Perfusion Index Estimation via Bayesian Capillary Transit Time Modeling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for estimating perfusion indices, such as singular value decomposition (SVD), underestimate high flow components, are sensitive to delays in the arterial input function, and are not robust against noise, which can compromise the accuracy of perfusion index measurements, especially in critical conditions like acute stroke.
Innovation Solution
A method using a parametric model for capillary transit time distributions combined with a minorize-maximization type procedure, specifically an expectation-maximization (EM) type procedure with regularization, to estimate perfusion indices, providing an exact analytical expression for the variance of the observation error, thereby improving robustness and reducing scan time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If SVD method is used for estimating perfusion indices, then the method is computationally simple and widely available, but it underestimates high flow components and is sensitive to delays in arterial input function
Solution Approach 1:
The patent changes the fundamental parameters of the estimation method by switching from SVD decomposition to a Bayesian framework with parametric capillary transit time distribution models. This allows the system to estimate multiple perfusion parameters (CBF, MTT, CTH) simultaneously while accounting for noise and delays through probabilistic modeling, thereby improving measurement precision without sacrificing computational feasibility through efficient algorithms.
Solution Approach 2:
The patent replaces the deterministic mechanical deconvolution approach of SVD with a statistical Bayesian inference system. This substitution introduces probability distributions and noise modeling that naturally handle delays and high flow components, improving accuracy while maintaining computational efficiency through optimized sampling and approximation techniques.
2Adaptability or versatility
If SVD method is used for estimating perfusion indices, then the method is widely implemented, but it is not robust against noise
Solution Approach 1:
The Bayesian framework implements feedback through iterative optimization where the estimated parameters are continuously refined based on how well they explain the observed concentration-time curves. The noise model provides feedback about measurement quality, allowing the system to adjust parameter estimates and uncertainty quantification accordingly, thereby improving reliability in noisy conditions while maintaining versatility across different imaging systems.
Solution Approach 2:
The patent combines multiple modeling components (capillary transit time distribution, noise models, prior distributions) into a composite Bayesian framework. This composite approach integrates various sources of information and uncertainty representations, creating a robust system that maintains reliability across different noise levels while remaining adaptable to various imaging modalities and clinical scenarios.
3Measurement precision
If Bayesian modeling with parametric capillary transit time distribution is used, then high flow components are not underestimated and noise sensitivity is reduced, but computing time increases and precision may be compromised
Solution Approach 1:
The patent performs preliminary actions by pre-defining physiologically plausible parametric models for capillary transit time distribution and pre-establishing efficient computational algorithms. These preparatory steps constrain the solution space to biologically realistic parameters, reducing the computational burden of Bayesian inference while maintaining accuracy for high flow components and improving computing efficiency through optimized sampling strategies.
Solution Approach 2:
The patent employs dynamic computational strategies that adapt the level of computational effort to the specific data characteristics. The system dynamically adjusts sampling density, optimization iterations, and model complexity based on signal quality and required precision, thereby optimizing the trade-off between computing time and measurement precision for each specific clinical case while maintaining accuracy for critical high flow measurements.
Data Source
Figure 1A
Figure 1B
Figure 2A
AI summary
The present invention relates to a method for estimating perfusion indices (PI) in a mammal (200), e.g. a human. Data (DAT, DAT') representative of a contrast agent concentration as a function of time of an injected contrast agent is obtained from a medical imaging system (100). Perfusion indices(PI) are found by applying a parametric model (PM) for capillary transit time distributions as a function of time, and a minorize-maximization (MM) type procedure, such as an expectation-maximization (EM) type procedure, with regularization. The minorize-maximization type procedure has an exact analytical expression for the variance of an observation error of the contrast agent (Cε) in a non-linear observation model for the contrast agent concentration used in the maximization step. Clinical tests performed for 7 patients show improved MTT mapping as compared to singular value decomposition (SVD), and reduced sensitivity to delay.