Yuan Andy, Podder Tarun, Yuan Jiankui, Zheng Yiran
Youngstown State University, Youngstown, United States.
University Hospitals, Cleveland Medical Center, Cleveland, United States.
J Contemp Brachytherapy. 2023 Feb;15(1):69-74. doi: 10.5114/jcb.2023.125512. Epub 2023 Feb 27.
To apply a deep learning approach to automatically detect implanted seeds on a fluoroscopy image in prostate brachytherapy.
Forty-eight fluoroscopy images of patients, who underwent permanent seed implant (PSI) were used for this study after our Institutional Review Boards approval. Pre-processing procedures that were used to prepare for the training data, included encapsulating each seed in a bounding box, re-normalizing seed dimension, cropping to a region of prostate, and converting fluoroscopy image to PNG format. We employed a pre-trained faster region convolutional neural network (R-CNN) from PyTorch library for automatic seed detection, and leave-one-out cross-validation (LOOCV) procedure was applied to evaluate the performance of the model.
Almost all cases had mean average precision (mAP) greater than 0.91, with most cases (83.3%) having a mean average recall (mAR) above 0.9. All cases achieved F1-scores exceeding 0.91. The averaged results for all the cases were 0.979, 0.937, and 0.957 for mAP, mAR, and F1-score, respectively.
Although there are limitations shown in interpreting overlapping seeds, our model is reasonably accurate and shows potential for further applications.
应用深度学习方法在前列腺近距离放射治疗的荧光透视图像上自动检测植入的种子。
经机构审查委员会批准后，本研究使用了48例接受永久性种子植入（PSI）患者的荧光透视图像。用于准备训练数据的预处理程序包括将每个种子封装在一个边界框中、重新归一化种子尺寸、裁剪到前列腺区域以及将荧光透视图像转换为PNG格式。我们采用了来自PyTorch库的预训练更快区域卷积神经网络（R-CNN）进行种子自动检测，并应用留一法交叉验证（LOOCV）程序来评估模型的性能。
几乎所有病例的平均精度均值（mAP）均大于0.91，大多数病例（83.3%）的平均召回率均值（mAR）高于0.9。所有病例的F1分数均超过0.91。所有病例的mAP、mAR和F1分数的平均结果分别为0.979、0.937和0.957。
尽管在解释重叠种子方面存在局限性，但我们的模型具有合理的准确性，并显示出进一步应用的潜力。
