发表机构
University of Strasbourg; CLCC Institut Strauss(斯特拉斯堡大学; CLCC施特劳斯研究所)
机构由 AI 辅助整理,请以论文原文为准。AI 中文总结
本文评估 nnU-Net 在 BraTS-GoAT 2026 异质脑肿瘤人群中的泛化能力,通过五折交叉验证训练,发现跨人群性能下降,且失败与 ET 体积小、成分分散相关。
AI 中文摘要
BraTS-GoAT 评估了异质人群中的肿瘤分割性能。我们使用五折交叉验证,在 1,351 个标注病例上训练了传统的 3D nnU-Net,每折训练 1,000 个 epoch。最终预测器对所有折的结果进行平均,并应用测试时镜像增强。在合并的官方验证集上,增强肿瘤(ET)、肿瘤核心(TC)和全肿瘤(WT)的全局 DSC 值分别为 0.7805、0.8288 和 0.8854。在匹配的折 0 推理下,区域平均 Dice 从源折外(OOF)病例的 0.9058 下降到合并验证集的 0.8310(差异为 -0.0747)。镜像增强在单折上带来小幅提升,但未显示明确的集成收益;残差编码器替代方案达到了 0.8282 的平均 Dice。在标注的 OOF 预测中,失败病例的参考 ET 体积明显更小;在调整 ET 和 WT 体积后,较低的 Dice 仍与更多不连通的 ET 成分以及最大成分中 ET 占比更小相关。
英文摘要
BraTS-GoAT evaluates tumor segmentation across heterogeneous populations. We trained a conventional 3D nnU-Net on 1,351 labeled cases using five-fold cross-validation and 1,000 epochs per fold. The final predictor averaged all folds and applied test-time mirroring. On pooled official validation, global DSC values were 0.7805, 0.8288, and 0.8854 for enhancing tumor (ET), tumor core (TC), and whole tumor (WT). Under matched fold-0 inference, mean regional Dice decreased from 0.9058 on source out-of-fold (OOF) cases to 0.8310 on pooled validation (difference--0.0747). Mirroring gave small single-fold gains but no clear ensemble benefit; a residual-encoder alternative reached 0.8282 mean Dice. In labeled OOF predictions, failure cases had substantially smaller reference ET volumes; after adjustment for ET and WT volume, lower Dice remained associated with more disconnected ET components and a smaller fraction of ET contained in the largest component.
Journal refBrainWorks 2026 -- The Brain Abnormality Workshop @MICCAI 2026, Sep 2026, Strasbourg, France