Past autoPET Challenges¶

The autoPET challenge series has been pushing the boundaries of fully automated lesion segmentation in whole-body PET/CT since 2022. With autoPETV now underway, we look back at four editions that shaped the field — each raising the bar from single-tracer automation to multitracer generalization and interactive, human-in-the-loop segmentation.

📄 New! We proudly present two new preprints for the autoPET III and autoPET IV challenges (see citations 3 & 4 below).

📚 A complete summary of all four challenges is available here: autoPET I–IV Summary (PDF)

How to Cite¶

If the autoPET challenges have supported your work, please cite the relevant publications:

1. autoPET I Gatidis S, Früh M, Fabritius MP, et al. Results from the autoPET challenge on fully automated lesion segmentation in oncologic PET/CT imaging. Nat Mach Intell. Published online October 30, 2024:1-10. doi:10.1038/s42256-024-00912-9

2. autoPET II Dexl J, Gatidis S, Früh M, et al. AutoPET Challenge on Fully Automated Lesion Segmentation in Oncologic PET/CT Imaging, Part 2: Domain Generalization. Journal of Nuclear Medicine. Published online December 30, 2025. doi:10.2967/jnumed.125.270260

3. autoPET III (new preprint) Dexl J, Jeblick K, Mittermeier A, et al. The autoPET3 Challenge: Automated Lesion Segmentation in Whole-Body PET/CT – Multitracer Multicenter Generalization. arXiv. Preprint posted online May 11, 2026:arXiv:2605.05775. doi:10.48550/arXiv.2605.05775

4. autoPET IV (new preprint) Megne Choudja PO, Dexl J, Marinov Z, et al. From Automation to Collaboration: The autoPET/CT-IV Challenge on Interactive Lesion Segmentation in Whole-Body PET/CT and Longitudinal CT. Social Science Research Network. Preprint posted online May 29, 2026:6841479. doi:10.2139/ssrn.6841479