MICCAI 2016 Daily - Wednesday

Mariana: What usually happens with cardiac patients is that they have some sort of symptoms, and they go to the doctor. Then the doctor gives them an ultrasound to see if there is a problem. Then, based on those findings, the doctor decides if they need to take an MRI. These are not emergency situations in which the patient is going to die in a day. Diseases of the heart take time, and it is very important to be able to detect them with time to be able to see before the person becomes really sick. Let’s look at what is happening with the flow in this person’s heart to see if we can do something to make this person survive longer than he or she would have. MICCAI Daily: How did you solve these challenges? What algorithmic or image processing techniques did you use? Mariana: Right now what I came to present at MICCAI is a way to extract angiographic data from this type of 4D flow MRI. The typical way of extracting angiographic data from 4D flow MRI is to average this 4-dimensional image into one 3- dimensional image. That 3D image is sort of the volume. It is static. Since it’s an average, you don’t know to which time frame within the cardiac cycle that this image belongs. What I came here to present is a way to calculate a 4D angiography from 4D flow MRI. It is an angiography, so it contains information about where the vessels are inside the cardiovascular system and where the heart is in the cardiovascular system. It’s also in 4D so the heart is moving. It’s beating. There is also motion vessels, which is small motion, but it is not usually taken into account when vessel analysis is done, at least in 4D flow MRI. MICCAI Daily: Are you also taking clinical tests? Mariana: The tests that we have done so far have been on healthy volunteers, but it is a technique which can for sure be applied to diseased patients as well. MICCAI Daily: What is the next step of your work? Mariana: The thing that we want to do with my PhD project is to do segmentation of the heart and vessels. We want to do it automatically. That is the goal to use these types of visualizations and analysis which we have done previously to somehow concoct a new way of segmenting the heart automatically and in 4D, keeping the motion of the heart. Since the type of imaging that we are using is 4D flow MRI, it would be specific for 4D flow MRI. It’s just a different type of acquisition technique that is not necessarily what most people in cardiac MRI are using. “ The challenge is how do we make this data useful, quickly. We want to make it as automatic as possible or entirely automatic ” Our Pick for Wednesday 6 MICCAI Daily: Wednesday