Deep learning for cerebral angiography segmentation from non-contrast computed tomography

Abstract

Cerebral computed tomography angiography is a widely available imaging technique that helps in the diagnosis of vascular pathologies. Contrast administration is needed to accurately assess the arteries. On non-contrast computed tomography, arteries are hardly distinguishable from the brain tissue, therefore, radiologists do not consider this imaging modality appropriate for the evaluation of vascular pathologies. There are known contraindications to administering iodinated contrast media, and in these cases, the patient has to undergo another examination to visualize cerebral arteries, such as magnetic resonance angiography. Deep learning for image segmentation has proven to perform well on medical data for a variety of tasks. The aim of this research was to apply deep learning methods to segment cerebral arteries on non-contrast computed tomography scans and consequently, generate angiographies without the need for contrast administration. The dataset for this research included 131 patients who underwent brain non-contrast computed tomography directly followed by computed tomography with contrast administration. Then, the segmentations of arteries were generated and aligned with non-contrast computed tomography scans. A deep learning model based on the U-net architecture was trained to perform the segmentation of blood vessels on non-contrast computed tomography. An evaluation was performed on separate test data, as well as using cross-validation, reaching Dice coefficients of 0.638 and 0.673, respectively. This study proves that deep learning methods can be leveraged to quickly solve problems that are difficult and time-consuming for a human observer, therefore providing physicians with additional information on the patient. To encourage the further development of similar tools, all code used for this research is publicly available.

Deep Learning for 3D Vascular Segmentation in Phase Contrast Tomography

“…As suggested by its name, TOF MRA relies on a principle known as flow-related enhancement, which happens when completely magnetised blood flows into a slab of magnetically saturated tissue whose signal has been muted by repeated RF-pulses 27 . However, the long acquisition time and high operational costs make it difficult to employ for cerebral artery visualisation 28,29 . Furthermore, MRA has been shown to overestimate stenosis compared to other modalities 30 .…”

Charting the potential of brain computed tomography deep learning systems

“…Deep learning systems can facilitate patient triage and improve clinical efficiency. O'Neill et al developed a model to detect ICH on NCCTB and showed that using this system as a triage tool reduced image interpretation turnaround times [52] [44]. Poirot et al applied deep learning to improve dual-energy CT (DECT) scan processing.…”

Automated anatomical labeling of the intracranial arteries via deep learning in computed tomography angiography

“…Deep learning (DL) has shown significant potential in medical image analysis tasks. Notably, the U-Net framework with symmetric network architecture is widely adopted in the field of medical image segmentation because of its flexibility and achieves remarkable successes ( Jin et al, 2020 ; Klimont et al, 2020 ; Mubashar et al, 2022 ). Previous models have focused on the segmentation of the specific artery (e.g., the carotid artery) or the whole 3D cerebral vessels ( Groves et al, 2020 ; Klimont et al, 2020 ; Bortsova et al, 2021 ; Guo et al, 2021 ).…”

“…Notably, the U-Net framework with symmetric network architecture is widely adopted in the field of medical image segmentation because of its flexibility and achieves remarkable successes ( Jin et al, 2020 ; Klimont et al, 2020 ; Mubashar et al, 2022 ). Previous models have focused on the segmentation of the specific artery (e.g., the carotid artery) or the whole 3D cerebral vessels ( Groves et al, 2020 ; Klimont et al, 2020 ; Bortsova et al, 2021 ; Guo et al, 2021 ). Only few studies strive for the automatic segmentation or labeling of intracranial arteries involved with detailed segments (i.e., automatic labelling of fine segments).…”