RebrAIn A New Clinical Targeting of VIM and STN Based on Supervised Statistical Learning: Study on Localization of the Prediction - Inria - Institut national de recherche en sciences et technologies du numérique Accéder directement au contenu
Communication Dans Un Congrès Année : 2023

RebrAIn A New Clinical Targeting of VIM and STN Based on Supervised Statistical Learning: Study on Localization of the Prediction

Résumé

Introduction: Targeting for DBS remains controversial. RebrAIn developed a clinical targeting based on localization of active contact in patients with DBS who had very good outcome. Aims & Objectives: To understand where the RebrAIn prediction is located within anatomical structures. Materials and Methods: In 88 hemispheres for the STN and 62 for the VIM we applied our artificial intelligence process to predict the STN or the VIM clinical targets. We segmented each marked MRI by both GuideXT and Suretune 4 software to which we compared the localization of the RebrAIn prediction in the segmented structures. Results: For the STN: 66 (75%) RebrAIn targets were in the STN in both GuideXT and Suretune. None (0%) were outside the STN for both software. In 17 (19%) hemispheres, the predictions were outside the STN for GuideXT only and 5 (6%) outside for Suretune only. For the VIM: with GuideXT, the predictions were in the inferior part of the VIM in 23 (37%) hemispheres, below in the PSA in 37 (60%) cases and in the STN in 2 (3%) cases. With Suretune the results were 21 (34%); 37 (60%), and 4 (6%) respectively. Among the 6 hemispheres where predictions were in the STN; 3 were located in the STN by both software. Conclusions: Targeting predictions based on clinical experience allow us to localize STN and VIM region of interest. The use of a single segmentation software to control the postoperative position of the electrodes exposes the risk of misinterpreting the location of the targets and electrodes.
Fichier non déposé

Dates et versions

hal-04155171 , version 1 (07-07-2023)

Licence

Paternité

Identifiants

Citer

Nejib Zemzemi, Julien Engelhardt, Emmanuel Cuny, Dominique Guehl, Pierre Burbaud. RebrAIn A New Clinical Targeting of VIM and STN Based on Supervised Statistical Learning: Study on Localization of the Prediction. INS 2022 Interim Meeting, Nov 2022, Mumbai (India), France. pp.S35, ⟨10.1016/j.neurom.2023.02.056⟩. ⟨hal-04155171⟩
28 Consultations
0 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More