用解剖结构自动过滤假阳性，提升脑动脉瘤检测准确率与可解释性。
Automated anatomy-based post-processing reduces false positives and improved interpretability of deep learning intracranial aneurysm detection
- 基于解剖知识的混合后处理方法，通过多层掩膜筛选减少误报。
- 最高可降低70.6%假阳性，且不丢失真阳性，每例假阳性从1.27降至0.62。
- 适合临床部署，提升深度学习模型的可解释性与医生信任度。
深度学习模型在CTA上检测颅内动脉瘤虽有进展，但高假阳性率仍是临床应用的主要障碍。本文提出一种自动化、基于解剖结构的启发式学习混合后处理方法，用于进一步降低假阳性。采用两个模型CPM-Net和3D-CNN-TR，在1,186例开源CTA（共1,373个标注动脉瘤）上训练，并在143例保留私有CTA（218个标注动脉瘤）上评估。通过脑组织、动脉、静脉及海绵窦（CVS）的分割掩膜，对深度学习输出进行五种规则过滤：(1) 脑掩膜；(2) 静脉掩膜；(3) 静脉>动脉区域；(4) 脑+静脉掩膜；(5) 脑+静脉>动脉掩膜。结果显示，3D-CNN-TR模型获得179个真阳性、39个假阴性、182个假阳性。假阳性主要为颅外（3D-CNN-TR占42.3%）、静脉（29.1%）、动脉（53.3%）及非血管结构（9.3%）。方法5表现最佳，使CPM-Net假阳性减少70.6%（126→37），3D-CNN-TR减少51.6%（182→88），且未损失真阳性，假阳性/例从1.27降至0.62，有效提升性能与可解释性。
原文摘要 · Abstract (English)
Introduction: Deep learning (DL) models can help detect intracranial aneurysms on CTA, but high false positive (FP) rates remain a barrier to clinical translation, despite improvement in model architectures and strategies like detection threshold tuning. We employed an automated, anatomy-based, heuristic-learning hybrid artery-vein segmentation post-processing method to further reduce FPs. Methods: Two DL models, CPM-Net and a deformable 3D convolutional neural network-transformer hybrid (3D-CNN-TR), were trained with 1,186 open-source CTAs (1,373 annotated aneurysms), and evaluated with 143 held-out private CTAs (218 annotated aneurysms). Brain, artery, vein, and cavernous venous sinus (CVS) segmentation masks were applied to remove possible FPs in the DL outputs that overlapped with: (1) brain mask; (2) vein mask; (3) vein more than artery masks; (4) brain plus vein mask; (5) brain plus vein more than artery masks. Results: CPM-Net yielded 139 true-positives (TP); 79 false-negative (FN); 126 FP. 3D-CNN-TR yielded 179 TP; 39 FN; 182 FP. FPs were commonly extracranial (CPM-Net 27.3%; 3D-CNN-TR 42.3%), venous (CPM-Net 56.3%; 3D-CNN-TR 29.1%), arterial (CPM-Net 11.9%; 3D-CNN-TR 53.3%), and non-vascular (CPM-Net 25.4%; 3D-CNN-TR 9.3%) structures. Method 5 performed best, reducing CPM-Net FP by 70.6% (89/126) and 3D-CNN-TR FP by 51.6% (94/182), without reducing TP, lowering the FP/case rate from 0.88 to 0.26 for CPM-NET, and from 1.27 to 0.62 for the 3D-CNN-TR. Conclusion: Anatomy-based, interpretable post-processing can improve DL-based aneurysm detection model performance. More broadly, automated, domain-informed, hybrid heuristic-learning processing holds promise for improving the performance and clinical acceptance of aneurysm detection models.
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。