3D Phantom Calibration for PET/SPECT Partial Volume Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Medical imaging systems like SPECT and PET suffer from poor resolution, leading to underestimated activity measurements due to the partial volume effect and structural noise, which complicates radiomic feature extraction in nuclear medicine and molecular imaging.
Innovation Solution
A 3D phantom with a fractal pattern is printed to represent the volume of interest, allowing for calibration of activity concentration uptake by correcting the partial volume effect using a PVC calibration factor determined through image reconstruction and comparison with a known test activity concentration.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If SPECT or PET imaging is used to detect radioactive tracer distribution, then organ function and cellular activity information can be obtained, but the resolution is poor leading to underestimated activity measurements due to partial volume effect
Solution Approach 1:
The patent creates a digital twin or virtual model of the tumor that replicates its physical characteristics including size, shape, and radioactive tracer distribution. This virtual copy allows for precise simulation and measurement of activity concentration without being limited by the physical resolution constraints of SPECT/PET imaging systems, thereby resolving the contradiction between measurement accuracy and imaging resolution.
Solution Approach 2:
The patent transforms the physical imaging problem into a computational parameter optimization problem. By adjusting parameters in the virtual model (such as tumor size, shape, and tracer distribution patterns) and comparing simulated images with actual SPECT/PET images, the system can infer accurate activity measurements even when the imaging resolution is poor, thus overcoming the partial volume effect limitation.
2Loss of information
If radiomics analysis is performed on SPECT or PET images, then quantitative radiomic features can be extracted, but structural noise patterns overwhelm the signal making feature extraction difficult
Solution Approach 1:
The patent extracts and isolates the relevant radiomic features from the noisy SPECT/PET images by using the virtual tumor model as a reference. The system identifies and extracts only the meaningful signal components while filtering out the structural noise patterns that would otherwise overwhelm the analysis, thereby preserving radiomic feature quality despite the noisy imaging environment.
Solution Approach 2:
The virtual tumor model serves as an intermediary between the noisy SPECT/PET images and the radiomics analysis. By comparing actual images with simulated images from the virtual model, the system can distinguish true biological signals from structural noise, enabling accurate radiomic feature extraction even in the presence of overwhelming noise patterns.
3Length of stationary object
If activity concentration is measured in tumors smaller than imaging resolution, then cellular level information can be obtained, but partial volume effect causes underestimation of actual activity
Solution Approach 1:
The patent moves the measurement problem from the physical imaging dimension to the computational simulation dimension. By creating a virtual model that can represent sub-resolution structures with precise geometric and radiological properties, the system can measure activity concentration in tumors smaller than the imaging resolution limit without suffering from partial volume effect, as the virtual model maintains exact dimensional accuracy regardless of physical imaging constraints.
Data Source
AI summary
A framework for calibrating activity concentration uptake. The framework generates a pattern that represents a volume of interest. A three-dimensional (3D) phantom may be printed based on the pattern. Activity concentration uptake in the volume of interest may then be calibrated by using the 3D phantom.