CT扫描参数差异会显著影响肺结节AI检测结果，但现有元数据无法捕捉此问题。
Acquisition state behaves as a structured, measurable variable governing lung-nodule AI: kernel-driven measurement instability and noise-driven detection fragility, invisible to DICOM metadata
- 通过重建核函数和噪声扰动实验，验证扫描参数是可测量的结构化变量。
- 不同核函数导致5.2%结节大小分类改变，而噪声主要影响小结节检测信心。
- 现有DICOM元数据无法识别重建信息，需引入输入侧验证机制。
医学影像AI治理正趋于规范化：2026年ACR-SIIM实践指南建议进行本地验收测试与持续漂移监控，且ACR Assess-AI注册库利用DICOM元数据监测AI输出上下文。我们认为，当前未被监控的一个必要层面存在于输出指标之下：待测影像是否仍处于模型验证时所覆盖的采集范围之内。我们以在LUNA16上训练的MONAI RetinaNet肺结节检测器为对象，检验采集状态是否为一种结构化、可度量的变量。在仅重构核函数不同的真实配对CT（NLST B30f vs B80f）中，核函数本身导致AI测量直径变化，并使155个结节中有8个（5.2%）的Fleischner大小类别发生改变，而检测置信度无显著变化（Wilcoxon p=0.22）。在受控的LIDC-IDRI扰动实验中，噪声轴显著降低检测置信度（p=5.9e-32，集中于小于6 mm的结节），但不影响测量；频率/核函数轴则破坏测量（p=8.6e-13），但不干扰检测。一个四特征像素指纹可恢复重建身份（真实CT患者级AUC约0.95，QIBA幻影上达0.995），而卷积核（ConvolutionKernel）DICOM标签则完全无区分能力（各重建间标签一致）。该核函数效应跨四个厂商迁移（留一厂商外验证AUC 0.94–0.98，匹配厂商内上限）。因此，采集状态映射到不同类型的AI失效模式：频谱内容决定测量可靠性，噪声水平影响检测敏感性，且无法从元数据中还原。面向采集状态的输入端验证，是当前影像AI认证中验收测试与漂移监控要求所缺失的关键环节。
原文摘要 · Abstract (English)
AI governance for medical imaging is formalizing: the 2026 ACR-SIIM Practice Parameter recommends local acceptance testing and ongoing drift monitoring, and the ACR Assess-AI registry monitors AI outputs using DICOM metadata for context. We argue that a necessary, currently unmonitored layer sits beneath output metrics: whether incoming studies remain within the acquisition envelope a model was validated on. Using a LUNA16-trained MONAI RetinaNet lung-nodule detector, we test whether acquisition state behaves as a structured, measurable variable. On real paired CT differing only in reconstruction kernel (NLST B30f vs B80f), kernel alone shifted AI-measured diameter and flipped a Fleischner size category in 5.2% (8 of 155) of nodules at fixed patient and acquisition, while detection confidence was unchanged (Wilcoxon p=0.22). Under controlled LIDC-IDRI perturbations the effects dissociated by axis: the noise axis degraded detection confidence (p=5.9e-32, concentrated in nodules under 6 mm) but not measurement, while the frequency/kernel axis corrupted measurement (p=8.6e-13) but not detection. A 4-feature pixel fingerprint recovered reconstruction identity (patient-level AUC about 0.95 on real CT, 0.995 on a QIBA phantom) where the ConvolutionKernel DICOM tag was uninformative (identical labels across reconstructions). The kernel axis transported across four manufacturers (leave-one-vendor-out AUC 0.94-0.98, matching the within-vendor ceiling). Acquisition state thus maps to distinct AI failure modes, frequency content to measurement reliability and noise to detection sensitivity, and is not recoverable from metadata. Acquisition-aware, input-side validation is the missing layer for the acceptance-testing and drift-monitoring requirements now entering imaging-AI accreditation.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。