ML-Based Survey Scan Parameter Optimization in CT Imaging
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In medical imaging, particularly in CT imaging, technologists face challenges in determining optimal scan parameters for diagnostic quality while minimizing radiation dose, as 3D survey images do not directly provide values for parameters like tube-voltage and tube-current, requiring manual configuration and potentially suboptimal settings.
Innovation Solution
A computer-implemented method using a machine learning model trained for image-to-image translation to simulate the appearance of diagnostic scan data based on 3D survey scan data, allowing for the adjustment of scan parameters before the full scan to improve image quality and anatomy coverage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual configuration of scan parameters is used based on 3D survey images, then technologists can perform scan planning, but the image quality may be suboptimal and radiation dose cannot be minimized effectively
Solution Approach 1:
The system performs self-service by automatically determining optimal scan parameters (tube voltage, tube current, scan range) through machine learning analysis of the 3D survey image, eliminating the need for manual technologist configuration and achieving consistent optimal results without human intervention
Solution Approach 2:
The system automatically adjusts multiple scan parameters (tube voltage, tube current, scan range boundaries) based on analysis of the 3D survey image, transforming the manual parameter-setting process into an automated parameter-optimization process that improves image quality while reducing radiation dose
2Object-affected harmful factors
If full diagnostic scan is performed without parameter optimization, then complete diagnostic data is acquired, but radiation dose is higher than necessary
Solution Approach 1:
The system performs preliminary analysis of the 3D survey image to determine optimal scan parameters and boundaries before executing the full diagnostic scan, ensuring that the subsequent scan is configured to deliver complete diagnostic data with minimized radiation dose
Solution Approach 2:
The system uses feedback from the 3D survey image analysis to automatically adjust scan parameters for the full diagnostic scan, creating a closed-loop process where the survey scan results directly inform and optimize the parameters of the subsequent diagnostic scan to minimize radiation while ensuring diagnostic quality
3Loss of information
If 3D survey images are used instead of 2D survey images, then more detailed information is revealed, but manual configuration of acquisition parameters is still required
Solution Approach 1:
The system performs self-service by automatically extracting anatomical information from the 3D survey image and using machine learning to determine optimal acquisition parameters (tube voltage, tube current, scan range) without requiring manual technologist input, fully leveraging the rich anatomical information contained in the 3D survey data
Data Source
AI summary
Methods related to survey scanning in diagnostic medical imaging. At least one aspect relates to a method for generating, using a machine learning model, a simulation of medical image data in accordance with a defined set of acquisition parameters for a medical imaging apparatus, based on processing of an initial 3D image data set (e.g. a survey image dataset).