Maruccio Federica C, Simões Rita, van Aalst Joëlle E, Brouwer Charlotte L, Sonke Jan-Jakob, van Ooijen Peter, Janssen Tomas M
Department of Radiation Oncology, The Netherlands Cancer Institute, Amsterdam, the Netherlands.
University of Groningen, University Medical Center Groningen, Department of Radiation Oncology, Groningen, the Netherlands.
Phys Imaging Radiat Oncol. 2025 May 8;34:100771. doi: 10.1016/j.phro.2025.100771. eCollection 2025 Apr.
While Deep Learning (DL) auto-segmentation has the potential to improve segmentation efficiency in the radiotherapy workflow, manual adjustments of the predictions are still required. Network uncertainty quantification has been proposed as a quality assurance tool to ensure an efficient segmentation workflow. However, the interpretation is often complicated due to various sources of uncertainty interacting non-trivially. In this work, we compared network predictions with both independent manual segmentations and manual corrections of the predictions. We assume that manual corrections only address clinically relevant errors and are therefore associated with lower aleatoric uncertainty due to less inter-observer variability. We expect the remaining epistemic uncertainty to be a better predictor of segmentation corrections.
We considered DL auto-segmentations of the mesorectum clinical target volume. Uncertainty maps of nnU-Net outputs were generated using Monte Carlo dropout. On a global level, we investigated the correlation between mean network uncertainty and network segmentation performance. On a local level, we compared the uncertainty envelope width with the length of the error from both independent contours and corrected predictions. The uncertainty envelope widths were used to classify the error lengths as above or below a predefined threshold.
We achieved an AUC above 0.9 in identifying regions manually corrected with edits larger than 8 mm, while the AUC for inconsistencies with the independent contours was significantly lower at approximately 0.7.
Our results validate the hypothesis that epistemic uncertainty estimates are a valuable tool to capture regions likely requiring clinically relevant edits.
虽然深度学习（DL）自动分割有潜力提高放射治疗工作流程中的分割效率，但仍需要对预测结果进行人工调整。网络不确定性量化已被提议作为一种质量保证工具，以确保高效的分割工作流程。然而，由于各种不确定性来源以非平凡的方式相互作用，其解释往往很复杂。在这项工作中，我们将网络预测结果与独立的手动分割以及对预测结果的人工校正进行了比较。我们假设人工校正仅处理临床相关误差，因此由于观察者间变异性较小，与较低的偶然不确定性相关。我们期望剩余的认知不确定性能更好地预测分割校正。
我们考虑了直肠系膜临床靶体积的DL自动分割。使用蒙特卡洛随机失活生成nnU-Net输出的不确定性图。在全局层面，我们研究了平均网络不确定性与网络分割性能之间的相关性。在局部层面，我们将不确定性包络宽度与来自独立轮廓和校正预测的误差长度进行了比较。不确定性包络宽度用于将误差长度分类为高于或低于预定义阈值。
在识别编辑大于8毫米的手动校正区域时，我们实现了AUC高于0.9，而与独立轮廓不一致的AUC显著较低，约为0.7。
我们的结果验证了以下假设，即认知不确定性估计是一种有价值的工具，可用于捕获可能需要临床相关编辑的区域。
