利用解剖先验和主动学习提高AGITG TOPGEAR临床试验中临床靶区的分割精度

中文摘要 AI 辅助

训练基于深度学习的医学图像分割模型面临着 curated 数据集有限的挑战。针对胃癌临床试验 AGITG TOPGEAR，其临床靶区（CTV）结构复杂，需由多个解剖学标志定义，这为自动轮廓 QA 分割模型的前期训练数据准备带来了困难。本研究探究了源自周围器官分割的解剖先验，以提供空间上下文并提升 TOPGEAR CTV 的分割精度，同时评估了主动学习方法——通过选择有望提升性能的病例，迭代扩充训练数据集。研究回顾性分析了 100 例 TOPGEAR CT 扫描，使用初始的 10 例经专家勾画的病例训练 nnU-Net 模型；TotalSegmentator 从周围结构生成体素级解剖先验图，作为额外输入通道。主动学习通过四轮迭代模拟，依据模型不确定性和分割性能选择病例，所有模型均采用五折交叉验证计算集成不确定性度量，评估使用 50 例保留的测试集。结果显示，解剖先验使 CTV 分割精度提升，平均 Dice 相似系数（DSC）从 0.84 增至 0.86；主动学习同样将性能提升至 0.86，在最后一轮迭代中获益最大；将解剖先验与主动学习结合，达到最高精度，DSC 为 0.87。模型不确定性与 DSC 相关，支持其用于识别次优预测并指导主动学习。解剖先验和主动学习均提升了 CTV 的分割精度与泛化能力，二者结合实现了最佳性能，支持将其整合至放疗临床试验自动轮廓 QA 的分割模型开发中。

英文摘要

Training deep learning-based medical image segmentation models is challenging with limited curated datasets. For AGITG TOPGEAR, a gastric cancer trial, the Clinical Target Volume (CTV) is complex and defined by multiple anatomical landmarks, making upfront training data preparation difficult for an automated contour QA segmentation model. We investigate anatomical priors, derived from surrounding organ segmentations, to provide spatial context and improve TOPGEAR CTV segmentation accuracy. We also evaluate active learning, iteratively expanding the training dataset by selecting cases expected to improve performance. One hundred TOPGEAR CT scans were retrospectively analyzed. An initial set of 10 expert-contoured cases was used to train an nnU-Net model. TotalSegmentator generated a voxel-wise anatomical prior map from surrounding structures as an additional input channel. Active learning was simulated over four iterations, selecting cases by model uncertainty and segmentation performance. All models used five-fold cross-validation for an ensemble uncertainty measure. Evaluation used a hold-out testing set of 50 cases. The anatomical prior improved CTV segmentation accuracy, increasing mean Dice Similarity Coefficient (DSC) from 0.84 to 0.86. Active learning similarly improved performance to 0.86, with greatest benefit in the final round. Combining the anatomical prior with active learning achieved the highest accuracy, with a DSC of 0.87. Model uncertainty correlated with DSC, supporting its use in identifying suboptimal predictions and guiding active learning. Anatomical priors and active learning each improved CTV segmentation accuracy and generalizability, with their combination achieving the best performance, supporting integration into segmentation model development for automated contour QA in radiotherapy clinical trials.

发表机构

• University of New South Wales（新南威尔士大学）
• Ingham Institute for Applied Medical Research（英厄姆应用医学研究所）
• Liverpool and Macarthur Cancer Therapy Centres（利物浦与麦克阿瑟癌症治疗中心）
• Peter MacCallum Cancer Centre（彼得·麦卡勒姆癌症中心）
• University of Melbourne（墨尔本大学）
• CSIRO Australian e-Health Research Centre（联邦科学与工业研究组织澳大利亚电子健康研究中心）
• University of Sydney（悉尼大学）
• The University of Western Australia（西澳大学）
• Sir Charles Gardiner Hospital（查尔斯·加德纳爵士医院）
• University of Wisconsin（威斯康星大学）
• Radformation Inc.（Radformation公司）

机构由 AI 辅助整理，请以论文原文为准。