XA-170K: 170,000 X-ray Angiograms for Vascular Anatomy Pre-training

XA-170K contains 170,000 X-ray angiogram images aggregated from CADICA, SYNTAX, XCAD, and other public sources. Created by researchers from the Chinese Academy of Sciences and Alibaba Group in 2026, it is designed for vascular anatomy-aware self-supervised pre-training.

Use Cases

• Self-supervised pre-training for vascular segmentation using X-ray angiogram frames
• Anatomical feature extraction for coronary artery disease analysis
• Cross-dataset benchmarking using aggregated CADICA and SYNTAX subsets

Strengths

• 170,000 image records
• Aggregates four public medical sources (CADICA, SYNTAX, XCAD)
• Designed specifically for vascular anatomy-aware pre-training

Limitations

• Potential domain shift between the four different source datasets
• Lack of detailed column-level metadata in the current documentation

Provenance

Source

Institute of Automation, Chinese Academy of Sciences and Alibaba Group

Collection Method

Aggregated from public sources including CADICA, SYNTAX, and XCAD

Freshness

Last updated February 2026.

Geography

United States