Periodic Reporting for period 2 - CardioZoom (High-fidelity Cardiovascular Modeling from Super-Fast Magnetic Resonance Imaging)
Reporting period: 2021-10-01 to 2023-03-31
Measuring detailed information on the mechanical properties of cardiovascular tissue on an individual patient basis could greatly improve the diagnosis, treatment, and prevention of cardiovascular diseases. These properties are often obtained from measurements of tissue motion, which are used to estimate the parameters of coupled fluid-solid mathematical and computational models.
In the CardioZoom project, the issue being addressed is the lack or low quality of kinematic information for thin cardiovascular structures such as cardiac valves and the arterial wall when measured from magnetic resonance imaging (MRI).
The overall objective is to propose a new paradigm in biophysical computational models of the cardiovascular system, which would replace long MRI scans with short time scans and obtain reliable information on delicate cardiac structures.
In the CardioZoom project, the issue being addressed is the lack or low quality of kinematic information for thin cardiovascular structures such as cardiac valves and the arterial wall when measured from magnetic resonance imaging (MRI).
The overall objective is to propose a new paradigm in biophysical computational models of the cardiovascular system, which would replace long MRI scans with short time scans and obtain reliable information on delicate cardiac structures.
There have been three sub-projects carried out with the following achievements:
Sub-project 1 aims to develop blood flow models from raw MRI data for simpler fluid and fluid-solid problems. The team found a new way to estimate the properties of blood flowing in the body using MRI measurements, including artifacts in the images. They also started developing a way to use raw MRI measurements to estimate blood flow properties, tested this approach on simulated data, and will soon use experimental data to confirm its accuracy. In addition, they developed a way to accelerate the MRI measurement process and found that they don't need to take many measurements to get accurate results in academic tests. However, further testing is still needed on fluid flow models.
Sub-project 2 aimed to develop new inverse problems for shape estimation of cardiac valves, first from velocity data, and then extending the results of sub-project 1 from raw MRI data. The team made a simple model of the aortic valve that can be used in computer simulations of blood flow and tested it on patients with aortic stenosis using images from CT scans. They also created a new method for finding obstacles in the flow of fluid inside the body, like in the heart. This method adds a term to the equations that makes it easier to find these obstacles, even when there is limited information from MRI or ultrasound scans. They tested this method on simulated heart valves.
Sub-project 3 aimed to develop new inverse problems for cardiac mechanics, first from velocity data, and then extending the results of sub-project 1 from MRI data. The team developed a new algorithm that can simulate how fluids and solids interact with each other in large, deformable structures like the heart. They tested the program and found that it works well for modeling how blood flows through the heart. They also created a new computer method that can estimate the properties of fluids and solids in the heart by analyzing how they interact with each other. They found that when they include measurements of how the fluid flows, they can get much better results than with just measurements of the solid's motion.
Sub-project 1 aims to develop blood flow models from raw MRI data for simpler fluid and fluid-solid problems. The team found a new way to estimate the properties of blood flowing in the body using MRI measurements, including artifacts in the images. They also started developing a way to use raw MRI measurements to estimate blood flow properties, tested this approach on simulated data, and will soon use experimental data to confirm its accuracy. In addition, they developed a way to accelerate the MRI measurement process and found that they don't need to take many measurements to get accurate results in academic tests. However, further testing is still needed on fluid flow models.
Sub-project 2 aimed to develop new inverse problems for shape estimation of cardiac valves, first from velocity data, and then extending the results of sub-project 1 from raw MRI data. The team made a simple model of the aortic valve that can be used in computer simulations of blood flow and tested it on patients with aortic stenosis using images from CT scans. They also created a new method for finding obstacles in the flow of fluid inside the body, like in the heart. This method adds a term to the equations that makes it easier to find these obstacles, even when there is limited information from MRI or ultrasound scans. They tested this method on simulated heart valves.
Sub-project 3 aimed to develop new inverse problems for cardiac mechanics, first from velocity data, and then extending the results of sub-project 1 from MRI data. The team developed a new algorithm that can simulate how fluids and solids interact with each other in large, deformable structures like the heart. They tested the program and found that it works well for modeling how blood flows through the heart. They also created a new computer method that can estimate the properties of fluids and solids in the heart by analyzing how they interact with each other. They found that when they include measurements of how the fluid flows, they can get much better results than with just measurements of the solid's motion.
For each of the parts of CardioZoom, there are achievements beyond the state that can be highlighted are:
We have introduced general theory for measuring tissue motion from MRI, increasing the quality of the images and hence potentially considerably improving clinical diagnosis. This is wide applications in several types of motion-encoding (flow, elastography).
We have also developed new ways of measuring missing geometrical anatomical structures in MRI images by analysing the effect they have on the flow surrounding it. These methods also include the capability of handling important artifacts in the images.
We have introduced general theory for measuring tissue motion from MRI, increasing the quality of the images and hence potentially considerably improving clinical diagnosis. This is wide applications in several types of motion-encoding (flow, elastography).
We have also developed new ways of measuring missing geometrical anatomical structures in MRI images by analysing the effect they have on the flow surrounding it. These methods also include the capability of handling important artifacts in the images.