🧠 TopAneu 2026 Challenge¶

Background¶

The TopAneu 2026 challenge focuses on automated analysis of intracranial aneurysms from CT angiography (CTA) and MR angiography (MRA). Intracranial aneurysms are relatively common and can lead to severe outcomes when ruptured, making early detection and accurate characterization clinically important.

While recent work has explored aneurysm detection and segmentation, most existing datasets and benchmarks either focus on binary segmentation or provide limited anatomical localization. In clinical practice, however, aneurysm location, size, and morphology are all relevant for risk assessment and treatment planning.

TopAneu 2026 addresses this gap by introducing a multi-center, multi-modality dataset with fine-grained vessel-specific labels, enabling more anatomically meaningful evaluation of algorithms.

Key Highlights¶

• Multi-modality: CTA and MRA data from routine clinical practice
• Multi-center dataset: ~850 scans from several international institutions
• Fine-grained anatomy: 50+ vessel-specific aneurysm location classes
• Two complementary tasks:
  • Vessel-specific classification
  • Vessel-specific segmentation
• Clinical relevance: Evaluation considers anatomical location and aneurysm size

🎯 The Tasks¶

Participants will develop algorithms for two related tasks:

Task 1 — Vessel-Specific Classification¶

Given a CTA or MRA scan, predict the presence or absence of aneurysms across more than 50 predefined vessel locations.

• Multi-label classification problem
• Focus on detection and anatomical localization
• Evaluated per vessel location

Task 2 — Vessel-Specific Segmentation¶

Given a scan, detect and segment aneurysms at the voxel level, assigning each aneurysm to its vessel location class.

• Multi-class segmentation problem
• Includes size-based analysis
• Evaluates both detection and segmentation quality

Together, the two tasks reflect a clinically meaningful workflow: from screening and localization to detailed morphological assessment.

📥 Inputs¶

• 3D CTA or MRA scans
• Optional auxiliary data (training only):
  • Vessel segmentation masks (gold- and silver-standard)

📤 Outputs¶

• Task 1: Multi-label aneurysm classification based on vessel location labels
• Task 2: Voxel-level segmentation masks with vessel-specific labels

🗓️ Schedule (Preliminary)¶

🧾 Participation¶

• Submission via Docker containers on Grand Challenge
• A sanity-check phase is provided for testing submissions
• One final submission per team (test phase)
• Teams may use additional data, with proper disclosure

📦 Data Usage¶

• Data are fully anonymized and collected under ethical approval
• Open for research use with attribution
• Commercial use requires permission from data owners

📝 Registration¶

[!IMPORTANT] Verified accounts are mandatory to be accepted to the challenge. Please ensure your Grand-Challenge profile is complete and verified before requesting to join.

📌 Summary¶

TopAneu 2026 provides a benchmark for anatomy-aware aneurysm analysis, combining detection, localization, and segmentation across modalities and centers. The challenge aims to support the development of methods that are not only technically strong but also aligned with clinical decision-making needs.