文献速递胸部肿瘤放疗放疗物理与技术

面向胸部放疗的心脏CT亚结构自动分割领域适应基础模型

Domain-adapted foundation model for automated cardiac CT substructure segmentation for thoracic radiotherapy.
PubMed 文献 · Radiother Oncol · 2026-06-03

本研究构建了含69例左侧肺癌患者、21个心脏亚结构专家标注的增强放疗模拟CT数据集CardiacSubstructSeg,提出基于CT领域自适应自监督预训练的分割框架DINO-CardiacSeg。五折交叉验证中该模型整体性能最高(DSC 65.95%±7.91%,HD95 7.91±4.40 mm),在TotalSegmentator心脏子集(n=603)微调迁移评估中保持优势(DSC 90.81%±8.23%,HD95 6.07±13.17 mm)。

要点

  • DINO-CardiacSeg在内部数据集上取得最高整体分割性能,小结构和低对比度结构增益一致。
  • 迁移至TotalSegmentator心脏子集微调后仍保持优越性能,显示跨标注协议与影像特征适应能力。
  • 该框架支持亚结构级心脏剂量学与心脏毒性风险分层,但尚需临床验证。

核心结论

  • DINO-CardiacSeg通过CT特异性基础模型预训练,可在胸部放疗模拟CT上实现准确且可扩展的临床相关心脏亚结构分割。
  • 该方法在内部评估和公开心脏CT数据集微调中均优于现有方法,支持亚结构级心脏剂量学及改善肺癌放疗心脏毒性风险分层的潜力。

研究设计与样本

  • 研究构建CardiacSubstructSeg数据集,含69例左侧肺癌患者的增强放疗模拟CT及21个心脏亚结构的专家手动标注。
  • 提出DINO-CardiacSeg分割框架,基于CT领域自适应自监督预训练,采用五折交叉验证并与现有最先进方法比较。
  • 通过微调并在TotalSegmentator心脏子集(n=603)上评估迁移能力。

主要发现

  • 在CardiacSubstructSeg上,DINO-CardiacSeg在所有指标上整体性能最高(DSC 65.95%±7.91%,HD95 7.91±4.40 mm)。
  • 性能提升在小结构和低对比度结构上保持一致。
  • 在TotalSegmentator迁移学习评估中,DINO-CardiacSeg保持优越性能(DSC 90.81%±8.23%,HD95 6.07±13.17 mm),显示微调后对不同标注协议和影像特征数据集的有效适应。

临床怎么用

  • 可用于胸部放疗模拟CT上心脏亚结构的自动分割,支持亚结构级心脏剂量学评估。
  • 有望辅助肺癌放疗中心脏毒性风险分层,但需进一步临床验证。

以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。

原文

PURPOSE: Radiation-induced cardiotoxicity contributes to non-cancer mortality in patients treated with thoracic radiotherapy. Radiation dose to specific cardiac substructures is more predictive of adverse cardiac outcomes than whole-heart dose. This study aimed to develop a robust, automated cardiac substructure segmentation framework tailored to thoracic RT imaging. METHODS AND MATERIALS: We curated CardiacSubstructSeg, a contrast-enhanced RT simulation CT dataset comprising 69 patients with left-sided lung cancer and expert manual annotations of 21 cardiac substructures. We proposed DINO-CardiacSeg, a novel segmentation framework built upon a CT domain-adaptive self-supervised pretraining. Model performance was assessed using five-fold cross-validation and compared with state-of-the-art methods. Transferability was assessed via fine-tuning and evaluation on the TotalSegmentator cardiac subset (n = 603). RESULTS: On CardiacSubstructSeg, DINO-CardiacSeg achieved the highest overall performance across all metrics (65.95% ± 7.91% DSC and 7.91 ± 4.40 mm HD95). Performance gains were consistent across small and low-contrast structures. In transfer-learning evaluation on TotalSegmentator, DINO-CardiacSeg maintained superior performance (DSC: 90.81% ± 8.23%, HD95: 6.07 ± 13.17 mm), demonstrating effective adaptation after fine-tuning to a dataset with differing annotation protocols and imaging characteristics. CONCLUSIONS: DINO-CardiacSeg enables accurate and scalable segmentation of clinically relevant cardiac substructures on thoracic RT simulation CT through CT-specific foundation model pretraining. This approach outperforms existing methods on internal evaluation and fine-tuning to a public cardiac CT dataset, supporting its potential for substructure-level cardiac dosimetry and improved cardiotoxicity risk stratification in lung radiotherapy.

来源:Radiother Oncol,查看原文。版权归原作者所有。

引用

  1. Li Y, Wildman VL, Balzo LD, et al. Domain-adapted foundation model for automated cardiac CT substructure segmentation for thoracic radiotherapy. Radiother Oncol. 2026. PMID: 42242380. doi:10.1016/j.radonc.2026.111619

供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。