Department of Otorhinolaryngology Head and Neck Surgery, the Peking Union Medical College Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China.
Int J Med Robot. 2023 Dec;19(6):e2567. doi: 10.1002/rcs.2567. Epub 2023 Aug 26.
Artificial intelligence (AI) techniques, especially deep learning (DL) techniques, have shown promising results for various computer vision tasks in the field of surgery. However, AI-guided navigation during microscopic surgery for real-time surgical guidance and decision support is much more complex, and its efficacy has yet to be demonstrated. We propose a model dedicated to the evaluation of DL-based semantic segmentation of chorda tympani (CT) during microscopic surgery.
Various convolutional neural networks were constructed, trained, and validated for semantic segmentation of CT. Our dataset has 5817 images annotated from 36 patients, which were further randomly split into the training set (90%, 5236 images) and validation set (10%, 581 images). In addition, 1500 raw images from 3 patients (500 images randomly selected per patient) were used to evaluate the network performance.
When evaluated on a validation set (581 images), our proposed CT detection networks achieved great performance, and the modified U-net performed best (mIOU = 0.892, mPA = 0.9427). Moreover, when applying U-net to predict the test set (1500 raw images from 3 patients), our methods also showed great overall performance (Accuracy = 0.976, Precision = 0.996, Sensitivity = 0.979, Specificity = 0.902).
This study suggests that DL can be used for the automated detection and segmentation of CT in patients with otosclerosis during microscopic surgery with a high degree of performance. Our research validated the potential feasibility for future vision-based navigation surgical assistance and autonomous surgery using AI.
人工智能（AI）技术，尤其是深度学习（DL）技术，在手术领域的各种计算机视觉任务中显示出了有前景的结果。然而，用于微创手术中实时手术指导和决策支持的 AI 引导导航要复杂得多，其疗效尚待证明。我们提出了一种专门用于评估基于深度学习的鼓索（CT）显微手术中语义分割的模型。
构建、训练和验证了各种卷积神经网络，用于 CT 的语义分割。我们的数据集有 36 名患者的 5817 张图像标注，进一步随机分为训练集（90%，5236 张图像）和验证集（10%，581 张图像）。此外，还使用了来自 3 名患者的 1500 张原始图像（每位患者随机选择 500 张图像）来评估网络性能。
在验证集（581 张图像）上评估时，我们提出的 CT 检测网络表现出色，改进的 U-net 表现最佳（mIOU=0.892，mPA=0.9427）。此外，当将 U-net 应用于预测测试集（3 名患者的 1500 张原始图像）时，我们的方法也表现出了出色的整体性能（准确性=0.976，精度=0.996，灵敏度=0.979，特异性=0.902）。
本研究表明，DL 可用于在显微镜下对耳硬化症患者的 CT 进行自动检测和分割，具有很高的性能。我们的研究验证了基于 AI 的未来视觉导航手术辅助和自主手术的潜在可行性。
