Methods for Computing Coronary Physiology Indexes Using a High Precision Registration Model

Abstract

This invention describes methods to compute coronary physiology indexes using a high precision registration model, which consists of acquiring coronary angiography images of coronary vessels, performing intravascular imaging, and registering the coronary angiography images with intravascular images to create a high precision registration model, based upon which the coronary flow, fractional flow reserve (FFR) and index of microcirculation resistance (IMR) can be computed. The methods described in this invention to compute coronary flow, FFR, IMR are based on both coronary angiography and intravascular images, and the accuracy is better than those derived from coronary angiography alone or intravascular imaging alone, and have high practical values.

Claims

1. A method to compute coronary physiology indexes, characterized in that comprising the acquisition of the coronary angiography images and intravascular images of coronary vessels, the registration of the coronary angiography images and the intravascular images into a high precision registration model, and the calculation of the coronary flow, the fractional flow reserve (FFR) and the index of microcirculation resistance (IMR) based on the high precision registration model.

2. The method of claim 1 for computing coronary physiology indexes based on the high precision registration model, wherein the method of performing co-registration of coronary angiography and intravascular images is realized by placing a radio opaque marker in the imaging catheter that moves together with the transducer, tracking the marker's position and pullback trajectory, locating the coronary vessel positions of the intravascular images in the corresponding coronary angiography, and generating a high precision registration model through signal synchronization and processing.

3. The method of claim 1 for computing coronary physiology indexes based on the high precision registration model, wherein the method for the computation of coronary blood flow based on the high precision registration model comprise selecting a segment of vessel from the high precision registration model, measuring the transit time of the contrast traveling through the vessel segment, calculating the vessel segment volume, and calculating the coronary flow using equation (1): $\begin{array}{cc}Q=\frac{V}{.Math..Math.T}& \left(1\right)\end{array}$ where Q is the coronary blood flow, T is the contrast transit time, and V is the lumen volume.

4. The method of claim 3 for computing coronary physiology indexes based on the high precision registration model, wherein the vessel segment volume V is calculated from the vascular shape obtained by the intravascular images based on the high precision registration model.

5. The method of claim 3 for computing coronary physiology indexes based on the high precision registration model, wherein the method of measuring the contrast transit time T for the vessel segment comprises injecting contrast from the proximal end of the vessel, recording the time of the first frame of the coronary angiography T.sub.1, the time of the second frame T.sub.2, and so forth, calculating the contrast transit time as T=T.sub.dT.sub.p, where T.sub.p is the contrast arriving time at the proximal end of the vessel segment, and T.sub.d is the contrast arriving time at the distal end of the vessel segment.

6. The method of claim 1 for computing coronary physiology indexes based on the high precision registration model, wherein the said fractional flow reserve is computed from the coronary blood flow and vessel morphological parameters measured based on the high precision registration model.

7. The method of claim 1 for computing coronary physiology indexes based on the high precision registration model, wherein the said index of microcirculation resistance is derived from the fractional flow reserve.

8. The method of claim 7 for computing coronary physiology index, where the index of microcirculation resistance is calcualted from equation (2): $\begin{array}{cc}\mathrm{IMR}=\mathrm{FFR}\frac{P_a}{Q}& \left(2\right)\end{array}$ where FFR is fractional flow reserve, P.sub.a is mean arterial pressure, Q is blood flow.

9. The method of claim 1 for computing coronary physiology indexes based on the high precision registration model, wherein the intravascular images comprise intravascular ultrasound images, intravascular optical coherence tomography images, and combined use of intravascular ultrasound images and intravascular optical coherence tomography images.

Description

BRIEF DESCRIPTION OF DRAWINGS

[0022] FIG. 1 is an illustration of the vessel model obtained from the high precision registration of coronary angiography and intravascular imaging.

[0023] FIG. 2 is an illustration of the transit time T of the contrast traveling through the vessel segment.

[0024] FIG. 3 is a schematic of the relationship between the errors of pressure difference calculation and the errors from diameter measurements.

DETAILED DESCRIPTION OF THE INVENTION

[0025] The technical approaches of certain embodiments of the present invention are described in detail and in completeness with the figures in the following description.

