Back
Results from the autoPET challenge on fully automated lesion\nsegmentation in oncologic PET/CT imaging
Automated detection of tumour lesions on positron emission tomography–computed tomography (PET/CT) image data is a clinically relevant but highly challenging task. Progress in this field has been hampered in the past owing to the lack of publicly available annotated data and limited availability of platforms for inter-institutional collaboration. Here we describe the results of the autoPET challenge, a biomedical image analysis challenge aimed to motivate research in the field of automated PET/CT image analysis. The challenge task was the automated segmentation of metabolically active tumour lesions on whole-body 18F-fluorodeoxyglucose PET/CT. Challenge participants had access to a large publicly available annotated PET/CT dataset for algorithm training. All algorithms submitted to the final challenge phase were based on deep learning methods, mostly using three-dimensional U-Net architectures. Submitted algorithms were evaluated on a private test set composed of 150 PET/CT studies from two institutions. An ensemble model of the highest-ranking algorithms achieved favourable performance compared with individual algorithms. Algorithm performance was dependent on the quality and quantity of data and on algorithm design choices, such as tailored post-processing of predicted segmentations. Future iterations of this challenge will focus on generalization and clinical translation.
@article{GatidisNatureMI24, title = {Results from the autoPET challenge on fully automated lesion\nsegmentation in oncologic PET/CT imaging}, journal = {Nature Machine Intelligence}, abstract = {Automated detection of tumour lesions on positron emission tomography–computed tomography (PET/CT) image data is a clinically relevant but highly challenging task. Progress in this field has been hampered in the past owing to the lack of publicly available annotated data and limited availability of platforms for inter-institutional collaboration. Here we describe the results of the autoPET challenge, a biomedical image analysis challenge aimed to motivate research in the field of automated PET/CT image analysis. The challenge task was the automated segmentation of metabolically active tumour lesions on whole-body 18F-fluorodeoxyglucose PET/CT. Challenge participants had access to a large publicly available annotated PET/CT dataset for algorithm training. All algorithms submitted to the final challenge phase were based on deep learning methods, mostly using three-dimensional U-Net architectures. Submitted algorithms were evaluated on a private test set composed of 150 PET/CT studies from two institutions. An ensemble model of the highest-ranking algorithms achieved favourable performance compared with individual algorithms. Algorithm performance was dependent on the quality and quantity of data and on algorithm design choices, such as tailored post-processing of predicted segmentations. Future iterations of this challenge will focus on generalization and clinical translation.}, volume = {6}, number = {11}, pages = {1396–1405}, year = {2024}, slug = {gatidisnaturemi24}, author = {Gatidis, Sergios and Frueh, Marcel and Fabritius, Matthias P. and Gu, Sijing and Nikolaou, Konstantin and La Fougere, Christian and Ye, Jin and He, Junjun and Peng, Yige and Bi, Lei and Ma, Jun and Wang, Bo and Zhang, Jia and Huang, Yukun and Heiliger, Lars and Marinov, Zdravko and Stiefelhagen, Rainer and Egger, Jan and Kleesiek, Jens and Sibille, Ludovic and Xiang, Lei and Bendazzoli, Simone and Astaraki, Mehdi and Ingrisch, Michael and Cyran, Clemens C. and Kuestner, Thomas}, url = {https://www.nature.com/articles/s42256-024-00912-9} }