WO2021253294A1

WO2021253294A1 - Ct image separation and reconstruction method and application thereof

Info

Publication number: WO2021253294A1
Application number: PCT/CN2020/096635
Authority: WO
Inventors: 郑海荣; 李彦明; 万丽雯; 胡战利; 陈子翔
Original assignee: 深圳高性能医疗器械国家研究院有限公司
Priority date: 2020-06-17
Filing date: 2020-06-17
Publication date: 2021-12-23

Abstract

A CT image separation and reconstruction method, comprising the following steps: step 1: separating, from a dynamic perfusion scanning image, a contrast-enhanced scanning image introduced by an imaging agent; step 2: reconstructing an enhanced image by using the contrast-enhanced scanning image; and step 3: fusing the enhanced image with a baseline image obtained by means of reconstructing a full-sampling baseline image collection image, so as to obtain a perfusion CT reconstruction image. Artifacts in a fused dynamic perfusion image can be eliminated to a large extent.

Description

A Method and Application of Separation and Reconstruction of CT Images

Technical field

This application belongs to the technical field of medical CT imaging, and particularly relates to a method and application of CT image separation and reconstruction.

Background technique

CT is a full-featured disease detection instrument, which is the abbreviation of electronic computer tomography technology. Scanners used in computed axial tomography (CAT) can generate X-rays, which are powerful electromagnetic energy. The photons of X-rays are basically the same as those of ordinary visible light, but they carry more energy. This higher energy level allows X-rays to pass directly through most soft tissues of the human body (please refer to X-rays to understand the principle of X-rays penetrating soft tissues and how X-ray machines generate X-ray photons). Conventional X-ray imaging technology uses the principle of light and shadow. The "light" is irradiated from one side of the human body. At this time, the film on the other side of the human body can record the outline of the bones.

Myocardial perfusion imaging is mainly suitable for the diagnosis of coronary heart disease, risk stratification, and efficacy judgment. This examination has very important application value. It is different from general CTCA or CAG coronary angiography. It judges whether there is coronary heart disease from the morphological changes. Myocardial perfusion imaging reflects whether the myocardium is ischemia, and can accurately diagnose whether the myocardium is ischemia. The principle is to observe the blood flow reserve of the myocardium through interventional tests, drug load, or exercise load, which can accurately determine whether the patient has coronary heart disease, as well as the location and degree of coronary heart disease, so that the patient can be diagnosed and the method should be adopted. Provide valuable information. Therefore, myocardial perfusion imaging has important applications for coronary heart disease diagnosis, treatment decision-making, and risk stratification. Using this method can achieve accurate diagnosis and treatment of coronary heart disease.

Myocardial dynamic perfusion CT imaging is a clinically important medical imaging technique for non-invasive examination of heart disease. However, multiple scans of the patient after the imaging agent reaches the pedestrian's myocardial tissue will inevitably lead to concerns about excessive radiation dose. The sparse angle X-ray scanning scheme can significantly reduce the radiation dose caused by repeated scanning, but there is no effective and feasible method for the image reconstruction of the sparse angle image acquisition data.

Summary of the invention

1. Technical problems to be solved

Based on the existing dynamic perfusion CT image reconstruction method, image reconstruction can only be performed based on the full-sampled scan data. In the process of rescanning the same part of the patient, there is a large accumulation of radiation dose in a short period of time, resulting in unknowable X-rays For the problem of radiation risk, this application provides a method and application for the separation and reconstruction of CT images.

2. Technical solution

In order to achieve the above objective, the present application provides a method for separating and reconstructing CT images. The method includes the following steps:

Step 1: Separate the contrast-enhanced scan image introduced by the imaging agent from the dynamic perfusion scan image;

Step 2: Use the contrast-enhanced scanned image to reconstruct an enhanced image;

Step 3: Fusion the enhanced image with the baseline image reconstructed from the full-sample baseline image acquisition image to obtain a perfusion CT reconstruction image.

