文献速递头颈部肿瘤放疗放疗物理与技术

不确定性量化检测头颈部危及器官深度学习分割中的分布外病例

Uncertainty quantification for detecting out-of-distribution cases in deep learning segmentation of head and neck organs of interest.
PubMed 文献 · Radiother Oncol · 2026-10-13

本研究评估不确定性量化(UQ)能否检测头颈部危及器官自动分割中因分布偏移产生的分布外输入。基于蒙特卡洛dropout的不确定性评分与nnU-Net几何精度在两项实验中被评估,结果显示不确定性评分能反映学习到的训练分布,可检测分布偏移并支持自动QA工作流。

要点

  • 不确定性评分可反映模型学习到的训练分布,用于检测分布偏移。
  • 无金属模型对金属伪影与非伪影病例的不确定性评分分离度大于基线模型。
  • 窄模型在确定病例上分割精度显著优于不确定病例,宽模型无差异。

核心结论

  • 不确定性量化能够检测放疗自动分割中的分布偏移。
  • 该能力可通过识别分布内病例并应用专用模型来支持自动化QA工作流。

研究设计与样本

  • 评估基于蒙特卡洛dropout的不确定性评分与nnU-Net模型的几何精度,共两项实验。
  • 第一项为概念验证,在基线模型(混合训练数据,N=620)与无金属模型(无伪影训练数据,N=620)上评估不确定性评分能否检测金属伪影(N=10)与非伪影病例(N=10)。
  • 第二项为临床现实实验,按不确定性评分将微调数据集分为窄模型(低不确定性病例,N=21)与宽模型(高不确定性病例,N=21),并用留出测试集(N=76)评估不同不确定性类别的分割精度差异。

主要发现

  • 两项实验均显示不确定性评分反映学习到的训练分布。
  • 在监督实验中,无金属模型对伪影与非伪影病例的不确定性评分分离度(1.1×10^5 vs 0.6×10^5)大于基线模型(1.1×10^5 vs 0.8×10^5)。
  • 在无监督实验中,窄模型在确定病例上的表现显著优于不确定病例(p=0.04),而宽模型无差异(p=0.34),后者训练中包含了这些病例。

临床怎么用

  • 可利用不确定性量化识别分布内病例,从而应用专用模型并支持自动QA工作流。

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: Quality assurance (QA) of auto-segmentations is necessary for safe clinical deployment. Deep learning models are optimised for data similar to their training set, but reliability drops with out-of-distribution (OOD) inputs. This study evaluates whether uncertainty quantification (UQ) can detect OOD inputs arising from distribution shift in head and neck organ of interest (OOI) auto-segmentation. MATERIALS AND METHODS: Monte Carlo dropout-based uncertainty scores and geometric accuracy of nnU-Net models were evaluated in two experiments. First, as a proof-of-concept, we assessed whether uncertainty scores detect metal artefact (N = 10) from non-artefact cases (N = 10) on a baseline model (mixed training data, N = 620) and a no-metal model (artefact-free training data, N = 620). Second, in a clinically realistic experiment, we split a fine-tuning dataset by uncertainty score to obtain a narrow model (low-uncertainty cases, N = 21) and a broad model (high-uncertainty cases, N = 21). A held-out test set (N = 76) evaluated whether segmentation accuracy differed between uncertainty categories. RESULTS: Both experiments showed that uncertainty scores reflect the learned training distribution. In the supervised experiment, the no-metal model demonstrated greater uncertainty score separation between artefact and non-artefact cases (1.1 × 105 vs 0.6 × 105) than the baseline model (1.1 × 105 vs 0.8 × 105). In the unsupervised experiment, the narrow model performed significantly better on certain than uncertain cases (p = 0.04), while no difference was seen for the broad model (p = 0.34), which did include these cases in training. CONCLUSIONS: Uncertainty quantification can detect distribution shifts in radiotherapy auto-segmentation. This could support automated QA workflows by identifying in-distribution cases and enabling the application of specialised models.

来源:Radiother Oncol,查看原文。版权归原作者所有。

引用

  1. van Aalst JE, Simões R, Maruccio FC, et al. Uncertainty quantification for detecting out-of-distribution cases in deep learning segmentation of head and neck organs of interest. Radiother Oncol. 2026. PMID: 42869247. doi:10.1016/j.radonc.2026.111826

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。