Dynamic Radiology Protocol Data Generation via AI-Human Merge
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current radiology reporting systems lack a coherent method for integrating expert knowledge and automatic analysis tools to effectively analyze medical measurement data, leading to static DICOM SR objects that are not updated and a gap between AI-generated and human-reported data, with inadequate peer review processes due to selection bias and lack of explainability.
Innovation Solution
A method for generating protocol data by creating first and second structured medical data objects from user-generated and AI-generated data, respectively, and merging them to form a final verified protocol data, using techniques like text analysis and graph structures to reconcile discrepancies and enhance peer review with explainable AI.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If DICOM SR objects are generated by algorithm automatically, then productivity is improved, but reliability deteriorates due to lack of updates after radiologist remeasurement
Solution Approach 1:
The DICOM SR object is transformed from a static to a dynamic structure that can be automatically updated. The system establishes update triggers that detect when radiologists perform remeasurements or add new measurements, and automatically refresh the DICOM SR object to reflect these changes, ensuring the structured report remains current without manual intervention.
Solution Approach 2:
A feedback mechanism is implemented where the system continuously monitors the radiology reporting process for remeasurement activities. When changes are detected in the measurement data, the system automatically feeds this information back to update the DICOM SR object, creating a closed-loop system that maintains data consistency between the free-text report and structured report.
2Manufacturing precision
If structured reporting is implemented to standardize terminology, then manufacturing precision is improved, but device complexity increases due to integration requirements
Solution Approach 1:
The system creates a universal integration layer that works across different radiology reporting workflows and AI measurement tools. This layer provides standardized DICOM SR object generation and update mechanisms that can be applied regardless of the specific radiology information system or AI tool being used, reducing integration complexity through a unified approach.
Solution Approach 2:
An intermediary component is introduced that acts as a mediator between the radiology reporting system and AI measurement tools. This intermediary automatically generates and updates DICOM SR objects, translating between different data formats and standards without requiring direct complex integration between all system components, thereby simplifying the overall system architecture.
3Reliability
If peer review is performed manually to ensure quality, then reliability is improved, but productivity deteriorates due to increased workload
Solution Approach 1:
The system implements self-service capabilities where the DICOM SR object automatically updates itself based on detected remeasurement activities. This eliminates the need for manual peer review intervention to ensure data consistency, as the system autonomously maintains accuracy by synchronizing structured and free-text reports, thereby improving productivity without compromising reliability.
Data Source
AI summary
A method for generating protocol data of a specific medical process is described. The specific medical process comprises at least one measurement process preferably a radiological imaging process. The method comprises generating a first structured medical data object based on user-generated electronic medical data. The user-generated electronic medical data are created by an expert documenting the specific medical process. Further, the method includes generating a second structured medical data object by applying an automatic processing algorithm to automatically created electronic medical data. The electronic medical data refer to the specific medical process. Furthermore, the method comprises generating the protocol data based on the first structured medical data object and on the second structured medical data object, wherein the first and second structured medical data objects are compared to determine deviations. Moreover, a protocol data generation device is described. Furthermore, a medical process system is described.