The autoPET3 Challenge: Automated Lesion Segmentation in Whole-Body PET/CT $\unicode{x2013}$ Multitracer Multicenter Generalization

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Dexl, Jakob, Jeblick, Katharina, Mittermeier, Andreas, Schachtner, Balthasar, Stüber, Anna Theresa, Topalis, Johanna, Rokuss, Maximilian, Isensee, Fabian, Maier-Hein, Klaus H., Kalisch, Hamza, Kleesiek, Jens, Seibold, Constantin M., Alasmawi, Hussain, Chan, Lap Yan Lennon, Yuan, Yixuan, Jaus, Alexander, Stiefelhagen, Rainer, Choudja, Pauline Ornela Megne, Nikolaou, Konstantin, La Fougère, Christian, Gatidis, Sergios, Fabritius, Matthias P., Heimer, Maurice, Abaci, Gizem, Sundar, Lalith Kumar Shiyam, Werner, Rudolf A., Ricke, Jens, Cyran, Clemens C., Küstner, Thomas, Ingrisch, Michael
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913109369683968
author Dexl, Jakob
Jeblick, Katharina
Mittermeier, Andreas
Schachtner, Balthasar
Stüber, Anna Theresa
Topalis, Johanna
Rokuss, Maximilian
Isensee, Fabian
Maier-Hein, Klaus H.
Kalisch, Hamza
Kleesiek, Jens
Seibold, Constantin M.
Alasmawi, Hussain
Chan, Lap Yan Lennon
Yuan, Yixuan
Jaus, Alexander
Stiefelhagen, Rainer
Choudja, Pauline Ornela Megne
Nikolaou, Konstantin
La Fougère, Christian
Gatidis, Sergios
Fabritius, Matthias P.
Heimer, Maurice
Abaci, Gizem
Sundar, Lalith Kumar Shiyam
Werner, Rudolf A.
Ricke, Jens
Cyran, Clemens C.
Küstner, Thomas
Ingrisch, Michael
author_facet Dexl, Jakob
Jeblick, Katharina
Mittermeier, Andreas
Schachtner, Balthasar
Stüber, Anna Theresa
Topalis, Johanna
Rokuss, Maximilian
Isensee, Fabian
Maier-Hein, Klaus H.
Kalisch, Hamza
Kleesiek, Jens
Seibold, Constantin M.
Alasmawi, Hussain
Chan, Lap Yan Lennon
Yuan, Yixuan
Jaus, Alexander
Stiefelhagen, Rainer
Choudja, Pauline Ornela Megne
Nikolaou, Konstantin
La Fougère, Christian
Gatidis, Sergios
Fabritius, Matthias P.
Heimer, Maurice
Abaci, Gizem
Sundar, Lalith Kumar Shiyam
Werner, Rudolf A.
Ricke, Jens
Cyran, Clemens C.
Küstner, Thomas
Ingrisch, Michael
contents We report the design and results of the third autoPET challenge (MICCAI 2024), which benchmarked automated lesion segmentation in whole-body PET/CT under a compositional generalization setting. Training data comprised 1,014 [18F]-FDG PET/CT studies from the University Hospital Tübingen and 597 [18F]/[68Ga]-PSMA PET/CT studies from the LMU University Hospital Munich, constituting the largest publicly available annotated PSMA PET/CT dataset to date. The held-out test set of 200 studies covered four tracer-center combinations, two of which represented unseen compositional pairings. A complementary data-centric award category isolated the contribution of data handling strategies by restricting participants to a fixed baseline model. Seventeen teams submitted 27 algorithms, predominantly nnU-Net-based 3D networks with PET/CT channel concatenation. The top-ranked algorithm achieved a mean DSC of 0.66, FNV of 3.18 mL, and FPV of 2.78 mL across all four test conditions, improving DSC by 8% and reducing the false-negative volume by 5 mL relative to the provided baseline. Ranking was stable across bootstrap resampling and alternative ranking schemes for the top tier. Beyond the benchmark, we provide an in-depth analysis of segmentation performance at the patient and lesion level. Three main conclusions can be drawn: (1) in-domain multitracer PET/CT segmentation is sufficient and probably approaching reader agreement; (2) compositional generalization to unseen tracer-center combinations remains an open problem mainly driven by systematic volume overestimation; (3) heterogeneity and case difficulty drive performance variation substantially more than the choice of algorithm among top-ranked teams.
format Preprint
id arxiv_https___arxiv_org_abs_2605_05775
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle The autoPET3 Challenge: Automated Lesion Segmentation in Whole-Body PET/CT $\unicode{x2013}$ Multitracer Multicenter Generalization
Dexl, Jakob
Jeblick, Katharina
Mittermeier, Andreas
Schachtner, Balthasar
Stüber, Anna Theresa
Topalis, Johanna
Rokuss, Maximilian
Isensee, Fabian
Maier-Hein, Klaus H.
Kalisch, Hamza
Kleesiek, Jens
Seibold, Constantin M.
Alasmawi, Hussain
Chan, Lap Yan Lennon
Yuan, Yixuan
Jaus, Alexander
Stiefelhagen, Rainer
Choudja, Pauline Ornela Megne
Nikolaou, Konstantin
La Fougère, Christian
Gatidis, Sergios
Fabritius, Matthias P.
Heimer, Maurice
Abaci, Gizem
Sundar, Lalith Kumar Shiyam
Werner, Rudolf A.
Ricke, Jens
Cyran, Clemens C.
Küstner, Thomas
Ingrisch, Michael
Computer Vision and Pattern Recognition
Artificial Intelligence
We report the design and results of the third autoPET challenge (MICCAI 2024), which benchmarked automated lesion segmentation in whole-body PET/CT under a compositional generalization setting. Training data comprised 1,014 [18F]-FDG PET/CT studies from the University Hospital Tübingen and 597 [18F]/[68Ga]-PSMA PET/CT studies from the LMU University Hospital Munich, constituting the largest publicly available annotated PSMA PET/CT dataset to date. The held-out test set of 200 studies covered four tracer-center combinations, two of which represented unseen compositional pairings. A complementary data-centric award category isolated the contribution of data handling strategies by restricting participants to a fixed baseline model. Seventeen teams submitted 27 algorithms, predominantly nnU-Net-based 3D networks with PET/CT channel concatenation. The top-ranked algorithm achieved a mean DSC of 0.66, FNV of 3.18 mL, and FPV of 2.78 mL across all four test conditions, improving DSC by 8% and reducing the false-negative volume by 5 mL relative to the provided baseline. Ranking was stable across bootstrap resampling and alternative ranking schemes for the top tier. Beyond the benchmark, we provide an in-depth analysis of segmentation performance at the patient and lesion level. Three main conclusions can be drawn: (1) in-domain multitracer PET/CT segmentation is sufficient and probably approaching reader agreement; (2) compositional generalization to unseen tracer-center combinations remains an open problem mainly driven by systematic volume overestimation; (3) heterogeneity and case difficulty drive performance variation substantially more than the choice of algorithm among top-ranked teams.
title The autoPET3 Challenge: Automated Lesion Segmentation in Whole-Body PET/CT $\unicode{x2013}$ Multitracer Multicenter Generalization
topic Computer Vision and Pattern Recognition
Artificial Intelligence
url https://arxiv.org/abs/2605.05775