AutoPET Challenge: Combining nn-Unet with Swin UNETR Augmented by Maximum Intensity Projection Classifier
This work addresses the problem of reliable tumor segmentation for cancer staging and re-staging in medical imaging, but it is incremental as it combines existing state-of-the-art models with a gating mechanism.
The paper tackles the challenging task of tumor segmentation in FDG-PET/CT scans, where high metabolism regions are not specific to tumors, by proposing an ensemble of nn-Unet and Swin UNETR models augmented with a maximum intensity projection classifier, achieving a Dice score of 72.12% on a cross-validation set for lung cancer, melanoma, and lymphoma patients.
Tumor volume and changes in tumor characteristics over time are important biomarkers for cancer therapy. In this context, FDG-PET/CT scans are routinely used for staging and re-staging of cancer, as the radiolabeled fluorodeoxyglucose is taken up in regions of high metabolism. Unfortunately, these regions with high metabolism are not specific to tumors and can also represent physiological uptake by normal functioning organs, inflammation, or infection, making detailed and reliable tumor segmentation in these scans a demanding task. This gap in research is addressed by the AutoPET challenge, which provides a public data set with FDG-PET/CT scans from 900 patients to encourage further improvement in this field. Our contribution to this challenge is an ensemble of two state-of-the-art segmentation models, the nn-Unet and the Swin UNETR, augmented by a maximum intensity projection classifier that acts like a gating mechanism. If it predicts the existence of lesions, both segmentations are combined by a late fusion approach. Our solution achieves a Dice score of 72.12\% on patients diagnosed with lung cancer, melanoma, and lymphoma in our cross-validation. Code: https://github.com/heiligerl/autopet_submission