The European Society for Hybrid, Molecular, and Translational Imaging (ESHIMT) is launching what it calls the first autoPET challenge of automated tumor lesion segmentation in whole-body FDG-PET/CT.
The challenge, which will award 15,000 euros in total, is being held in collaboration with the Medical Image Computing and Computer Assisted Intervention (MICCAI).
The goal is to foster and promote research on machine learning-based automation and data evaluation. The seven highest-ranking submissions will compete for the prize money and will be invited to present their methods at MICCAI 2021. They will also be invited to write a peer-reviewed journal paper that performs a full analysis of the results and highlights the key findings and methods.
AutoPET gives a publicly available dataset of 1,014 studies from 900 patients on a single site. The task is to develop an algorithm to accurately assess lesion segmentation of whole-body FDG-PET/CT while avoiding false positives.
Evaluation will be performed on held-out test cases of 200 patients, which will be split into two subgroups. One will be drawn from the same hospital as the training cases (University Hospital Tübingen, Germany) while the other will be drawn from a different hospital (University Hospital of LMU in Munich, Germany) with similar acquisition protocols. Both subgroups will have 100 patients each.
Submissions open on 25 April. Further details are available on the ESHIMT website.