2D Topogram AI Analysis to Capture Missed CT Findings
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current medical imaging techniques often overlook important findings in 2D topogram images due to reliance on 3D image analysis, leading to missed diagnoses in computed tomography (CT) scans, particularly in abdominal and trauma imaging.
Innovation Solution
Implementing a computer-implemented method that utilizes a machine learning algorithm for topogram analysis to generate 2D annotation data, providing medical imaging decision support data to adjust scan protocols and capture findings in 3D images, including suggestions for scan range enlargement and additional scans.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If 3D image analysis is used for diagnostic reading, then diagnostic workflow efficiency is improved, but findings in 2D topogram images are overlooked
Solution Approach 1:
The diagnostic workflow is segmented into two independent analysis streams: one for 3D images and one for 2D topogram images. The AI system separately analyzes topogram images and generates findings that are then integrated with 3D image results, ensuring neither modality is overlooked while maintaining efficient workflow.
Solution Approach 2:
An AI-based image reading system acts as an intermediary between 2D topogram analysis and 3D image analysis. This intermediary automatically processes topogram images, extracts relevant findings, and integrates them with 3D image diagnostic results, preventing information loss while preserving workflow efficiency.
2Loss of time
If scan range is reduced for 3D reconstruction, then scan time and radiation dose are reduced, but findings outside reconstruction FoV are missed
Solution Approach 1:
The system performs preliminary analysis of the 2D topogram images before 3D reconstruction to identify potential findings. Based on these preliminary findings, the reconstruction field of view is adaptively adjusted to ensure all relevant findings are captured in the 3D images, preventing information loss while optimizing scan parameters.
Solution Approach 2:
The AI system provides feedback about findings detected in topogram images to adjust the 3D reconstruction field of view. This feedback loop ensures that the reconstruction FoV is optimized to include all clinically relevant findings identified in the preliminary 2D analysis, balancing scan efficiency with comprehensive diagnostic coverage.
3Loss of time
If topogram images are read sporadically, then reading time is reduced, but diagnostic accuracy decreases
Solution Approach 1:
The AI-based image reading system performs automatic analysis of topogram images without requiring manual review by radiologists. The system independently processes topogram images, generates findings, and integrates them with 3D image analysis results, eliminating the need for sporadic manual reading while maintaining high diagnostic accuracy.
Solution Approach 2:
The manual mechanical process of sporadic topogram review by radiologists is replaced with an automated AI-based image reading system. This substitution eliminates variability and fatigue associated with manual reading, ensuring consistent and accurate analysis of topogram images without increasing reading time.
Data Source
AI summary
One or more example embodiments of the present invention relates in one aspect to a computer-implemented method includes receiving 2D topogram data of a patient; generating 2D topogram annotation data by applying a machine learning algorithm for topogram analysis onto the 2D topogram data; generating the medical imaging decision support data based on the 2D topogram annotation data; and providing the medical imaging decision support data.