Coronary artery lesion functional quantitative method based on deep learning and neutrosophy theory

Abstract

The invention provides a coronary artery lesion functional quantitative method based on deep learning and a neutrosophy theory, which belongs to the field of biomedicine and combines respective characteristics of CTA and CAG images. The method comprises: firstly using the CTA image to obtain three-dimensional image data of each cardiac cycle of the coronary artery; then registering and projecting the three-dimensional image data to the coronary artery area of the CAG image, and determining the position of the coronary artery stenosis lesion area according to the information of a contrast agent. According to the method, the defect of the accuracy of reconstructing the blood vessel only by the CAG two-dimensional image is overcome, the coronary artery area can be accurately and automatically identified, the flow velocity of the blood in different coronary artery areas is calculated according to the tracking route of the contrast agent and the time of the CAG video sequence, the blood flow of each point of the coronary artery is calculated. Therefore, the ratio of the far-end blood flow to the near-end blood flow of the lesion is obtained, FFRCAD of computer-aided diagnosis is obtained, comprehensive evaluation of coronary artery lesion function is achieved, and the problem of non-invasive quantitative measurement of coronary artery lesion function is solved.

Description

Technical field[0001]The present invention belongs to the field of biomedical, and specifically involves a determination method for coronary artery disease based on deep learning and smart theory.Background technique[0002]Since 1990, the study of Coronary Artry Angiography, CAGs, and their disease diagnosis have gradually been widely concerned by scholars at home and abroad. However, due to the limitations of the cardiac special motion mode and image shutter speed, the collected image adjacent frame is different, and the gradation of the CAG image is low, and the coronary branch overlaps and intersects. Accurate detection and measurement have become a problem that has been plaguing researchers at home and abroad. At present, coronary assault is mainly summarized as seven categories:[0003](1) Traditional mode identification method: It is the automatic detection and classification of blood vessels or vascular characteristics by using traditional model recognition means, and more commo...

AI Technical Summary

Problems solved by technology

But the tracking method is less accurate due to overlapping heart beat patterns and crossing blood vessels

[0005] (3) Model method: it is a method of extracting blood vessels with a model or template that can clearly describe the characteristics of blood vessels. The disadvantage of this method is that it requires more manual participation, and the edge or centerline of the blood vessel that is relatively close will interfere, and the amount of calculation is large. , time-consuming

[0013] Coronary CTA imaging obtains indirect coronary images with hundreds of cross-sectional reconstructions, which cannot fully simulate the real state of maximum coronary congestion. The accuracy of judging the functional significance of severe calcified lesions and tortuous lesions is low, and the model reconstruction And the calculation time is long, and it cannot be used online in real time in the cath lab

In addition, CTA can only observe the static characteristics of coronary arteries, and cannot provide real-time dynamic coronary information

[0014] Although FFR QCA and QFR have improved the diagnostic performance of coronary angiography, they still have their limitations: ① adenosine is still needed to induce the coronary artery to reach the maximum hyperemia state when obtaining angiography images; ② all side branch vessels need to be reconstructed, branch The overlap and intersection of blood vessels have a great influence on the results, and the use is relatively complicated; ③The accuracy of functional judgment of branch vascular lesions is low

[0015] In view of the lack of clinical imaging methods for non-invasive real-time evaluation of coronary artery stenosis function, the purpose of the present invention is to develop a quantitative analysis system and platform for coronary artery lesion function of CTA and CAG radiomics based on deep learning and neutrosophic theory , to solve key problems such as automatic detection of coronary artery lesions and functional quantitative evaluation

Images