Machine-Learned MRI Scanner Model Updates for Reproducible Quantification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Magnetic Resonance Imaging (MRI) is plagued by non-reproducibility and non-quantitativeness due to MRI scanner-induced imperfections such as static and dynamic B0-variations, eddy currents, and patient-induced effects like B0-variations and coil loading, leading to variability in image quality and hindering quantification of relaxation parameters.
Innovation Solution
A computer-implemented method using a machine learning model (MLM) updates an initial scanner model with actual MRI measurement data to predict an updated scanner model, which determines the MRI configuration, incorporating corrections for magnetic field deviations and gradient fields, improving reproducibility and quantitativeness.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If very fast and very accurate adjustment and calibration scans are performed to incorporate scanner-induced imperfections and patient-induced effects, then measurement precision and reliability of MRI quantification are improved, but loss of time increases due to additional scan requirements
Solution Approach 1:
The system performs calibration scans and updates the scanner model in advance before actual diagnostic imaging. The machine learning model is trained beforehand on calibration data to predict scanner imperfections and patient-induced effects, so that when diagnostic scans are performed, the pre-trained model can quickly apply corrections without requiring additional calibration time during the diagnostic process.
Solution Approach 2:
The system creates a digital copy or virtual model of the scanner's magnetic field characteristics and gradient behavior through the scanner model. This virtual model is updated using calibration data and machine learning predictions, allowing the system to simulate and correct for imperfections in the actual hardware without requiring physical re-calibration or additional measurement time during diagnostic scanning.
2Reliability
If the scanner model is continuously updated with actual measurement data using machine learning to account for B0-variations, eddy currents, and patient-specific effects, then reliability and measurement precision are improved, but device complexity increases due to the implementation of machine learning models and additional data processing requirements
Solution Approach 1:
The scanner model updates itself automatically using the machine learning model and calibration data without requiring manual intervention or complex external calibration procedures. The system performs self-calibration by processing calibration scans through the trained ML model to generate updated scanner model parameters, eliminating the need for operator expertise in manual calibration and reducing operational complexity.
Solution Approach 2:
The system updates the scanner model by changing its parameters based on calibration data and machine learning predictions. Instead of redesigning the entire system, the approach modifies specific model parameters (such as B0 field map, gradient correction terms, coil sensitivity profiles) that capture scanner imperfections and patient-specific effects, allowing incremental improvements in reliability without proportionally increasing overall system complexity.
Data Source
Figure 1
Figure 2~3
Figure 4~5
AI summary
For automatically determining a configuration (25) for magnetic resonance imaging, MRI, an initial scanner model (22) for an MRI scanner (1) is received, the initial scanner model (22) specifying a deviation of a main magnetic field from a predefined target main magnetic field and/or a specifying a deviation of a magnetic field gradient from a predefined target gradient field. MRI measurement data (23) measured by using the MRI scanner (1) is received. A first updated scanner model (24) is generated by applying a trained first machine learning model, MLM, (20) to first input data (22, 23), which depends on the initial scanner model (22) and the MRI measurement data (23). The configuration (25) for MRI is determined depending on the first updated scanner model (24).