Another implementation manner provided by this application is that the dynamic perfusion scan image in step 1 includes a fully sampled baseline scan image, a sparse angle perfusion scan image, and a down-sampled sparse baseline scan image.

Another embodiment provided by this application is that: the fully sampled baseline scan image is acquired by using an X-ray scan scheme with a normal radiation dose before the imaging agent reaches the heart tissue; the sparse angle perfusion scan image is After the imaging agent reaches the heart tissue, the low-dose image is acquired at a sparse angle.

Another implementation manner provided by this application is that the down-sampled sparse baseline scan image is to downsample the fully sampled baseline scan image according to the sparse angle scheme of the perfusion scan, and retain the same scan angle as the sparse perfusion scan image. The baseline scan image is obtained.

Another implementation manner provided by this application is that the down-sampled sparse baseline scan image and the perfusion scan image of each frame have the same image size, and the corresponding scan angle is the same.

Another implementation manner provided by this application is that the contrast-enhanced scan image in step 1 includes a contrast-enhanced scan image with a sparse angle.

Another implementation manner provided by the present application is that the sparse angle contrast-enhanced scan image is obtained by subtracting the sparse angle perfusion scan image and the down-sampled sparse baseline scan image.

Another implementation manner provided by the present application is that the enhanced image includes normal-dose baseline image reconstruction, low-dose contrast enhanced image reconstruction and dynamic perfusion image fusion.

Another implementation manner provided by this application is: the normal-dose baseline image reconstruction adopts a commercial reconstruction algorithm, and the low-dose contrast-enhanced image reconstruction adopts a statistical iterative algorithm; the commercial reconstruction algorithm includes a filtered back projection or a traditional algebraic iterative algorithm The statistical iterative algorithm includes the maximum likelihood-expected value maximization of full variational regularization or the weight penalty least squares method of full variational regularization.

The present application also provides an application of a method for separation and reconstruction of CT images, which is applied to image reconstruction of low-dose brain dynamic perfusion CT imaging or low-dose myocardial dynamic perfusion CT imaging.

3. Beneficial effects

Compared with the prior art, the method for separating and reconstructing CT images provided by this application has the following beneficial effects:

The CT image separation and reconstruction method provided in this application is an image acquisition scheme and reconstruction method suitable for sparse angle low-dose myocardial dynamic perfusion CT imaging.

The CT image separation and reconstruction method provided in this application is a low-dose dynamic perfusion scanning scheme and its subsequent image separation and reconstruction method.

The method for separating and reconstructing CT images provided in the present application effectively reduces the radiation dose in the process of myocardial dynamic perfusion imaging, and uses sparse angle image acquisition data to reconstruct a myocardial perfusion CT image that meets the requirements of clinical diagnosis.

The separation and reconstruction method of CT images provided in this application can solve the problem of excessive radiation dose caused by repeated continuous X-ray scanning in dynamic perfusion CT imaging and image artifacts caused by the reconstruction of sparse angle image acquisition data using traditional methods problem.

The method for separating and reconstructing CT images provided by this application separates and reconstructs baseline scan data and contrast-enhanced scan data, while reducing the radiation dose of continuous perfusion scans, and making full use of the image structure information in the normal-dose baseline image to help dynamics Reconstruction of perfusion image. The process of using sparse angle contrast-enhanced scan data to reconstruct contrast-enhanced images is less prone to image artifacts than the process of directly reconstructing perfusion images using unprocessed sparse perfusion scan data, so artifacts in the fused dynamic perfusion image Can be eliminated to a large extent.

Description of the drawings

Figure 1 is a schematic diagram of the overall flow of the CT image separation and reconstruction method of the present application;

Figure 2 is a schematic diagram of the experimental results of the dynamic perfusion CT separation and reconstruction method of the present application.

detailed description

Hereinafter, specific embodiments of the application will be described in detail with reference to the accompanying drawings. According to these detailed descriptions, those skilled in the art can clearly understand the application and can implement the application. Without violating the principle of the present application, the features in different embodiments can be combined to obtain new implementations, or replace some features in some embodiments to obtain other preferred implementations.

