Automated Breast Image Texture Analysis for Cancer Risk Assessment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for assessing cancer risk, such as Gail risk factors and mammographic breast density assessment, are limited by their reliance on population statistics and subjective density assessments, which do not accurately predict individual cancer risk and are prone to human error.
Innovation Solution
A system that analyzes texture features from breast images, such as skewness, coarseness, and contrast, combined with personal risk factors, to provide a probabilistic cancer risk assessment using logistic and linear regression models, with the option for automatic ROI selection to minimize human error.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If automated ROI selection and texture analysis are implemented, then measurement precision and reliability of cancer risk assessment are improved, but device complexity and difficulty of detecting and measuring increase
Solution Approach 1:
The system performs automatic ROI selection and texture feature extraction without requiring manual intervention. The computer automatically identifies regions of interest in breast images and computes texture metrics, eliminating the need for subjective manual assessment while maintaining high measurement precision through algorithmic consistency
Solution Approach 2:
The patent replaces manual mechanical assessment methods with automated computational analysis. Texture features are extracted through algorithmic processing of digital images rather than human visual inspection, substituting mechanical human judgment with computational mechanisms that provide more precise and reproducible measurements
2Ease of operation
If automated ROI selection is used, then ease of operation is improved by minimizing human error, but device complexity increases due to image comparison and mapping requirements
Solution Approach 1:
The system automatically compares the current breast image with reference images and performs ROI mapping without user intervention. This self-service capability eliminates operational errors associated with manual ROI selection while providing consistent, reproducible results across different users and sessions
Solution Approach 2:
The system performs preliminary image comparison and ROI identification before the user needs to make any decisions. By pre-processing the images and automatically establishing anatomic correspondences, the system prepares the analysis in advance, making the overall operation simpler and more efficient
Data Source
AI summary
Methods and systems for determining a probabilistic assessment of a person developing cancer are disclosed. The probabilistic assessment may include receiving a digital breast image of a person, selecting a region of interest within the received breast image, and analyzing this selected region of interest with respect to texture analysis. A probabilistic assessment may then be determined through the use of a logistic regression model based on the texture analysis within the region of interest and personal risk factors. A probabilistic assessment may also be determined through the use of a linear regression model based on the texture analysis within the region of interest and a known cancer indicator or risk factor.