DI-UMONS : Dépôt institutionnel de l’université de Mons

Recherche transversale
Rechercher
(titres de publication, de périodique et noms de colloque inclus)
2019-11-22 - Colloque/Présentation - communication orale - Anglais - page(s)

Amkrane Yassine , El Adoui Mohammed , Benjelloun Mohammed , "Analysing Breast cancer reaction to chemotherapy using Radiomics" in Grascomp Doctoral Day (GDD’19), UNamur, Belgique, 2019

  • Codes CREF : Imagerie médicale, radiologie, tomographie (DI3243)
  • Unités de recherche UMONS : Informatique (F114)
  • Instituts UMONS : Institut de Recherche en Technologies de l’Information et Sciences de l’Informatique (InforTech), Institut NUMEDIART pour les Technologies des Arts Numériques (Numédiart)
Texte intégral :

Abstract(s) :

(Anglais) Breast cancer presents one of the most important diseases of women in the world. Indeed, Breast cancer is the second cause of death in the world with 8,8 million of deaths only in 2015. So, more than 70% are caused by breast cancer, occur in low-income countries since they are not equipped provided by efficient solutions for breast cancer diagnosis and prediction. The main challenge is diagnosis breast cancer as soon as it appears in other order to have the possibility ability to provide convenient and efficient treatments. In this context, several image modalities are used for breast tumor diagnosis such as echography, mammography and, PET (positron emission tomography) scans and MRI (Magnetic Resonance Images). One of the main treatments of this kind of pathologies is neoadjuvant chemotherapy, which attacks cancer cells and reduce the breast tumor’s size to facilitate the surgery. However, chemotherapy is hampered by several secondary effects (hair loss, osteoporosis, vomiting, etc.) and the cancer may not respond to it after several years of treatment. In this paper, we study propose a new method for breast cancer response prediction using bases on three steps: 1. Breast cancer segmentation from MR images. 2. Features extraction from segmented tumors in order to generate a complete and exploitable database. 3. Exploitation of deep learning architectures inf order to compute models allowing to the prediction of the tumor response. Experimental results will be conducted using a dataset of 42 breast cancer patients having local breast cancer, provided by our collaborator, Jules Bordet Institute - (Brussels) – Belgium. Key