Scholarlink

<p>To develop and validate a two-level hierarchical fusion architecture enabling efficient few-shot domain adaptation for cervical cancer clinical target volume (CTV) auto-segmentation across different institutional imaging and contouring settings. We propose DNQ-UNet, a dual-encoder architecture with two fusion levels: spatially adaptive normalization (SPADE) for shallow appearance alignment and cross-attention for deep semantic transfer. A progressive two-stage training strategy with grouped learning rates was employed. Few-shot adaptation used 3, 5, 10, 20, 40, and 60 fine-tuning cases, and additional 3D fine-tuning baselines, including nnU-Net V2, SwinUNETR, and MedNeXt, were evaluated under identical target-domain test settings. Ablation studies assessed each fusion component. On the target-domain test set, zero-shot DSC was <InlineEquation ID="IEq1"> <EquationSource Format="TEX">\(0.8040\pm 0.0549\)</EquationSource> </InlineEquation>. With 3, 5, and 10 fine-tuning cases, DSC improved to <InlineEquation ID="IEq2"> <EquationSource Format="TEX">\(0.818\pm 0.073\)</EquationSource> </InlineEquation>, <InlineEquation ID="IEq3"> <EquationSource Format="TEX">\(0.830\pm 0.050\)</EquationSource> </InlineEquation>, and <InlineEquation ID="IEq4"> <EquationSource Format="TEX">\(0.838\pm 0.056\)</EquationSource> </InlineEquation>, respectively. DNQ-UNet achieved the highest DSC among the compared 3D baselines at 5-shot (0.830 vs. 0.805–0.817) and 10-shot (0.838 vs. 0.823–0.833) settings. Ablation showed significant DSC reductions after removing either fusion level or using a single U-Net (all <InlineEquation ID="IEq5"> <EquationSource Format="TEX">\(p&lt;0.001\)</EquationSource> </InlineEquation>). The proposed framework enables efficient few-shot domain adaptation for cervical cancer CTV segmentation using 5–10 annotated cases, providing a locally adapted contour-initialization approach that may reduce annotation and repeated correction burden during cross-institutional deployment.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Dnq-unet: a two-level fusion framework for few-shot domain adaptation in cervical cancer CTV segmentation

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年6月3日

查看全文
  • Boying Li,
  • Yihui Liu,
  • Yanfei Pan,
  • Yuhui Zhang

摘要

To develop and validate a two-level hierarchical fusion architecture enabling efficient few-shot domain adaptation for cervical cancer clinical target volume (CTV) auto-segmentation across different institutional imaging and contouring settings. We propose DNQ-UNet, a dual-encoder architecture with two fusion levels: spatially adaptive normalization (SPADE) for shallow appearance alignment and cross-attention for deep semantic transfer. A progressive two-stage training strategy with grouped learning rates was employed. Few-shot adaptation used 3, 5, 10, 20, 40, and 60 fine-tuning cases, and additional 3D fine-tuning baselines, including nnU-Net V2, SwinUNETR, and MedNeXt, were evaluated under identical target-domain test settings. Ablation studies assessed each fusion component. On the target-domain test set, zero-shot DSC was \(0.8040\pm 0.0549\) . With 3, 5, and 10 fine-tuning cases, DSC improved to \(0.818\pm 0.073\) , \(0.830\pm 0.050\) , and \(0.838\pm 0.056\) , respectively. DNQ-UNet achieved the highest DSC among the compared 3D baselines at 5-shot (0.830 vs. 0.805–0.817) and 10-shot (0.838 vs. 0.823–0.833) settings. Ablation showed significant DSC reductions after removing either fusion level or using a single U-Net (all \(p<0.001\) ). The proposed framework enables efficient few-shot domain adaptation for cervical cancer CTV segmentation using 5–10 annotated cases, providing a locally adapted contour-initialization approach that may reduce annotation and repeated correction burden during cross-institutional deployment.