PET Random Coincidence Correction via Delay Scaling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current PET imaging systems face challenges in accurately correcting for random coincidences, which can degrade image quality by introducing noise and bias, and require trade-offs between noise amplification and data-processing requirements.
Innovation Solution
A PET imaging system comprising a singles unit, delay unit, and scaling unit generates a correction sinogram for random coincidences using the singles rate between detector locations, determines delay coincidences, and iteratively reconstructs images based on the scaled random sinogram and prompt sinogram, effectively reducing noise and bias.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If delay coincidences are used to correct random coincidences, then bias is reduced, but noise is amplified
Solution Approach 1:
The patent combines two correction approaches: the singles-based method (which provides a noiseless estimate but introduces bias) and the delay coincidence method (which reduces bias but amplifies noise). By merging these methods and applying optimal weighting, the system achieves both low noise and low bias in random coincidence correction.
Solution Approach 2:
The patent optimizes the weighting parameter that balances the contribution of singles-based correction and delay coincidence-based correction. By adjusting this parameter, the system can control the trade-off between noise amplification and bias reduction, achieving optimal correction performance.
2Object-affected harmful factors
If singles-based estimation is used to correct random coincidences, then noise is reduced, but bias is introduced
Solution Approach 1:
The patent combines two correction approaches: the singles-based method (which provides a noiseless estimate but introduces bias) and the delay coincidence method (which reduces bias but amplifies noise). By merging these methods and applying optimal weighting, the system achieves both low noise and low bias in random coincidence correction.
Solution Approach 2:
The patent uses the delay coincidence data as feedback to adjust and optimize the weighting of the singles-based correction. This feedback mechanism allows the system to minimize bias while maintaining the noise-reduction benefits of the singles-based approach.
3Quantity of substance
If a longer coincidence time window is used, then more true coincidences are detected, but more random coincidences are also detected
Solution Approach 1:
The patent performs preliminary estimation of random coincidences using the singles rate before reconstructing the final image. This preliminary action allows the system to account for and correct random coincidences that were detected due to the longer time window, ensuring accurate quantification.
Solution Approach 2:
The patent converts the harmful effect of increased random coincidences (caused by longer time window) into a beneficial correction signal. By using the singles-based estimation method, the system transforms the noise from random coincidences into useful information for correcting the prompt sinogram, thereby improving image quantification accuracy.
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
This approach provides a more accurate quantification of radiopharmaceutical uptake, improves correction for detected prompts, and offers a noiseless measure of random coincidences, leading to enhanced image quality and quantification.
Implementation Method 1
True coincidences are the result of a positron emission and subsequent annihilation which generates two gamma photons along a line of response (LOR)
Implementation Method 2
Detectors of a PET scanner, typically in a ring configuration, detect the photons
Implementation Method 3
scatter events are coincidences in which one or both gamma photons are Compton scattered resulting in the LOR between the two photon detector locations being different from the actual defined path followed by the photons. Scatter events can be identified based on a lower energy level of a photon, e.g. less than 511 keV
Data Source
AI summary
A positron emission tomography (PET) imaging system (10) includes a singles unit (24), a delay unit (26), a scaling unit (28), and a reconstruction unit (30). The singles unit (24) is configured to generate a correction sinogram of random coincidences defined by a coincidence time window tw and a time period t based on rij=2 t si sj where rij is an estimate of the random coincidences in the time period t between singles at detector locations si and sj. The delay unit (26) is configured to determine delay coincidences in the time period t defined by a delay coincidence time window of a paired coincidences including tw to 2 tw. The scaling unit (28) is configured to scale the correction sinogram of random coincidences based on the delay coincidences. The reconstruction unit (30) is configured to iteratively reconstruct one or more images based on the scaled random sinogram and a prompt sinogram for the time period t.