Challenge Banner

Data

The form for the dataset application can be found here form

⚠️ HECKTOR 2026 – Data Access & Compliance Policy

The HECKTOR 2026 Challenge dataset includes data related to the United States. As a result, all participants must comply with applicable U.S. regulations governing access to sensitive personal data.

📜 Regulatory Compliance Requirements

Participation in HECKTOR 2026 requires compliance with:

Additional details:
👉 https://miccai.org/index.php/events/upcoming-conferences/miccai-2026-challenge-requirements/

✅ Participant Agreement

By registering for HECKTOR 2026, participants:

  • Confirm that they meet these updated requirements, and
  • Agree to comply with all applicable data use regulations listed above

Failure to meet these requirements will result in denial or revocation of dataset access.

Please note that, to ensure fairness, no data will be shared before April 15 for challenge participants. If you already have the HECKTOR 2025 dataset, please apply again for the HECKTOR 2026 version as it includes more data for the challenge.

Please refer to the Dataset page for a detailed description of the data structure.

Center Mask T stage N Stage RFS Available for All
CHUM 5656565656
CHUP 7272724444
CHUS 7272727272
MDA 398395396397392
MDA New 4646462727
HGJ 5555555555
HMR 1817181817
USZ 1111111111
CHUV 5353534444
TOTAL 781777779724691

Licensing and Acknowledgement

If you use this dataset and/or results derived from participating in this challenge, please cite the following paper in any resulting publication or public disclosure

Saeed et al., “A Multimodal and Multi-centric Head and Neck Cancer Dataset for Segmentation, Diagnosis and Outcome Prediction,” arXiv:2509.00367 (2025)

This dataset is licensed under the CC BY-NC-SA license.

✅ Participants can:

  • Use your dataset to develop and evaluate models for the challenge.
  • Publish papers based on their results (with proper attribution).
  • Share code or results using the dataset under the same license.

❌ Participants cannot:

  • Use the data to train models for commercial medical software or services.
  • Sell or integrate the data into proprietary datasets.
  • Redistribute a modified dataset under a more restrictive or commercial license.