Please use this identifier to cite or link to this item: http://hdl.handle.net/10609/134366
Title: Caracterització i detecció precoç de la ventriculomegalia aïllada a partir de dades pre i post-natals
Author: Bosch Garcia, Meritxell
Tutor: Bonet-Carne, Elisenda  
Burgos-Artizzu, Xavier Paolo  
Others: Prados Carrasco, Ferran  
Abstract: Ventriculomegaly (VMG) is a rare disease that affects between 0.3 and 22 out of every 1000 pregnancies. 25-60% of these are isolated cases, which means that it has no other associated pathology. The evolution of the disease and the effect on neurodevelopment is not predictable. The detection of VMG is done through the measurement of only two parameters: the size of the Left Atrium and the Right Atrium. The aim is to use a database of 81 participants (41 affected and 40 unaffected) to improve characterisation of the condition and the prognosis and to provide an early detection to predict postnatal effects based on the available prenatal data. The small size of the database has been a challenge to find new characterisations of the disease, although some correlation has been detected between VMG and two other parameters: the Left Anterior Horn and the Right Anterior Horn. A binary classification model has been trained with a value of F1 of about 90%. This has been achieved applying neural networks or by cascading neural networks with the most powerful supervised learning models. The conclusion is that a small number of even unbalanced data allows to provide models with sufficient performance but more knowledge on the context might improve the results.
Keywords: ventriculomegaly
machine learning
data mining
detection
fetal medicine
Document type: info:eu-repo/semantics/masterThesis
Issue Date: Jun-2021
Publication license: http://creativecommons.org/licenses/by-nc-nd/3.0/es/  
Appears in Collections:Trabajos finales de carrera, trabajos de investigación, etc.

Files in This Item:
File Description SizeFormat 
mboschgaTFM0621memòria.pdfMemòria del TFM12,25 MBAdobe PDFThumbnail
View/Open