Clinical myocardial dynamic perfusion CT imaging of the patient’s image acquisition includes baseline scans before the imaging agent reaches the heart tissue (including the aorta, coronary arteries, left and right ventricles, left and right atriums and myocardial tissue, etc.) and after the imaging agent reaches the heart tissue Continuous dynamic perfusion scan. Under the existing clinical conditions, the baseline scan and the dynamic perfusion scan use the same full sampling scan plan, which brings a larger X-ray scan radiation dose to the patient.

A sample sequence is sampled once at intervals of several samples, so that the new sequence obtained is a down-sampling of the original sequence.

Referring to Figures 1-2, the present application provides a method for separating and reconstructing CT images. The method includes the following steps:

Before the imaging agent reaches the heart tissue, use the normal radiation dose X-ray scanning plan to collect the baseline image of the patient to obtain the baseline scan data y _{baseline, full sampling} , after the imaging agent reaches the heart tissue, use the low sparse angle The dose image acquisition program scans the patient continuously to obtain the sparse angle perfusion scan data Y _{perfusion, sparse} . The full-sampling baseline scan data is down-sampled according to the sparse angle scheme of the perfusion scan, and the baseline scan data with the same scan angle as the sparse perfusion scan data is retained to ensure the down-sampled baseline scan data y _{baseline, sparse} and perfusion of each frame Scan data

Have the same data size, and the corresponding scan angles are the same. Subtract the perfusion scan data of the sparse angle with the sparse baseline scan data after downsampling to obtain the contrast-enhanced scan data of the sparse angle

The above-mentioned image data collection process can be expressed as:

in,

For the time series of dynamic perfusion scan data, the same goes for

Time series of scan data for contrast enhancement. T is the number of time frames of the dynamic perfusion image.

Further, the dynamic perfusion scan image in the step 1 includes a fully sampled baseline scan image, a sparse angle perfusion scan image, and a down-sampled sparse baseline scan image.

Further, the fully sampled baseline scan image is acquired using an X-ray scanning scheme with a normal radiation dose before the imaging agent reaches the heart tissue; the sparse angle perfusion scan image is obtained after the imaging agent reaches the heart tissue Obtained by low-dose image acquisition with sparse angle.

Further, the down-sampled sparse baseline scan image is obtained by down-sampling the fully sampled baseline scan image according to the sparse angle scheme of the perfusion scan, and retains the baseline scan image with the same scan angle as the sparse perfusion scan image.

Further, the down-sampled sparse baseline scan image and the perfusion scan image of each frame have the same image size, and the corresponding scan angle is the same.

Further, the contrast-enhanced scan image in the step 1 includes a contrast-enhanced scan image with a sparse angle.

Further, the sparse-angle contrast-enhanced scan image is obtained by subtracting the sparse-angle perfusion scan image and the down-sampled sparse baseline scan image.

Further, the enhanced image includes normal-dose baseline image reconstruction, low-dose contrast enhanced image reconstruction and dynamic perfusion image fusion.

Further, the normal-dose baseline image reconstruction uses a commercial reconstruction algorithm, and the low-dose contrast-enhanced image reconstruction uses a statistical iterative algorithm; the commercial reconstruction algorithm includes filtered back projection or a traditional algebraic iterative algorithm, and the statistical iterative algorithm includes full The maximum likelihood-expected value maximization of variational regularization or the weight-penalized least squares method of total variational regularization.

