Model-based computation of blood pressure from tomographic data
Motivation and goals
In cardiovascular diagnostics, the knowledge of blood pressure can be of great help for the physician. Nowadays, blood pressure values in the vessels close to the human heart can only be obtained by the use of invasive catheter measurements, which lead to increased risk of injury and infection for the patient.
The goal of this project is to build a process chain for pressure data assessment from phase contrast magnetic resonance imaging (PC-MRI). The developed methods should provide robust and fast computations of patient-specific, spatially and temporally resolved blood pressure values of the blood vessels close to the human heart.
Initially, we use PC-MRI to measure blood flow velocities of the particular patient. After treatment of imaging artifacts, a semi-automatic method for segmentation of the relevant blood vessels is applied. The following computation of blood pressure values is divided into two major steps. First, we compute pressure gradients from the measured blood flow velocities by using discretized physical equations from fluid mechanics (Navier-Stokes equations). During these computations, additional physical a-priori knowledge about the blood flow is incorporated. Thereby we’re able to minimize the effect of image noise and other measurement errors of MRI. In a second step the blood pressure values can be obtained from the pressure gradients by applying numerical integration methods.
We examine the performance of the developed methods in cooperation with radiologists from the University Hospital Heidelberg and the German Cancer Research Center, Heidelberg. Experiments include MRI measurements of a flow phantom as well as of healthy subjects and patients suffering from cardiovascular diseases.