L’angiografia per ressonància magnètica (MRA) és capaç de capturar variacions en les artèries cerebrals, ja siguin morfològiques, geomètriques o de densitat, amb alta resolució. Aquestes variacions s'utilitzen per a identificar els factors de risc de malalties cerebrals de manera precoç i, per aquest motiu, és necessari el coneixement morfològic dels vasos sanguinis en pacients sans. Per tant, l'atles estadístic del cervell proporcionat per aquest projecte pot ser un recurs de neuroimatge útil per a identificar aquests canvis vasculars.
La angiografía por resonancia magnética (MRA) es capaz de capturar variaciones en las arterias cerebrales, ya sean morfológicas, geométricas o de densidad, con alta resolución. Estas variaciones se utilizan para identificar los factores de riesgo de enfermedades cerebrales de manera precoz y, por este motivo, es necesario el conocimiento morfológico de los vasos sanguíneos en pacientes sanos. Por lo tanto, el atlas estadístico del cerebro proporcionado por este proyecto puede ser un recurso de neuroimagen útil para identificar estos cambios vasculares.
Magnetic resonance angiography (MRA) is capable of capturing variations in the cerebral arteries, whether morphological, geometric or density, with high resolution. These
variations are used to identify early risk factors for brain diseases and, for this reason,
morphological knowledge of blood vessels in healthy patients is necessary. Therefore, the
statistical brain atlas provided by this project can be a useful neuroimaging resource to
identify these vascular changes.
This atlas is generated using 3D Time-of-Flight (TOF) MRA images of 10 healthy
subjects extracted from the public and free NITRC database. These images have been
pre-processed before being taken for the study, to be able to correct the differences in
intensity between the different datasets acquired, by applying a method of normalization
of intensities.
In order to carry out the analysis of the brain structures, manual segmentation was
first performed using the ITK-Snap software. Each structure was identified and painted,
producing a labeled image, where each brain structure corresponded to a different label.
Then, it was possible to start the automatic analysis of segmented images, focusing on
the features extraction of brain vascular system (arteries and bifurcations that join them)
and optics. The idea was to create algorithms that use biomedical image processing
tools (Insight Segmentation and Registration Toolkit (ITK)) to be applied to all images
automatically.
From this analysis, the aim of this project is to design a statistical atlas that can
serve as a reference in the field of medicine. Several intensity and geometrical features
were computed for several anatomical structures. Descriptive statistics were also reported,
considering all 10 selected NITRC datasets.
To make the atlas really helpful a large dataset is required to fully capture the variability of these small structures in a healthy population, so it would be interesting to
apply the algorithms designed in this project to more brain MRA-ToF images with the
same characteristics in a future work.
gmail.com
