Advanced Tumor Segmentation in PET/CT Imaging: A Training Strategy Study with nnU-Net for AutoPET III
For medical imaging researchers, this work provides an incremental improvement in tumor segmentation by systematically evaluating training strategies, though the method is based on existing nnU-Net framework.
The authors developed a whole-body tumor segmentation method for PET/CT imaging using nnU-Net with a ResNet encoder, achieving a Dice score of 0.80 and third place in the AutoPET III challenge by optimizing training strategies like intensity normalization, batch dice optimization, and CraveMix augmentation.
Tumor segmentation in whole-body PET/CT imaging is crucial for precise disease evaluation and treatment planning. However, it remains challenging due to variability in lesion size, contrast, and anatomical distribution. Relying on manual segmentation makes the process time-consuming and prone to intra- and inter-observer variability. This work presents a whole-body tumor segmentation method developed for the AutoPET III challenge, where the goal is to build models that generalize across tracers and multi-center data. We employ the nnU-Net framework with a ResNet-based encoder as our baseline and systematically investigate the impact of training strategies, including intensity normalization, batch dice optimization, and data augmentation using CraveMix. Our experiments show that these strategies significantly influence model performance, particularly in reducing false positives and improving robustness to lesion variability. The best-performing configuration achieves a Dice score of up to 0.80 on the preliminary test phase, and our method ranked third in the AutoPET III challenge. The code is publicly available here.