Lesion Marking Consistency via Protocol Enforcement
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Diagnostic imaging technologies generate large amounts of data, leading to inconsistencies in lesion/cancer markings between radiologists, resulting in intra-observer and inter-observer inconsistencies, which complicates diagnostic decision-making and lacks effective solutions.
Innovation Solution
A system and method for electronically enforcing lesion marking protocols, utilizing a consensus panel to analyze and combine multiple markings, and providing interactive tools for quality assurance, aiming to reduce inconsistencies by creating a consensus lesion marking.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If radiologists manually create lesion markings to annotate medical images, then ground truth data is generated for algorithm validation, but intra-observer and inter-observer inconsistencies occur due to variability in marking performance
Solution Approach 1:
The system implements feedback mechanisms where radiologists receive feedback on their markings through consistency checks against previous markings or consensus markings. This feedback loop allows them to correct inconsistencies and improve marking reliability, directly addressing the intra- and inter-observer variability problem while maintaining the need for expert annotation.
Solution Approach 2:
The system introduces an intermediary computing system that mediates between different radiologists' markings. This intermediary processes multiple markings, performs consistency checks, and generates consensus markings that reconcile differences. The intermediary acts as a mediator that harmonizes varying expert opinions into a unified reference standard.
2Quantity of substance
If multiple lesion markings are created by different radiologists or at different times, then more data is available for analysis, but inconsistencies increase the complexity of diagnostic decision making
Solution Approach 1:
The system merges multiple lesion markings from different radiologists or time points into a single consensus marking through automated processing. This combining operation aggregates the quantity of data while simplifying the output to a unified reference standard, thereby reducing the complexity of diagnostic decision making without losing the benefits of multiple perspectives.
Solution Approach 2:
The computing system serves as an intermediary that processes and reconciles multiple markings into a consistent reference. This intermediary layer absorbs the complexity of handling multiple data sources and presents a simplified, consistent output to radiologists, making the diagnostic decision making process less complex despite having access to multiple markings.
3Ease of manufacture
If approximate ground truth markings are used instead of validated biopsies, then the cost and difficulty of acquiring ground truth are reduced, but marking consistency deteriorates
Solution Approach 1:
The system uses feedback from consistency checks and consensus marking generation to improve the quality of approximate ground truth markings. By continuously comparing markings against each other and providing feedback to radiologists, the system enhances the precision of markings that would otherwise be inconsistent, while maintaining the ease of acquisition advantage.
Solution Approach 2:
The computing system acts as an intermediary that transforms approximate ground truth markings into more consistent reference standards. It processes the easily acquired approximate markings through consistency algorithms and outputs refined consensus markings that maintain accessibility while improving measurement precision.
Data Source
AI summary
Method and system for quality assurance in lesion marking is disclosed. A marking is created based on a patient study on display in a manner in accordance with an electronically enforced protocol. One or more features are generated based on the marking. Quality of the marking is assessed based on such computed one or features.