively analyze multiple lesions based on image data and predict biomarker expression, automatically determining the most representative lesion. This substitution eliminates subjective bias while maintaining ease of operation through automated processing.
4Device complexity
If traditional subjective lesion selection is used, then clinical workflow is simple, but tissue collection efficiency decreases
Solution Approach 1:
The system replaces the simple but inefficient subjective selection workflow with an automated machine learning-based system that analyzes multiple lesions and predicts biomarker expression. Although the system is more complex, it significantly improves tissue collection efficiency by identifying the most representative lesion in advance, eliminating the need for multiple repeated collections and confirming drug effectiveness through more accurate initial selection.
Data Source
AI summary
The present disclosure relates to a method for predicting biomarker expression from a medical image. The method for predicting biomarker expression includes receiving a medical image. and outputting indices of biomarker expression for the at least one lesion included in the medical image by using a first machine learning model.