Fractional flow reserve (FFR) index

Patent metadata

Field	Value
Publication number	US-9757073-B2
Application number	US-201314437990-A
Country	US
Kind code	B2
Filing date	Oct 24, 2013
Priority date	Nov 6, 2012
Publication date	Sep 12, 2017
Grant date	Sep 12, 2017

Abstract

As described herein, an unknown FFR is classified based on certain extracted features. In addition, an estimation of the unknown FFR can be determined based on certain extracted features. Furthermore, a confidence interval can be determined for the estimated FFR. In another instance, boundary conditions for determining an FFR via simulation are determined. The boundary conditions can be used to classify the unknown FFR.

First claim

The invention claimed is:

1. A method, comprising:
estimating an outlet flow rate of a stenosis based on an effective diameter or radius at an outlet of the stenosis, a diameter or radius at an inlet of the stenosis, and a flow rate at the inlet of the stenosis, where at least one of the effective diameter or radius, the diameter or radius or the flow rate is determined based on image data produced by one of a computed tomography, an X-ray, or a magnetic resonance imaging system;
iteratively estimating a resistance of the by:
calculating an initial resistance as a function of aortic blood pressure at the inlet of the stenosis and the outlet flow rate;
performing a coarse computational fluid dynamic simulation;
generating a subsequent resistance based on the initial resistance and a result of the coarse computational fluid dynamic simulation, wherein the subsequent resistance is the estimated resistance; and
repeating the acts of performing and generating using a current resistance until stopping iteration stopping criteria is satisfied;

classifying an unknown fractional flow reserve metric for a cardiac vessel with the stenosis as one of a plurality of different pre-defined classes based on extracted features and a learning model, wherein the extracted features include one or more estimated boundary conditions of the stenosis, including at least one of the estimated outlet flow rate or the estimated outlet resistance of the stenosis; and
generating a signal indicative of the classification.

2. The method of claim 1 , wherein the classification includes generating a numerical value for the unknown FFR and, further comprising:
classifying the unknown fractional flow reserve metric based on the numerical value.

3. The method of claim 1 , wherein features include one or more of a stenosis percentage, a stenosis length, a distance between the aorta and the stenosis, a stenosis position, a heart chamber size, a myocardium mass, a geometry of a coronary artery, a center line of a coronary artery.

4. The method of claim 3 , wherein the features include one or more of a subset of voxels corresponding to a set of intensities of interest around at least one of the inlet of the stenosis, a middle region of the stenosis or the outlet to the stenosis, a median intensity value of at least one of after the stenosis or before the stenosis, an intensity profile in around at least one of the inlet, the outlet or a center region of the stenosis, or an intensity profile along the vessel, one or more of test results, vital signs, subject history, or subject family history, or a result of a computational fluid dynamic simulation.

5. The method of claim 1 , further comprising:
optimizing the estimated resistance by minimizing an energy function including a resistance term and a flow velocity term.

6. The method of claim 5 , further comprising:
performing a subsequent computational fluid dynamics simulation based on the estimated outlet flow rate of the stenosis and the estimated resistance of the stenosis.

7. The method of claim 1 , further comprising:
estimating a point estimation of a fractional flow reserve for the stenosis and generating a second signal indicative thereof.

8. The method of claim 7 , the estimating, comprising:
using a weighted interpolation based on at least one of a predetermined number of nearest neighbor samples of a training set or only from a related class set from the training set, where the related class set is a class corresponding to the classification class.

9. The method of claim 7 , further comprising:
determining a confidence interval of the estimated fractional flow reserve and generating a third signal indicative thereof.

10. The method of claim 1 , further comprising:
determining the flow rate at the inlet of the stenosis using a machine learning algorithm trained with patient data, including at least one of image or non-image data, and flow rates measurements of other patients determined via cardiac catheterization.

11. The method claim 10 , further comprising:
determining a velocity based on the outlet flow rate.

12. The method of claim 1 , further comprising:
performing a subsequent computational fluid dynamics simulation based on the estimated outlet flow rate of the stenosis and the estimated resistance of the stenosis.

13. The method of claim 12 , further comprising:
determining a fractional flow reserve for the steno sis based on a result of the subsequent computational fluid dynamics simulation.

14. A system, comprising:
a data analyzer with a processor configured to determine a fractional flow reserve classification of an unknown fractional flow reserve for a stenosis, the processor, including:
a boundary condition estimator configured to estimate at least one boundary condition of a stenosis of a vessel, including at least one of an estimated outlet flow rate of the stenosis or an estimated outlet resistance of the stenosis, based on image data that includes a representation of the vessel and the stenosis, wherein the image data is generated by one of a one a computed tomography, an X-ray, or a magnetic resonance imaging modality,
wherein the boundary condition estimator estimates an outlet flow rate of the stenosis as a function of an effective diameter or radius at an outlet of the stenosis, a diameter or radius at an inlet of the stenosis, and a flow rate at the inlet of the stenosis, and the boundary condition estimator estimates a resistance of the stenosis based an aortic blood pressure at the inlet of the stenosis and the outlet flow rate using an iterative algorithm;
a feature extractor configured to extract one or more features from at least one of segmented tissue of interest in image data representing the stenosis and corresponding vessel, the estimated boundary conditions of the stenosis, and intensity information from the image data or subject data;
a classifier configured to classify the unknown fractional flow reserve into one of a plurality of different pre-defined classes based on the extracted features, including the estimated outlet flow rate of the stenosis and the estimated outlet resistance of the stenosis extracted features, and a learning model; and
a FFR estimator configured to estimate a point estimation of the fractional flow reserve for the stenosis using a weighted interpolation based on at least one of a predetermined number of nearest neighbor samples of a training set or only from a related class set from the training set, where the related class set is a class corresponding to the classification class.

15. The system of claim 14 , the data analyzer, further comprising:
a confidence interval determiner configured to determine a confidence interval of the estimated fractional flow reserve.

16. The system of claim 14 , further comprising:
a parameter determiner configured to determine the flow rate at the inlet of the stenosis using a machine learning algorithm trained with patient data, including at least one of image or non-image data, and flow rates measurements of other patients determined via cardiac catheterization.

17. The system of claim 14 , wherein the boundary condition estimator estimates a velocity for the stenosis based on the flow rate.

18. The system of claim 14 , the data analyzer, further comprising:
a CFD processor configured to perform a computational fluid dynamics simulation based on the at least one boundary condition; and
a FFR determiner configured to determine an FFR for the stenosis based on a result of the CFD processor.

Assignees

• Koninklijke Philips Nv

Inventors

• Goshen Liran
• Lamash Yechiel
• Gilboa Guy

Classifications

• G16H50/30

  for calculating health indices; for individual health risk assessment · CPC title

• A61B6/504

  for diagnosis of blood vessels, e.g. by angiography · CPC title

• A61B5/021

  Measuring pressure in heart or blood vessels · CPC title

• A61B5/1075

  for measuring dimensions by non-invasive methods, e.g. for determining thickness of tissue layer (A61B8/0858 takes precedence) · CPC title

• A61B5/0263

  using NMR · CPC title