[0026] As illustrated in FIG. 1, a method to compute coronary indexes based on a high precision model, comprises acquisition of coronary angiography of coronary vessels, and intravascular images of vessels inside, and registration between coronary angiography and intravascular images into a high precision registration model, and calculation of coronary blood flow, fractional flow reserve and index of microcirculation resistance based on the high precision registration model.

[0027] Coronary angiography and intravascular imaging are two different approaches to estimate the disease severity of coronary arteries. Coronary angiography uses X-ray to generate projections of human body along a certain direction by injecting contrast through vessels, and the output is a projected two dimensional image with the maximum vessel diameter along this direction. Intravascular imaging uses an optical or ultrasound catheter to generate pipe-like circular images over all axial directions inside the vessel.

[0028] Coronary angiography and intravascular imaging complement each other for making a diagnosis of the stenosis of a diseased vessel. Coronary angiography has relatively low resolution, and has limited precision for quantifying the vessel diameter, area and stenosis, and is unable to differentiate between different atherosclerotic plaque types, but can provide the overall morphological information of coronary vascular trees. Intravascular imaging has higher resolution, and is able to compute the vessel area and stenosis precisely, and can effectively differentiate and make a diagnosis of atherosclerotic plaques inside the artery, but is unable to see the overall coronary vascular structures.

[0029] In summary, both coronary angiography and intravascular imaging have certain limitations, and each of them alone cannot perform real precise measurement. Hence, the present invention proposes to use both coronary angiography and intravascular imaging, and methods for achieving a high precision registration between the two images.

[0030] The method to register the coronary angiography and intravascular images into a high precision registration model is described in detail as follows:

[0031] First, a radio-opaque marker is placed on the intravascular imaging catheter, and in the initial stage, by locating the positions of the radio-opaque marker and guide wire in the coronary angiography images, and the insert directions of the guide wire, the possible range of the pullback trajectory of the radio-opaque marker or the guide wire during the subsequent intravascular imaging procedure can be roughly estimated.

[0032] In the second step, the coronary angiography console is turned on, and contrast is injected through the vessels via a catheter, and after the contrast is released, the time-stamped videos of coronary angiography is acquired. By detecting the vessel locations in the coronary angiography images, precise reference of the location information of the radio-opaque marker in the coronary angiography can be obtained. In one embodiment, the vessel location detection can be performed using the eigenvalues of Hessian matrices or other filtering methods.

[0033] In the third step, the locations of the vessels and the guide wire determined from the previous steps provide a rough range of the possible radio-opaque marker positions. The next step is to precisely detect the radio-opaque marker pullback trajectory. One embodiment is to use a matched filter to detect the radio-opaque position in every frame of the coronary angiography. The matched filter can be designed based on the unique features of the radio-opaque marker from pre-acquired coronary angiography images. In another embodiment, an objective function is used to locate the radio-opaque marker, and the optimal trajectory in the time-stamped coronary angiography images is determined using graph-cuts or Markov chain or Bayesian methods by globally optimizing the accumulated objective function. Another embodiment is to select one or multiple frames of coronary angiography images after contrast injection and manually mark the radio-opaque marker positions, and determine the optimal pullback trajectory using livewire or intelligent scissor algorithms.

[0034] In the fourth step, using the optimal pullback trajectory determined from the previous step, registration between the intravascular images and coronary angiography is completed, and every frame of the intravascular images is matched to a location in the corresponding coronary angiography frame.

[0035] One embodiment of the method to compute the coronary blood flow based on the high precision registration model is to select a vessel segment from the high precision registration model, and measure the transit time of the contrast traveling through the vessel segment, and obtain the lumen volume of the vessel segment in the high precision registration model, and compute coronary blood flow using equation (1):

[00003]$\begin{array}{cc}Q=\frac{V}{.Math..Math.T}& \left(1\right)\end{array}$

As illustrated in FIG. 1, the lumen volume V is calculated from the morphological parameters measured using intravascular imaging based on the high precision registration model.

[0036] As illustrated in FIG. 2, one embodiment of the method to measure the contrast transit time T inside the vessel segment is to inject contrast at the proximal end of the vessel, and record the time of the first frame of the coronary angiography T.sub.1, the time of the second frame T.sub.2, and so forth. Then the contrast transit time is obtained as T=T.sub.dT.sub.p, where T.sub.p is the contrast arriving time at the proximal end of the vessel segment, and T.sub.d is the contrast arriving time at the distal end of the vessel segment. Based on the registration methods between coronary angiography and intravascular imaging described previously, the three dimensional locations of the vessel L.sub.p and L.sub.d in the intravascular images corresponding to the contrast leading edge at T.sub.p and T.sub.d, respectively, can be obtained. The lumen volume between L.sub.p and L.sub.d can be determined based on the three dimensional models of the vessel. The coronary blood flow Q determined from equation (1) can be used in subsequent calculations of pressure drop and FFR.

