Vessel diameters of 14 basal cerebral arteries assessed in 1000 digital subtraction angiographies

Till Gumbel1, Cindy Richter2, Christian Martin3

  • 1Department of Neurosurgery, University Hospital Leipzig, Liebigstrasse 20, 04103, Leipzig, Germany.

Scientific Data
|September 8, 2025
PubMed

Insights

Establishing normative values for intracranial vessel size is challenging due to variations in patient factors and disease effects. This study analyzed over 1000 cerebral angiographies to create a dataset for approximating normal vessel dimensions.

Area of Science:

  • Neurology
  • Radiology
  • Medical Imaging

Background:

  • Normative values for intracranial vessel size are difficult to establish.
  • Factors like gender, height, weight, and cerebrovascular diseases influence vessel diameters.
  • Cerebral angiography is typically reserved for severe conditions, complicating the definition of physiological values.

Purpose of the Study:

  • To approximate "normal" values for intracranial vessel size.
  • To create a comprehensive dataset of cerebral angiographic measurements.
  • To enable the computation of intraindividual indices and train machine learning models.

Main Methods:

  • Analysis of over 1000 contemporary cerebral angiographies from a single neurovascular center.
  • Recording of diameters for 14 basal cerebral arteries, patient age, gender, and underlying disease.
  • Utilized SPSS 29 (IBM) for data management of 1010 digital subtraction angiographies.

Main Results:

  • A significant difference (p < 0.001) was found in the size of the left carotid artery between male and female patients.
  • Male patients had a larger average left carotid artery size (3.23 mm) compared to female patients (3.09 mm).
  • The dataset provides detailed measurements for statistical analysis and potential correlations.

Conclusions:

  • The developed dataset offers a valuable resource for approximating normative intracranial vessel sizes.
  • It can be used to compute intraindividual indices for specific diseases, aiding in the assessment of conditions like cerebral aneurysms.
  • The dataset is suitable for training machine learning algorithms to predict neurological events such as ischemic stroke or cerebral hemorrhage.