标题： 针对PET-CT肿瘤分割的3D nnU-Net的跨癌症泛化能力的压力测试：使用新的食管癌和肺癌数据集的多队列评估 显示英文标题

标题： Stress-testing cross-cancer generalizability of 3D nnU-Net for PET-CT tumor segmentation: multi-cohort evaluation with novel oesophageal and lung cancer datasets

摘要： 鲁棒的泛化对于在临床PET-CT工作流程中部署基于深度学习的肿瘤分割至关重要，其中解剖部位、扫描仪和患者群体存在广泛变化。 本研究首次对PET-CT上的nnU-Net进行了跨癌症评估，并引入了两个新颖的专家标注的全身数据集。 279名食管癌患者（澳大利亚队列）和54名肺癌患者（印度队列）。 这些队列补充了公共AutoPET数据集，并实现了跨领域性能的系统性压力测试。 我们在三种范式下训练和测试了3D nnUNet模型。 仅目标（食管）、仅公共（AutoPET）和联合训练。 对于测试集，仅食管模型在内部域准确性方面表现最佳（平均DSC，57.8），但在外部印度肺癌队列上失败（平均DSC低于3.4），表明存在严重的过拟合。 仅公共模型泛化范围更广（在AutoPET上平均DSC为63.5，在印度肺癌队列上平均DSC为51.6），但在食管澳大利亚队列中表现不佳（平均DSC为26.7）。 联合方法提供了最平衡的结果（平均DSC，肺癌（52.9），食管（40.7），AutoPET（60.9）），减少了边界错误并在所有队列中提高了鲁棒性。 这些发现表明，数据集多样性，特别是多人口统计、多中心和多癌症整合，比架构新颖性更能推动鲁棒的泛化。 这项工作提出了基于人口统计的跨癌症深度学习分割评估，并强调数据集多样性，而不是模型复杂性，是临床鲁棒分割的基础。 显示英文摘要

摘要： Robust generalization is essential for deploying deep learning based tumor segmentation in clinical PET-CT workflows, where anatomical sites, scanners, and patient populations vary widely. This study presents the first cross cancer evaluation of nnU-Net on PET-CT, introducing two novel, expert-annotated whole-body datasets. 279 patients with oesophageal cancer (Australian cohort) and 54 with lung cancer (Indian cohort). These cohorts complement the public AutoPET dataset and enable systematic stress-testing of cross domain performance. We trained and tested 3D nnUNet models under three paradigms. Target only (oesophageal), public only (AutoPET), and combined training. For the tested sets, the oesophageal only model achieved the best in-domain accuracy (mean DSC, 57.8) but failed on external Indian lung cohort (mean DSC less than 3.4), indicating severe overfitting. The public only model generalized more broadly (mean DSC, 63.5 on AutoPET, 51.6 on Indian lung cohort) but underperformed in oesophageal Australian cohort (mean DSC, 26.7). The combined approach provided the most balanced results (mean DSC, lung (52.9), oesophageal (40.7), AutoPET (60.9)), reducing boundary errors and improving robustness across all cohorts. These findings demonstrate that dataset diversity, particularly multi demographic, multi center and multi cancer integration, outweighs architectural novelty as the key driver of robust generalization. This work presents the demography based cross cancer deep learning segmentation evaluation and highlights dataset diversity, rather than model complexity, as the foundation for clinically robust segmentation.