基于图像的乳房假体检测用于乳腺X线摄影数据集整理与近实时部署：基础模型与任务特定卷积模型的比较
Image-Based Breast Implant Detection for Mammography Dataset Curation and Near-Real-Time Deployment: Comparing Foundation Models and Task-Specific Convolutional Models
- Emory University School of Medicine（埃默里大学医学院）
- University of Maryland（马里兰大学）
机构由 AI 辅助整理，请以论文原文为准。
AI总结：
本研究比较基础模型与任务特定CNN在二维乳腺X线摄影中检测乳房假体的性能，发现MammoCLIP和ResNetLite分别适合GPU医院和资源受限环境，实现近实时高精度部署。
AI中文摘要：
目的：评估基础模型（FMs）和从零训练的任务特定卷积神经网络（CNNs）在二维乳腺X线摄影中乳房假体分类的性能-可行性权衡，重点关注其是否适合近实时临床部署。方法：我们评估了四种模型：两个基础模型（RAD-DINO和MammoCLIP）和两个从零训练用于假体预测的CNN（ResNet18和我们的轻量级ResNetLite）。使用埃默里乳腺成像数据集，5,000张单侧筛查乳腺X线摄影用于训练/验证，1,000张人工审查的单侧图像作为测试集。对于基础模型，使用支持向量机（SVM）对预训练编码器的全局图像嵌入进行分类。CNN在二维乳腺X线摄影上进行端到端训练，ResNetLite通过网格搜索优化深度和宽度以平衡准确性和效率。使用AUROC、敏感性、特异性、准确性、嵌入可视化和推理延迟评估性能。结果：所有模型在保留测试数据（n=1,000）上均表现出强劲性能。MammoCLIP达到了最高的AUROC（0.999），训练时间最短，为493秒。RAD-DINO达到了最高的敏感性（0.980；准确性0.989），但推理和训练时间最慢。ResNet18和MammoCLIP的准确性相当（0.985）。ResNetLite与ResNet18相比无统计学显著差异（AUROC 0.993；准确性0.976），尽管仅使用ResNet18参数的1.4%，且推理时间最快。结论：基础模型和任务特定CNN模型能可靠地在二维乳腺X线摄影上检测乳房假体。模型选择最好由部署环境指导：MammoCLIP适用于配备GPU的医院环境，需要可扩展集成；轻量级CNN如ResNetLite适用于资源受限或边缘部署。
英文摘要：
Purpose: To evaluate the performance-feasibility tradeoffs of foundation models (FMs) and task-specific convolutional neural networks (CNNs) trained from scratch for breast implant classification in 2D mammography, with emphasis on suitability for near real-time clinical deployment. Methods: We evaluated four models: two FMs (RAD-DINO and MammoCLIP) and two CNNs trained from scratch for implant prediction (ResNet18 and our lightweight ResNetLite). Using the Emory Breast Imaging Dataset, 5,000 unilateral screening mammograms were used for training/validation and 1,000 manually reviewed unilateral images were held out for testing. For the FMs, global image embeddings from the pretrained encoder were classified using a support vector machine (SVM). The CNNs were trained end-to-end on 2D mammograms, with ResNetLite optimized via grid search over depth and width to balance accuracy and efficiency. Performance was evaluated using AUROC, sensitivity, specificity, accuracy, embedding visualization, and inference-latency. Results: All models demonstrated strong performance on held-out test data (n = 1,000). MammoCLIP achieved the highest AUROC (0.999) with the quickest training time of 493 seconds. RAD-DINO achieved the highest sensitivity (0.980; accuracy 0.989) but had the slowest inference and training times. ResNet18 and MammoCLIP achieved comparable accuracy (0.985). ResNetLite showed no statistically significant difference from ResNet18 (AUROC 0.993; accuracy 0.976) despite using only 1.4% of ResNet18's parameters, and had the fastest inference time. Conclusion: FMs and task-specific CNN models reliably detect breast implants on 2D mammography. Model selection is best guided by deployment context: MammoCLIP for GPU-equipped hospital settings requiring scalable integration, and lightweight CNNs such as ResNetLite for resource-constrained or edge deployments.
