基于深度学习在中国成人人群中自动定量评估正常胰腺
本研究基于2778例胰腺CT图像构建级联两个3D U-Net网络的胰腺自动分割模型,测试集平均DSC为0.94,外部验证集3189例正常胰腺CT图像自动分割客观评估满意率为90%,并测量正常胰腺各期径线、体积及CT值,分析其与年龄、性别、身高、体重的关系。
要点
- 级联3D U-Net模型在测试集分割胰腺平均DSC为0.94。
- 外部验证自动分割客观评估满意率为90%。
- 正常胰腺体积在18–38岁最大,69–85岁降至最小。
核心结论
- 基于深度学习的胰腺分割工具能在CT图像上分割胰腺并准确有效地测量其正常径线、体积和CT值。
研究设计与样本
- 回顾性收集2778例胰腺图像(数据集1),随机分为训练集2252例、验证集245例和测试集281例。
- 通过级联应用两个3D U-Net网络构建胰腺分割模型,以Dice相似系数(DSC)评估分割效率。
- 另获取3189例正常胰腺CT图像(数据集2)进行外部验证,包括平扫1063例、动脉期1063例和门静脉期1063例。
- 数据集2的胰腺分割由两名放射科医师进行客观评估和手动修正,并计算各期胰腺体积、径线和平均CT值,分析胰腺体积与年龄、性别、身高、体重的关系。
关键影像表现/征象
- 数据集2中胰腺径线分别为43.71–44.28 mm、67.40–68.15 mm和114.53–117.06 mm。
- 平均胰腺体积为63,969.06–65,247.75 mm³,在18–38岁最大,69–85岁降至最小。
- 胰腺CT值随年龄增长而降低,从最大值38.87±9.70 HU降至最小值27.72±10.85 HU。
诊断标准与关键数值
- 数据集1测试集胰腺分割平均DSC为0.94。
- 数据集2外部验证自动分割胰腺的客观评估满意率为90%。
临床应用与阅片要点
- 该工具可自动分割CT图像上的胰腺并测量正常径线、体积和CT值,可用于正常胰腺的定量评估。
局限性
- 原文未报告局限性。
以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。
英文原文(摘要)· 点击展开
OBJECTIVE: To develop a 3D U-Net-based model for the automatic segmentation of the pancreas using the diameters, volume, and density of normal pancreases among Chinese adults.
METHODS: A total of 2778 pancreas images (dataset 1) were retrospectively collected and randomly divided into training (n = 2252), validation (n = 245), and test (n = 281) datasets. The segmentation model for the pancreas was constructed through cascaded application of two 3D U-Net networks. The segmentation efficiency for the pancreas was evaluated by the Dice similarity coefficient (DSC). Another dataset of 3189 normal pancreas CT images (dataset 2) was obtained for external validation, including 1063 non-contrast images, 1063 arterial phase images, and 1063 portal venous phase images. The pancreas segmentation in dataset 2 was assessed objectively and manually revised by two radiologists. Then, the pancreatic volume, diameters, and average CT value for each phase of pancreas images in dataset 2 were calculated. The relationships between pancreas volume and age, sex, height, and weight were analyzed.
RESULTS: In dataset 1, a mean DSC of 0.94 for the test dataset was achieved. In dataset 2, the objective assessment yielded a 90% satisfaction rate for the automatic segmentation of the pancreas as external validation. The diameters of the pancreas were 43.71-44.28 mm, 67.40-68.15 mm, and 114.53-117.06 mm, respectively. The average pancreatic volume was 63,969.06-65,247.75 mm3, which was greatest at the age of 18-38 and then decreased to a minimum at the age of 69-85. The CT value of the pancreas also decreased with age, from a maximum value of 38.87 ± 9.70 HU to a minimum of 27.72 ± 10.85 HU.
CONCLUSION: The pancreas segmentation tool based on deep learning can segment the pancreas on CT images and measure its normal diameter, volume, and CT value accurately and effectively.
引用
- Cai, Guo, Wang, et al. Automatic quantitative evaluation of normal pancreas based on deep learning in a Chinese adult population. Abdom Radiol (NY). 2022. PMID: 35064795. doi:10.1007/s00261-021-03327-x
供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。