Bayesian TNBC Recurrence Prediction From Mammograms and Biopsy Images

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing methods struggle to accurately predict the recurrence of Triple Negative Breast Cancer (TNBC) using routinely collected clinical data, including mammograms and biopsy samples, due to the lack of integration and interpretation of diverse data sources, leading to inefficiencies in prognosis prediction.

Innovation Solution

A Bayesian learning framework combines recurrence probabilities from pathomics, radiomics, and clinicopathological data using a disease prediction model that includes a pathomics model for H&E biopsy samples, a radiomics model for mammograms, and in situ imaging, refining latent spaces with generative adversarial networks to enhance predictive accuracy.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If multiple data sources (pathomics, radiomics, clinicopathological data) are integrated using a Bayesian learning framework, then predictive accuracy for TNBC recurrence is improved, but system complexity increases

Engineering Contradiction:
Improvepredictive accuracyVSAvoidsystem complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The system divides the complex prediction task into separate specialized models: a pathomics model for analyzing H&E biopsy samples, a radiomics model for processing mammogram images, and a disease prediction model that integrates their outputs. Each model processes specific data types independently before consolidation, reducing overall system complexity while maintaining high predictive accuracy through modular architecture.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The disease prediction model acts as an intermediary that receives and integrates recurrence probabilities from the specialized pathomics and radiomics models along with clinicopathological data. This mediator structure allows diverse data sources to be combined systematically through Bayesian learning without requiring direct integration of all components, simplifying the overall system architecture.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Measurement precision

If diverse data sources are integrated to predict TNBC recurrence, then prognosis prediction accuracy is improved, but data processing time increases

Engineering Contradiction:
Improveprognosis prediction accuracyVSAvoiddata processing time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The system performs preliminary processing of diverse data sources through specialized models that extract recurrence probabilities before final integration. The pathomics model pre-processes biopsy data and the radiomics model pre-processes mammogram data independently, preparing refined outputs that accelerate the final prediction stage by the disease prediction model, thus reducing overall processing time while maintaining accuracy.

Inventive Principle:
Principle #10Preliminary action

3Reliability

If a Bayesian learning framework is used to combine recurrence probabilities, then sensitivity and specificity in prediction are improved, but computational requirements increase

Engineering Contradiction:
Improvesensitivity and specificityVSAvoidcomputational requirements
Core Design Contradiction:
ReliabilityVSUse of energy by moving object

Solution Approach 1:

The system transforms complex multi-source data into standardized recurrence probability parameters that can be efficiently processed by the Bayesian learning framework. By converting diverse inputs (biopsy images, mammograms, clinical data) into unified probability outputs from specialized models, the computational complexity is reduced to manageable parameter integration, maintaining high sensitivity and specificity while lowering computational requirements.

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentEP4254327B1System and method for detecting recurrence of a disease
Publication Date: 2026.01.21 GE PRECISION HEALTHCARE LLC
  • EP4254327B1 patent drawingFigure 1~2
  • EP4254327B1 patent drawingFigure 3
  • EP4254327B1 patent drawingFigure 4

AI summary

A method for determining a recurrence of a disease in a patient is presented. The method includes generating a plurality of medical images of an organ of the patient and determining a plurality of recurrence probabilities from the plurality of medical images. A recurrence of the disease is determined based on the plurality of recurrence probabilities and clinicopathological data of the patient using a Bayesian network.