The reconstruction of the baseline image is to use traditional commercial reconstruction algorithms (filtered back-projection, traditional algebraic iterative algorithm, etc.) to reconstruct the full-sampled baseline scan data y _{baseline, full-sampling} to obtain the normal dose baseline image x _{baseline, normal dose} . Contrast-enhanced image reconstruction is to use a regularized statistical iterative algorithm with image artifact suppression capabilities (full variational regularization maximum likelihood-expected value maximization (MLEM-TV), full variational regularization weight penalty minimum two Multiplication (PWLS-TV), etc.) contrast enhancement scan data for sparse angles

Perform reconstruction to obtain low-dose contrast-enhanced images

Dynamic perfusion image fusion is the reconstructed

Add with x _{baseline and normal dose to} get low-dose dynamic perfusion image

The above image reconstruction process can be expressed as:

The full-sampling baseline image only needs to be reconstructed once, and it is contrast enhanced with the contrast-enhanced image time series X. _{Each frame of the low-dose} image is added separately to obtain the myocardial dynamic perfusion CT image time series X _{perfusion, low-dose} .

The separation and reconstruction method of dynamic perfusion CT images has been verified by computer simulation experiments, and the effect is obvious:

Fig. 2 shows the computer simulation experiment result of the method for separating and reconstructing the myocardial dynamic perfusion CT image proposed by the present invention. The baseline image scanning scheme is 984 frames/360°, and the perfusion image scanning scheme is 41 frames/360°. It can be seen from the results that the separation reconstruction algorithm proposed by the present invention can correctly reconstruct the sparse angle low-dose myocardial dynamic perfusion CT image, and there is nothing in the reconstructed image. Obviously sparse artifacts.

Although the application is described above with reference to specific embodiments, those skilled in the art should understand that within the principles and scope disclosed in the application, many modifications can be made to the configuration and details disclosed in the application. The protection scope of this application is determined by the appended claims, and the claims are intended to cover all modifications included in the literal meaning or scope of equivalent technical features in the claims.

Claims

A method for separating and reconstructing CT images, characterized in that: the method includes the following steps:

Step 1: Separate the contrast-enhanced scan image introduced by the imaging agent from the dynamic perfusion scan image;

Step 2: Use the contrast-enhanced scanned image to reconstruct an enhanced image;

Step 3: Fusion the enhanced image with the baseline image reconstructed from the full-sample baseline image acquisition image to obtain a perfusion CT reconstruction image.
The method for separating and reconstructing CT images according to claim 1, wherein the dynamic perfusion scan image in step 1 includes a fully sampled baseline scan image, a sparse angle perfusion scan image, and a sparse baseline after downsampling Scan the image.
The method for separating and reconstructing CT images according to claim 2, characterized in that: the fully sampled baseline scan image is acquired by using an X-ray scan scheme with a normal radiation dose before the imaging agent reaches the heart tissue; the sparse The angled perfusion scan image is acquired by using a sparse angle low-dose image after the imaging agent reaches the heart tissue.
The method for separating and reconstructing CT images according to claim 2, wherein the sparse baseline scan image after downsampling is to downsample the fully sampled baseline scan image according to the sparse angle scheme of the perfusion scan, and retain and sparse The perfusion scan image is obtained from the baseline scan image with the same scan angle.
The method for separating and reconstructing a CT image according to claim 2, wherein the down-sampled sparse baseline scan image and the perfusion scan image of each frame have the same image size, and the corresponding scan angle is the same.
The method for separating and reconstructing a CT image according to claim 1, wherein the contrast-enhanced scan image in the step 1 includes a contrast-enhanced scan image with a sparse angle.
8. The method for separating and reconstructing CT images according to claim 6, wherein the sparse angle contrast-enhanced scan image is obtained by subtracting the sparse angle perfusion scan image and the down-sampled sparse baseline scan image.
The method for separating and reconstructing CT images according to claim 1, wherein the enhanced image includes normal-dose baseline image reconstruction, low-dose contrast enhanced image reconstruction and dynamic perfusion image fusion.
8. The method for separating and reconstructing CT images according to claim 8, wherein the normal-dose baseline image reconstruction uses a commercial reconstruction algorithm, and the low-dose contrast-enhanced image reconstruction uses a statistical iterative algorithm.
An application of a method for separation and reconstruction of CT images, characterized in that the method is applied to image reconstruction of low-dose brain dynamic perfusion CT imaging or low-dose myocardial dynamic perfusion CT imaging.