Quantitative Analysis Of Myocardial Perfusion Magnetic Resonance Image

Over the past two decades, the morbidity of cardiac disease is risingconstantly, along with the accelerated pace of life. Cardiovascular diseasehas already become one of the deadliest diseases. The development andmaturity of nuclear magnetic resonance provides the diagnosis andtreatment of cardiovascular disease with critical image evidence and makesit possible to conduct quantitative analysis of myocardial perfusion. Theresearch of the magnetic resonance image of myocardial perfusion is ofgreat significance in the treatment and prognosis of cardiovascular disease.The quantitative analysis of magnetic resonance image of myocardialperfusion serves as the basis of computer-aided treatment and prognosis ofcardiovascular disease.Traditionally, a semi-quantitative method was adopted thequantitative analysis of magnetic resonance image of myocardial perfusion,with certain parameters derived from myocardium contrast agent fillingspeed curve as its analytic object. It did not take into consideration theimpact on myocardium brought by the filling of contrast agent into the leftventricle. Thus, without reference object, the parameters could not bequantitatively analyzed, and the relationship between the parameters andthe blood flow volume was merely defined through experience andobservation.This paper adopts an absolute quantitative analysis method, withcentral volume principle as its theoretical basis. First, it amplifies thedifference value of the magnetic resonance image of myocardial perfusionobtained through clinical acquisition, and then conducts image registration and disposes the artifacts and noise of the image by using the meanfiltering method, and thus obtains the signal intensity curves of myocardialperfusion and left ventricular perfusion. Then it decouples the perfusioncurves and establishes the related mathematical model-based transferfunction, also called residual function, as well as identifies the parametersof the mathematical model by the nonlinear optimization method.The choice of mathematical model is the key of absolute quantitativeanalysis. Some researchers choose the spline model which can fit well withthe residual function. However, this spline model contains a huge numberof parameters, which are difficult to be compared and analyzed. Someother researchers select a Fermi model to match the residual function, butthe parameters of Fermi model is lack of practical physical significance.This paper adopts the exponential model to match the residual function.The exponential model, containing two parameters, not only decreases thevolume of parameters, but also increases their physical significance. Thispaper conducts analysis and comparison of the parameters of theexponential model. Through comparing with the traditionalsemi-quantitative method, it reveals the stability of exponential modelparameters. In addition, by comparing the different parameters of normaland abnormal people, it proves the significance of exponential modelquantitative method in aiding clinical diagnosis. Experimental results showthat the method proposed in this paper is of good reliability andsignificance in aiding clinical diagnosis, and therefore has clinical researchvalue.According to the demand of the magnetic resonance image analysis ofmyocardial perfusion in clinical practice, this paper designs software forquantitative analysis. The input of the software is myocardial perfusionmagnetic resonance image data acquired through clinical practice, whilethe output is graphical interface of the quantitative analysis of theparameters. The software is composed of five modules: DICOM fileparsing, image registration, image processing, parameter identification and user interface. This software, whose design framework is based on MFC,adopts a combined programming method of C++and MATLAB, andprovides the doctor with diagnostic information of the patient in anintuitive, eye diagram-data combined manner.