Method for monitoring myocardial wall thickness
View Patent ↗A method for monitoring left ventricular (LV) myocardial wall thickness. The method includes: obtaining real time images of a periodically spatially changing myocardium and segmenting the myocardium in such images; calculating wall thickness of the myocardium from each one of the obtained images; and performing a dynamic harmonic analysis of the calculated thickness to determine spatial changes in the thickness of the wall of the myocardium. The method applies the calculated wall thickness to a predictor to determine changes in the thickness of the wall of the myocardium. The method applies the calculated wall thickness to a predictor to determine the periodicity the myocardium.
1. A computer-implemented method for real time monitoring of a myocardium during a cardiovascular procedure, the method implemented by the computer comprising the steps of:
acquiring real time images of a left ventricle (LV) myocardium for a plurality of time steps and segmenting in real time the myocardium in each said image to estimate the endocardial and epicardial contours of the LV myocardium;
calculating a wall thickness of the myocardium in real-time from the estimated endocardial and epicardial contours for each said segmented image;
using a first state space model with a first regressive filter to predict the myocardium wall thickness for each acquired real time image over the plurality of time steps based on the previously acquired real time images and an estimated cardiac beating frequency, wherein the first state space model of the myocardium wall thickness is
S
t
=
c
0
+
∑
j
=
1
,
2
(
c
j
cos
(
j
wt
+
σ
j
)
)
,
wherein c 0 and c 1 are constants w=2πf wherein f is the estimated cardiac beating frequency, and σ j is a phase; and
using the predicted myocardium wall thickness with a second state space model with a second regressive filter to predict a cardiac beating frequency for each acquired real time image over the plurality of time steps based on the previously acquired real time images,
wherein at each subsequent time step the estimated cardiac beating frequency is updated with the predicted cardiac beating frequency in the first state space model.
2. The method of claim 1 , further comprising initializing endocardial and epicardial contours for a first real time image from the estimated endocardial and epicardial contours, propagating said initialized endocardial and epicardial contours to a subsequently acquired real time image and applying a local deformation to said contours.
3. The method of claim 1 , wherein the first regressive filter estimates at each time step a coefficient vector (c 0 , a 1 , b 1 , a 2 , b 2 ) based on the first state space model from the myocardium wall thicknesses calculated for each segmented image, wherein
a 1 =c 1 cos(σ 1 ), b 1 =c 1 sin(σ 1 ), a 2 =c 2 cos(σ 2 ),and b 2 =c 2 sin(σ 2 ).
4. The method of claim 1 , further comprising applying a CUSUM algorithm to a difference between the predicted myocardium wall thickness and the calculated myocardium wall thickness to detect a rapid change in the differences.
5. The method of claim 1 , further comprising applying a CUSUM algorithm to a mean and variance of the predicted cardiac beating frequency and the estimated cardiac beating frequency to detect changes between the predicted cardiac beating frequency and the estimated cardiac beating frequency, and to detect significant changes between predicted cardiac beating frequencies for consecutive time steps and between estimated cardiac beating frequencies for consecutive time steps.
6. The method of claim 5 , wherein the estimated cardiac beating frequency is determined using a fast Fourier transform of the calculated myocardium wall thickness.
7. A computer-implemented method for real time monitoring of a myocardium during a cardiovascular procedure, the method implemented by the computer comprising the steps of:
acquiring real time images of a left ventricle (LV) myocardium for a plurality of time steps and segmenting in real time the myocardium in each said image to estimate the endocardial and epicardial contours of the LV myocardium;
calculating a wall thickness of the myocardium in real time from the estimated endocardial and epicardial contours for each said segmented image;
using a first state space model with a first regressive filter to predict the myocardium wall thickness for each acquired real time image over the plurality of time steps based on the previously acquired real time images and an estimated cardiac beating frequency,
wherein the first state space model of the myocardium wall thickness is
S
t
=
c
0
+
∑
j
=
1
,
2
(
c
j
cos
(
j
wt
+
σ
j
)
)
,
wherein c 0 and c j are constants, w=2πf wherein f is the estimated cardiac beating frequency, and σ j is a phase, and the first regressive filter estimates at each time step a coefficient vector (c 0 , a 1 , b 1 , a 2 , b 2 ) based on the first state space model from the myocardium wall thicknesses calculated for each segmented image, wherein
a 1 =c 1 cos(σ 1 ), b 1 =c 1 sin(σ 1 ), a 2 =c 2 cos(σ 2 ),and b 2 =c 2 sin(σ 2 ).
8. The method of claim 7 , further comprising using the predicted myocardium wall thickness with a second state space model with a second regressive filter to predict a cardiac beating frequency for each acquired real time image over the plurality of time steps based on the previously acquired real time images, wherein at each subsequent time step the estimated cardiac beating frequency is updated with the predicted cardiac beating frequency first state space model.
9. A non-transitory program storage device readable by a computer, tangibly embodying a program of instructions executable by the computer to perform the method steps for real time monitoring of a myocardium during a cardiovascular procedure, the method comprising the steps of:
acquiring real time images of a left ventricle (LV) myocardium for a plurality of time steps and segmenting in real time the myocardium in each said image to estimate the endocardial and epicardial contours of the LV myocardium;
calculating a wall thickness of the myocardium in real-time from the estimated endocardial and epicardial contours for each said segmented image;
using a first state space model with a first regressive filter to predict the myocardium wall thickness for each acquired real time image over the plurality of time steps based on the previously acquired real time images and an estimated cardiac beating frequency, wherein the first state space model of the myocardium wall thickness is
S
t
=
c
0
+
∑
j
=
1
,
2
(
c
j
cos
(
j
wt
+
σ
j
)
)
,
wherein c 0 and c j are constants, w=2πf wherein f is the estimated cardiac beating frequency, and π j is a phase; and
using the predicted myocardium wall thickness with a second state space model with a second regressive filter to predict a cardiac beating frequency for each acquired real time image over the plurality of time steps based on the previously acquired real time images,
wherein at each subsequent time step the estimated cardiac beating frequency is updated with the predicted cardiac beating frequency first state space model.
10. The computer readable program storage device of claim 9 , the method further comprising initializing endocardial and epicardial contours for a first real time image from the estimated endocardial and epicardial contours, propagating said initialized endocardial and epicardial contours to a subsequently acquired real time image and applying a local deformation to said contours.
11. The computer readable program storage device of claim 9 , wherein the first regressive filter estimates at each time step a coefficient vector (c 0 , a 1 , b 1 , a 2 , b 2 ) based on the first state space model from the myocardium wall thicknesses calculated for each segmented image, wherein
a 1 =c 1 cos(σ 1 ), b 1 =c 1 sin( a 2 =c 2 cos(σ 2 ),and b 2 =c 2 sin(σ 2 ).
12. The computer readable program storage device of claim 9 , the method further comprising applying a CUSUM algorithm to a difference between the predicted myocardium wall thickness and the calculated myocardium wall thickness to detect a rapid change in the differences.
13. The computer readable program storage device of claim 9 , the method further comprising applying a CUSUM algorithm to a mean and variance of the predicted cardiac beating frequency and the estimated cardiac beating frequency to detect changes between the predicted cardiac beating frequency and the estimated cardiac beating frequency, and to detect significant changes between predicted cardiac beating frequencies for consecutive time steps and between estimated cardiac beating frequencies for consecutive time steps.
14. The computer readable program storage device of claim 13 , wherein the estimated cardiac beating frequency is determined using a fast Fourier transform of the calculated myocardium wall thickness.