Hemodynamic Parameter Estimation via Bayesian Perfusion Modeling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for estimating hemodynamic parameters in perfusion imaging, such as PW-MRI and CT scans, face challenges in accurately determining arterial input functions, leading to biased and imprecise estimates due to the need for human intervention, additional measurements, and discrepancies between global and local input functions, which are time-consuming and difficult to reproduce.
Innovation Solution
A processing unit implements a method for estimating hemodynamic parameters by evaluating the a posteriori marginal distribution of parameters within a global perfusion model, using a prior assignment of joint distributions for arterial input and transit time models, eliminating the need for manual input and providing objective, reproducible estimates through Bayes methods or maximum likelihood approaches.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual selection of global arterial input function is used, then hemodynamic parameter estimation can be performed, but human intervention and additional measurements are required which increases time consumption and reduces reproducibility
Solution Approach 1:
The system automatically extracts the arterial input function from the perfusion image data itself without requiring manual selection or external measurements. The processing unit computes the arterial input function by analyzing the contrast agent concentration curves from the perfusion images, enabling the system to serve itself rather than requiring human intervention for this critical parameter estimation.
Solution Approach 2:
The method transforms the problem from manual parameter selection to automatic parameter computation. By changing the approach from requiring manual input function selection to automatically computing the arterial input function from the same perfusion data, the system eliminates time-consuming manual intervention while maintaining estimation accuracy through mathematical modeling and optimization.
2Device complexity
If global arterial input function is used, then processing can be simplified, but discrepancies arise between global and local input functions leading to biased estimates
Solution Approach 1:
The system extracts the arterial input function specifically from the perfusion images of the region of interest, creating a local arterial input function that corresponds to the specific vascular territory being analyzed. This local approach eliminates the discrepancies between global and local input functions by ensuring that the arterial input function is derived from the same local tissue characteristics and vascular anatomy that define the perfusion parameters being estimated.
3Productivity
If deconvolution is performed without proper arterial input function, then parameter estimation can be attempted, but results become unstable and biased
Solution Approach 1:
The system performs preliminary extraction and characterization of the arterial input function from the perfusion images before performing the deconvolution operation. By preparing the arterial input function in advance through automatic extraction and validation, the system ensures that the deconvolution process receives accurate, reliable input data, thereby stabilizing the overall estimation process and reducing biases that would otherwise result from using incorrect or manually selected input functions.
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 eliminates human intervention, reduces bias and instability in deconvolution, and provides precise, objective estimates of hemodynamic parameters, overcoming issues related to global vs. local input functions and improving the accuracy of perfusion modeling.
Implementation Method 1
evaluating the a posteriori marginal distribution of parameters within a global perfusion model, using a prior assignment of joint distributions for arterial input and transit time models, eliminating the need for manual input and providing objective, reproducible estimates through Bayes methods or maximum likelihood approaches
Implementation Method 2
evaluating the a posteriori marginal distribution of parameters within a global perfusion model, using a prior assignment of joint distributions for arterial input and transit time models, eliminating the need for manual input and providing objective, reproducible estimates through Bayes methods or maximum likelihood approaches
Implementation Method 3
a nuclear magnetic resonance or computed tomography imaging device is used. This delivers a plurality of sequences of digital images of a part of the body, in particular of the brain. Said device applies a combination of high-frequency electromagnetic waves to the part of the body in question and measures the signal re-emitted by certain atoms
Implementation Method 4
Said device applies a combination of high-frequency electromagnetic waves to the part of the body in question and measures the signal re-emitted by certain atoms
Implementation Method 5
a first relationship between the perfusion signal S(t) and a focus C(t) of a contrast agent circulating in said voxel over time t
Implementation Method 6
In Nuclear Magnetic Resonance Perfusion Imaging, there is an exponential relationship S(t)=S0.e-k.TE.C(t)S0 is the average signal intensity before the arrival of the contrast agent, YOU is the echo time (echo time in English) and k is a constant depending on the relationship between the paramagnetic susceptibility and the concentration of the contrast agent in the tissue
Implementation Method 7
In Perfusion CT, the signal for each voxel is directly proportional to the concentration: S(t)=α.C(t), α being a nonzero constant
Implementation Method 8
Concentration C(t) can then be expressed by the standard infusion convolution model C(t)=BF Ca(t)⊗R(t) where That(t) is the concentration of contrast agent in the artery supplying the tissue volume in a voxel (Arterial Input Function or Arterial Input Function (AIF) in English), BF is Blood Flow in tissue volume (Blood Flow in English), R(t) is the Complementary Distribution Function of the transit time in the volume of tissue (residue function in English) and ⊗ denotes the convolution product
Data Source
Figure 1~4
Figure 5a~7
Figure 8
AI summary
The invention relates to a method for estimating haemodynamic perfusion parameters of an elementary volume - termed a voxel - of an organ, from perfusion signals by jointly estimating the parameters of an optionally limited comprehensive perfusion model. The invention moreover relates to a processing unit of a perfusion imaging analysis system, adapted for carrying out such a method and for delivering the estimated parameters according to an appropriate format to a human-machine interface able to represent said estimated parameters for a user.