[0037] Fractional flow reserve is calculated using equation FFR=P.sub.d/P.sub.a, where the distal end pressure P.sub.d of the target vessel is determined by subtracting the pressure drop P from the proximal end pressure P.sub.a. The pressure drop of a fluid after passing through a pipe consists of the pressure drop from friction alone the path, gravity, acceleration and local resistance. In normal vessels, the friction pressure drop is the dominant factor for laminar flow. Assume the vessel length is L, vessel dimeter is d, and blood viscosity is , blood flow is Q (obtained previously), according to the Poiseuille's law, the pressure drop along the path takes the following form:

[00004]$.Math..Math.P=\frac{64.Math..Math.\mathrm{.Math.}.Math..Math.\mathrm{QL}}{.Math..Math.d^4}$

Therefore, in order to accurately calculate the pressure drop, it is necessary to precisely determine the blood flow Q, the vessel length L, and the vessel diameter d. In particular, the precision of vessel diameter d is of paramount importance. FIG. 3 illustrates the relationship between the computation errors of pressure drop and the measurement errors of diameter. The resolution of coronary angiography is around 0.5 mm, and the resulting computation errors of the pressure drop are significant, indicating that the result based on coronary angiography alone is unreliable. Intravascular imaging methods such as OCT with a resolution around 0.02 mm is able to control the computation errors of the pressure drop well. But because the penetration depth of OCT is limited, and blood clearance is required for imaging, it is sometimes challenging to acquire high quality images at all locations. On the other hand, IVUS does not require blood clearance during imaging, and the combination of OCT and IVUS can provide better intravascular imaging results.

[0038] Specifically, there are usually two ways to compute pressure drop. The first method is analytical, which divides the target vessel into small segments according to certain standard, and determine the overall pressure drop by summing over all the pressure drop from individual segments. The other method is numerical, based on computational fluid dynamic analysis, the pressure drop of the vessel segment is determined from calculating the pressure and flow of every unit volume inside the vessel using standard finite element analysis methods.

[0039] As illustrated in FIG. 2, the lumen volume between L.sub.p and L.sub.d and the pressure drop calculation method require intravascular imaging to accurately determine vessel area at each cross-section. One embodiment is to locate the frame locations of intravascular images between L.sub.p and L.sub.d, and perform segmentation of intravascular images and determine the lumen borders of the vessel in each frame, based on which reconstruction of the blood vessel model can be conducted, and pressure drop and FFR can be computed from the vessel lumen model utilizing the blood flow Q.

[0040] In one embodiment, the said index of microcirculation resistance is determined from fractional flow reserve.

[0041] In one embodiment, the said index of microcirculation resistance is computed from equation (2):

[00005]$\begin{array}{cc}\mathrm{IMR}=\mathrm{FFR}\frac{P_a}{Q}& \left(2\right)\end{array}$

Where FFR is the fractional flow reserve, P.sub.a is the mean arterial pressure, Q is the blood flow.

[0042] When there is coronary collateral flow that can not be neglected, calculation of IMR should be corrected using equation (3):

[00006]$\begin{array}{cc}{\mathrm{IMR}}^{}=\frac{P_a\left(P_d-P_w\right)}{Q_d\left(P_a-P_w\right)}={\mathrm{FFR}}_{\mathrm{cor}}\frac{P_a}{Q_d}& \left(3\right)\end{array}$

Where P.sub.w is the coronary wedge pressure, and is typically determined during coronary balloon angioplasty, or measured using a pressure wire at the distal end of the coronary artery after it is totally occluded. FFR.sub.cor is the radio between the distal end pressure by considering only the stenosis of the coronary artery and the mean arterial pressure P.sub.a. P.sub.d is mean venous pressure.

[0043] The embodiments described above are only part, but not all, of the possible embodiments of the present invention. The embodiments based on the present invention, and all other embodiments generated by regular technical people in the relevant field without creative work, are within the scope of this invention.