多任务形态学概念学习用于3D CT肺结节恶性评估
原标题：Evaluating Multi-Task Morphological Concept Learning for Pulmonary Nodule Malignancy Assessment in 3D CT
AI 摘要
该研究基于LIDC-IDRI数据集（742名患者、3918条阅片者级结节标注），比较单任务3D CNN与同时预测恶性风险、毛刺征和分叶征的多任务模型在肺结节恶性评估中的表现。单任务模型平衡准确率0.548、ROC-AUC 0.552，多任务模型分别为0.539和0.558，患者级bootstrap分析显示ROC-AUC差异仅0.005（95%CI -0.087至0.090）。结果表明加入形态学辅助任务并未明显改善恶性风险分类，凸显类别平衡、标签构建和阅片者级标注结构在多任务肺部CT分析中的重要性。
正文节选
Evaluating Multi-Task Morphological Concept Learning for Pulmonary Nodule Malignancy Assessment in 3D CT Abstract Morphological characteristics such as spiculation and lobulation play an important role in assessing pulmonary nodules on computed tomography (CT), particularly in relation to malignancy risk. This study examines whether learning radiologist-annotated morphological features together with malignancy risk from lesion-centred 3D CT volumes improves classification performance. The Lung I