Liver Fat Quantification Using DEXA and Machine Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for assessing liver fat content, such as MRI, are expensive and time-consuming, while DEXA scans lack the detail and specificity needed for accurate liver fat analysis, limiting their use in large-scale screenings and quick result settings.
Innovation Solution
Training machine-learning models using DEXA scan data and corresponding MRI-based adiposity scores to predict liver fat content, allowing for efficient and economic estimation of liver fat based on DEXA image data, enabling diagnosis, monitoring, and treatment of conditions like NAFLD.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If MRI is used to calculate PDFF for liver fat quantification, then measurement precision is improved, but cost and time consumption increase significantly
Solution Approach 1:
The patent creates a computational model that copies the liver fat quantification capability from MRI to DEXA imaging. By training a machine learning model on paired DEXA and MRI data, the system replicates MRI's PDFF measurement accuracy using the faster, cheaper DEXA modality, effectively copying the diagnostic capability without the associated time and cost penalties
Solution Approach 2:
The patent replaces the physical MRI scanning mechanism with a computational approach using DEXA images. Instead of using magnetic resonance physics to directly measure liver fat, the system uses machine learning algorithms to infer liver fat content from DEXA bone density images, substituting a mechanical/physical measurement system with an information-processing system
2Ease of manufacture
If DEXA scans are used for body composition analysis, then cost and accessibility are improved, but measurement precision for liver fat is insufficient
Solution Approach 1:
The patent copies the liver fat measurement capability from MRI to DEXA by training a machine learning model to translate DEXA images into accurate liver fat estimates. This allows DEXA's cost and accessibility advantages to be preserved while gaining MRI-level precision through computational enhancement
Solution Approach 2:
The patent changes the interpretation parameters of DEXA images by applying machine learning algorithms that extract liver fat information from bone density images. The model transforms DEXA's traditional bone density parameters into liver fat composition parameters, enabling the same imaging modality to provide different diagnostic information with improved precision
3Ease of manufacture
If traditional DEXA analysis methods are used, then cost is reduced, but detail and specificity for liver fat assessment are lost
Solution Approach 1:
The patent copies the detailed liver fat assessment capability from MRI to DEXA through machine learning. The trained model transfers the information extraction capability to DEXA images, allowing cost-effective scanning to produce detailed liver fat-specific information that was previously only available through expensive MRI
Solution Approach 2:
The patent introduces a machine learning model as an intermediary between DEXA imaging and liver fat assessment. This computational mediator extracts and translates liver fat information from DEXA bone density images, preserving cost effectiveness while recovering the detailed liver fat specificity that traditional DEXA analysis loses
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
The approach enables accurate prediction of liver fat content using DEXA scans, overcoming the limitations of traditional DEXA analysis by specifically evaluating the liver portion of the images, thus providing a cost-effective and efficient method for liver fat assessment.
Implementation Method 1
Dual-energy X-ray absorptiometry (DEXA) scans use two different X-ray energies to estimate bone density and body composition
Data Source
AI summary
An exemplary method for predicting one or more adipose depots for a patient includes receiving one or more Dual-energy X-ray Absorptiometry (DEXA) scans comprising at least a portion of a torso of the patient; providing at least one or more portions of the one or more DEXA scans to a trained machine-learning model, wherein the machine-learning model is trained using a training dataset comprising: a plurality of training DEXA scans of a plurality of subjects and a plurality of corresponding Magnetic Resonance Imaging (MRI)-image-based adiposity scores of the plurality of subjects; and predicting the one or more adipose depots for the patient utilizing the trained machine-learning model.