Brain-vascular segmentation for SEEG planning via a 3D fully-convolutional neural network

Abstract

Three dimensional visualization of vascular structures can assist clinicians in preoperative planning, intraoperative guidance, and post-operative decision-making. The goal of this work is to provide an automatic, accurate and fast method for brain vessels segmentation in Contrast Enhanced Cone Beam Computed Tomography (CE-CBCT) dataset based on a residual Fully Convolutional Neural Network (FCNN). The proposed NN embeds in an encoder-decoder architecture residual elements which decreases the vanishing effect due to deep architecture while accelerating the convergence. Moreover, a two-stage training has been proposed as a countermeasure for the unbalanced nature of the dataset. The FCNN training was performed on 20 CE-CBCT volumes exploiting minibatch gradient descent andthe Adam optimizer. Binary crossentropy was used as loss function. Performance evaluation was conducted considering 5 datasets. A median value of Dice, Precision and Recall of 0.79, 0.8 and 0.69 were obtained with respect to manual annotations.

A Fully Automatic Method for Vascular Tortuosity Feature Extraction in the Supra-Aortic Region: Unraveling Possibilities in Stroke Treatment Planning

“…Table 5 shows a comprehensive evaluation of the acquisition methods performance for each tortuosity feature. 2020) (Fu et al, 2020) ResU-Net CTA Head + neck + AA 18,259 0.95 Fantazzini et al (2020) (Fantazzini et al, 2020) 2D U-Nets CTA Aorta 80 0.92 ± 0.01 Fan et al (2020) (Fan et al, 2020) HMRF + U-Net TOF MRA Cerebral arteries 100 0.79 ± 0.05 ElHadji et al ( 2019) (Hadji et al, 2019) ResU-Net CE-CBCT Cerebral arteries 25 0.79 ± 0.13 Livne et al (2019) (Livne et al, 2019) Half U-Net TOF MRA Cerebral arteries 66 0.92 Phellan et al (2017) (Phellan et al, 2017) Deep CNN TOF MRA Cerebral arteries 4 0.77 ± 0.01 Isensee et al (2021) (Isensee et al, 2021) nnU-Net CTA Hepatic vessels 443 0.69…”