Dixon MRI Fat Fraction Mapping Without Negative Values
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Dixon MRI methods often result in negative fat fraction values, which are physically impossible, leading to biased and inaccurate measurements, and lack of confidence in fat fraction maps displayed to medical professionals.
Innovation Solution
A method to calculate and display a modified fat fraction map limited to values between 0 and 1, combined with pixel-wise fat fraction maps and interactive ROI selection using neural networks to enhance accuracy and reduce noise bias.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If Dixon MRI methods are used to calculate fat fraction, then fat fraction quantification is achieved, but negative fat fraction values are produced which are physically impossible
Solution Approach 1:
The patent applies parameter changes by transforming the fat fraction values through mathematical operations (such as taking the absolute value or applying a transformation function) to convert negative values into positive values. This ensures that the displayed fat fraction values are physically meaningful while preserving the quantitative information extracted from the Dixon MRI images.
2Loss of information
If raw fat fraction maps are displayed to medical professionals, then measurement data is provided, but lack of confidence arises due to negative values
Solution Approach 1:
The patent converts the harmful effect of negative fat fraction values into a benefit by applying a transformation that eliminates negative values while preserving the quantitative measurement capability. The transformed values are then displayed to medical professionals, providing both accurate measurement data and diagnostic confidence since the values now reflect physically possible fat fractions.
3Measurement precision
If fat fraction maps with negative values are displayed, then measurement information is provided, but anatomical structure visibility is reduced
Solution Approach 1:
The patent applies parameter changes by transforming the fat fraction values to eliminate negative numbers. This transformation improves the visibility of anatomical structures in the displayed maps, as negative values can create artifacts and reduce the dynamic range that enhances anatomical detail. The transformed values maintain measurement precision while improving visual quality.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Enables accurate and reliable fat fraction measurements by ensuring only physically possible values are displayed, improving anatomical structure visibility and reducing measurement bias, thereby enhancing diagnostic confidence.
Implementation Method 1
Dixon MRI employs chemical shift imaging to differentiate between water and fat signals within the body. By acquiring images at different echo times, Dixon MRI effectively separates and quantifies the water and fat content in tissues
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Disclosed herein is a method of determining a fat fraction value (130). The method comprises receiving (200) a water Dixon image (122) and a fat Dixon image (124) descriptive of a region of a subject (318). The method further comprises calculating (202) a modified fat fraction map (126) using the water Dixon image and the fat Dixon image, wherein the modified fat fraction map is limited to fat fraction values greater than or equal to zero and preferably less than or equal to one. The method further comprises displaying (204) the modified fat fraction map on a user interface (108) and receiving (206) a selection of a measurement ROI (128) descriptive of an area within the region of the subject from the user interface. The method further comprises calculating (208) the fat fraction value for the measurement ROI using the water Dixon image and the fat Dixon image and providing (210) the fat fraction value. Systems are disclosed, configured to perform the method.