Related Experiment Video
Updated: Jul 5, 2025
Author Spotlight: Noninvasive Cerebral Blood Flow Determination in Human Functional Brain Region for Diagnosis of Neurological Disorders
Published on: May 31, 2024
Automated anatomical labeling of the intracranial arteries via deep learning in computed tomography angiography
Ting Chen1,2, Wei You1,3, Liyuan Zhang4
1Department of Interventional Neuroradiology, Beijing Tiantan Hospital and Beijing Neurosurgical Institute, Capital Medical University, Beijing, China.
Insights
A novel deep learning model automatically labels intracranial arteries on CT angiography, improving aneurysm detection accuracy and reducing diagnosis time for clinicians.
Area of Science:
- Medical Imaging Analysis
- Artificial Intelligence in Radiology
- Neurovascular Anatomy
Background:
- Accurate anatomical labeling of cerebral vasculature is vital for understanding morphological variations.
- Manual labeling of intracranial arteries is challenging due to anatomical variability and large imaging datasets.
- Existing methods are time-consuming and prone to errors, necessitating automated solutions.
Purpose of the Study:
- To develop and validate a U-Net-based deep learning model for automated anatomical labeling of intracranial arteries in computed tomography angiography (CTA).
- To assess the model's performance in segmenting 42 distinct intracranial artery segments, including the internal carotid artery (ICA).
- To evaluate the clinical utility of the deep learning model in enhancing the precision and efficiency of intracranial aneurysm (IA) localization.
Main Methods:
- A U-Net deep learning architecture was trained and validated on 457 CTA examinations.
- The model was evaluated using Dice Similarity Coefficient (DSC), Mean Surface Distance (MSD), and Hausdorff Distance (HD).
- A clinical experiment involved 5 clinicians assessing IA localization with and without algorithmic assistance on 96 IA-containing CTA scans.
Main Results:
- The deep learning model achieved a mean DSC of 0.88, MSD of 0.82 mm, and HD of 6.59 mm for 42 intracranial artery segments.
- Superior performance was observed in healthy subjects compared to patients with stenosis (DSC: 0.91 vs. 0.89; HD: 4.75 vs. 6.19 mm).
- Clinician performance improved significantly with model assistance: accuracy increased by 0.04, diagnosis time decreased by 9.76 s, and interrater agreement (Fleiss' Kappa) rose by 0.07.
Conclusions:
- The developed deep learning model is proficient for labeling intracranial arteries using a large CTA dataset.
- The model demonstrates significant clinical utility by improving IA localization accuracy and efficiency.
- This automated approach can aid in prioritizing high-risk patients and reducing clinical workload.
Abstract:
Background and purpose: Anatomical labeling of the cerebral vasculature is a crucial topic in determining the morphological nature and characterizing the vital variations of vessels, yet precise labeling of the intracranial arteries is time-consuming and challenging, given anatomical structural variability and surging imaging data. We present a U-Net-based deep learning (DL) model to automatically label detailed anatomical segments in computed tomography angiography (CTA) for the first time. The trained DL algorithm was further tested on a clinically relevant set for the localization of intracranial aneurysms (IAs). Methods: 457 examinations with varying degrees of arterial stenosis were used to train, validate, and test the model, aiming to automatically label 42 segments of the intracranial arteries [e.g., 7 segments of the internal carotid artery (ICA)]. Evaluation metrics included Dice similarity coefficient (DSC), mean surface distance (MSD), and Hausdorff distance (HD). Additionally, 96 examinations containing at least one IA were enrolled to assess the model's potential in enhancing clinicians' precision in IA localization. A total of 5 clinicians with different experience levels participated as readers in the clinical experiment and identified the precise location of IA without and with algorithm assistance, where there was a washout period of 14 days between two interpretations. The diagnostic accuracy, time, and mean interrater agreement (Fleiss' Kappa) were calculated to assess the differences in clinical performance of clinicians. Results: The proposed model exhibited notable labeling performance on 42 segments that included 7 anatomical segments of ICA, with the mean DSC of 0.88, MSD of 0.82 mm and HD of 6.59 mm. Furthermore, the model demonstrated superior labeling performance in healthy subjects compared to patients with stenosis (DSC: 0.91 vs. 0.89, p < 0.05; HD: 4.75 vs. 6.19, p < 0.05). Concurrently, clinicians with model predictions achieved significant improvements when interpreting the precise location of IA. The clinicians' mean accuracy increased by 0.04 (p = 0.003), mean time to diagnosis reduced by 9.76 s (p < 0.001), and mean interrater agreement (Fleiss' Kappa) increased by 0.07 (p = 0.029). Conclusion: Our model stands proficient for labeling intracranial arteries using the largest CTA dataset. Crucially, it demonstrates clinical utility, helping prioritize the patients with high risks and ease clinical workload.
