探索临床因素在零样本分割VLMs中的提示对齐用于NSCLC肿瘤分割
Exploring Prompt Alignment with Clinical Factors in Zero-Shot Segmentation VLMs for NSCLC Tumor Segmentation
AI总结：
研究通过分析VoxTell在NSCLC肿瘤数据集上的提示对齐方向，发现解剖位置主导空间注意力，零样本分割VLMs在NSCLC肿瘤分割中表现优于其他模型。
摘要：
零样本视觉-语言模型（VLMs）为非小细胞肺癌（NSCLC）的大肿瘤体积（GTV）勾画提供了一种可提示的替代方法，但其空间行为的提示维度仍不明确。我们通过子提示分解、属性扰动鲁棒性分析、特异性梯度和跨病例提示交换等方法，研究VoxTell在保留内部NSCLC肿瘤数据集上的对齐方向。通过Dice相似性系数（DSC）和Wilcoxon符号秩检验及Benjamini-Hochberg校正，与微调和零样本基线模型进行比较。对齐分析显示，解剖位置是VoxTell空间注意力的主要驱动因素：63.4%的定位扰动导致灾难性下降，提示特异性从通用到完整描述提升，但诊断-only提示除外。无关提示正确导致零分割，跨病例提示交换确认了患者特异性条件（匹配DSC 0.906 vs. 不匹配 0.406）。组织学和分期替换影响较小，表明模型优先考虑“在哪里寻找”而非“寻找什么”。在这一背景下，VoxTell在零样本条件下实现了平均DSC 0.613，统计上与nnUNet（0.690，调整p=0.156）和Ahmed等人（0.675，调整p=0.679）不可区分，但显著优于所有其他零样本模型。这些发现表明，分割VLMs应不仅通过Dice评估，还应通过其对齐的提示维度进行评估。
英文摘要：
Zero-shot vision-language models (VLMs) offer a promptable alternative to task-specific training for gross tumor volume (GTV) delineation in non-small-cell lung cancer (NSCLC), but the prompt dimensions that govern their spatial behavior remain poorly understood. We study this question by probing alignment directions in VoxTell on a held-out internal NSCLC tumor dataset through sub-prompt decomposition into diagnosis, demographic, staging, anatomical, generic, and irrelevant controls; attribute-wise perturbation robustness; specificity ladders; and cross-case prompt swaps, while benchmarking against fine-tuned and zero-shot baselines using the Dice Similarity Coefficient (DSC) with Wilcoxon signed-rank tests and Benjamini-Hochberg correction. Alignment analyses revealed that anatomical location is the dominant driver of VoxTell's spatial attention: 63.4 percent of location perturbations caused catastrophic drops, prompt specificity improved from generic to full descriptions except for diagnosis-only prompts, irrelevant prompts correctly yielded zero segmentation, and cross-case prompt swaps confirmed patient-specific conditioning (matched DSC 0.906 vs. mismatched 0.406). Histology and stage substitutions had minimal effect, indicating that the model prioritizes "where to look" over "what to look for." In this context, VoxTell, operating fully zero-shot, achieved a mean DSC of 0.613, statistically indistinguishable from nnUNet (0.690, adjusted p = 0.156) and Ahmed et al. (0.675, adjusted p = 0.679), while significantly outperforming all other zero-shot models. Together, these findings argue that segmentation VLMs should be evaluated not only by Dice, but also by the prompt dimensions to which they align.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
