Skip to content


Medical Imaging

WASSABI: Automatic brian Structures Segmentation As potential imaging Biomarkers

Project reference: TIN2015-73563-JIN
Project Leader
: Arnau Oliver i Malagelada
Budget: 199.892,00€
Duration: 01/12/16 – 30/11/19
Research team: Arnau Oliver, Xavier Lladó

Background: In Spain there are more than 1,200,000 people and families affected by a neurodegenerative disease, being Alzheimer, Parkinson, multiple sclerosis, amyotrophic lateral sclerosis, and neuromuscular diseases the most important ones. Around 500.000 people is suffering schizophrenia in Spain, while epilepsy cases in Spain reached 400,000 affected. Rapidly diagnosing and effective treatment of all these diseases is important in health, social, and economic terms. Accurate volumetric measurement of brain structures, which can be done using magnetic resonance imaging, is a crucial step to infer a proper diagnosis and to study the evolution of all these diseases.

Main hypothesis: The main hypothesis of wASSABI is that the accurate segmentation of brain structures will provide robust imaging biomarkers providing a better prognosis for multiple sclerosis, epilepsy, and schizophrenia. Robust volumetric measures of structures like white and grey matter, cerebellum, cortex, hippocampi, hypothalamus, corpus callosum, or substantia innominate are crucial to predict the clinical evolution of these diseases. In order to transfer these tools the clinical practice there is the need of new technologies and computerised tools, not currently available nowadays, capable of providing precise and robust measures for the evaluation of the atrophy and evolution of brain structures in a reproducible and efficient way.

Objective: The main objective of wASSABI is to investigate, develop, validate, and implement in the clinical practice fully automated and robust tools to measure the volume changes of brain structures. This objective could be divided into two. Firstly, we will study and compare in a single time point (transversal study) the structures volume of healthy and diseased brains, in order to extract the most significant structures affected by the disease. These structures should be different according to the disease studied. In a second step, the longitudinal changes of these structures will be analysed, providing accurate and reliable structure volume loss over time. This factor will also help on understanding the patient evolution and the effects of the patient treatments.

Methodology: The objectives of wASSABI have been divided into four essential blocks that can be considered as the major milestones of this proposal: data collection, development of tools for brain structures segmentation and volumetry, analysis of brain structures biomarkers, and validation of the tools and study of the correlation between MRI biomarkers, and clinical records. A final step is added for the transference of the prototype to each clinical centre.

Expected results: The technology developed in wASSABI will provide objectivity and a set of biomarkers which would simplify the everyday clinical practice for radiologists and neurologists, providing an step forward towards personalised treatments and medicine for multiple sclerosis, schizophrenia, and epilepsy patients. At a long-term result, this project could establish the roots for a possible use of magnetic resonance imaging as a screening tool to predict the potential disease of current healthy brains.

Share it!

Medical Imaging Projects



UW vision and Robotics


UW vision and Robotics


UW vision and Robotics


Educational Robotics