Pathological Slide Biomarker Analysis With Pathologist Feedback

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing machine learning models for detecting or segmenting biological elements from pathological slide images often perform poorly, adversely affecting biomarker analysis and the accuracy of treatment plans.

Innovation Solution

A computing device and method that generates first biomarker expression information through analysis, allows user input for updating analysis results, and outputs a report based on biomarker expression information to enhance accuracy.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Extent of automation

If machine learning models are used for detecting or segmenting biological elements from pathological slide images, then automation is improved, but measurement precision deteriorates

Engineering Contradiction:

Improveautomation of biological element detectionVSAvoidprecision of biomarker analysis

Core Design Contradiction:

Extent of automationVSMeasurement precision

Solution Approach 1:

The system implements feedback by allowing pathologists to review and correct AI-generated segmentation results. The corrected results are then fed back to improve the model's performance, creating a continuous improvement loop that maintains automation while enhancing precision through human-in-the-loop validation.

Inventive Principle:

Principle #23Feedback

Solution Approach 2:

The system introduces an intermediary review process where pathologists act as mediators between the AI model and final diagnostic decisions. This intermediary layer allows automated processing to continue while adding a human verification step that ensures measurement precision is maintained or improved.

Inventive Principle:

Principle #24Intermediary (Mediator)

2Device complexity

If machine learning models with limited performance are used, then device complexity is reduced, but reliability of biomarker analysis deteriorates

Engineering Contradiction:

Improvecomplexity of analysis systemVSAvoidreliability of treatment plan accuracy

Core Design Contradiction:

Device complexityVSReliability

Solution Approach 1:

The system performs preliminary action by pre-processing images and generating initial segmentation results using the machine learning model before clinical analysis. This allows the simpler model to handle routine tasks while pathologists focus on critical review, maintaining reliability without requiring complex models for all processing.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system segments the analysis workflow into distinct stages: automated initial analysis by the machine learning model, human review and correction, and final diagnostic decision-making. This segmentation allows a simpler model to be used while maintaining overall system reliability through distributed task allocation.

Inventive Principle:

Principle #1Segmentation

Data Source

PatentUS20260057517A1Method and device for processing pathological slide image

Publication Date: 2026.02.26 LUNIT

AI summary

A computing device includes at least one memory, and at least one processor configured to generate, based on first analysis on a pathological slide image, first biomarker expression information, generate, based on a user input for updating at least some of results of the first analysis, second biomarker expression information about the pathological slide image, and control a display device to output a report including medical information about at least some regions included in the pathological slide image, based on at least one of the first biomarker expression information or the second biomarker expression information.