深度学习计算机视觉可识别妊娠早期颈项透明层增厚
研究收集560例孕11至13周6天胎儿标准正中矢状面超声图像,其中88例(15.7%)颈项透明层(NT)厚于3.5毫米。训练U-Net卷积神经网络分割胎儿与NT区域并计算NT与胎儿比值,模型分割所得比值在测试集检测厚NT的AUROC为0.96,在91%特异度下检出94%厚NT病例。
要点
- 模型分割所得NT与胎儿比值检测厚NT的AUROC为0.96,在91%特异度下检出94%厚NT病例。
- 图像质量高者检测率高于低者(AUROC 0.98 vs 0.90,p=0.003)。
- 作者认为该可解释深度学习方法可用于妊娠早期其他胎儿异常的筛查。
核心结论
- 该模型提供了一种可解释的深度学习方法来检测NT增厚,该技术可用于筛查妊娠早期其他胎儿异常。
研究设计与样本
- 收集560例孕11至13周6天胎儿标准正中矢状面胎儿头部和胸部超声图像,其中88例(15.7%)NT厚于3.5毫米。
- 图像质量由两名胎儿医学专家评定为高或低。
- 图像分为训练集451例(55例厚NT)和测试集109例(33例厚NT)。
- 训练U-Net卷积神经网络分割胎儿和NT区域并计算NT与胎儿比值,先用真实分割验证指标,再用预测分割验证算法,均使用受试者工作特征曲线下面积(AUROC)。
主要发现
- 真实分割的NT与胎儿比值在训练集和测试集检测厚NT的AUROC均为0.97。
- 测试集中胎儿和NT区域检测的Dice分数为0.94。
- 基于模型分割的NT与胎儿比值相对临床医生标签检测厚NT的AUROC为0.96。
- 测试集中,在91%特异度下,94%的厚NT病例被检出(敏感度)。
- 图像质量高者检测率高于低者(AUROC 0.98 vs 0.90,p=0.003)。
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)。
原文
OBJECTIVE: Many fetal anomalies can already be diagnosed by ultrasound in the first trimester of pregnancy. Unfortunately, in clinical practice, detection rates for anomalies in early pregnancy remain low. Our aim was to use an automated image segmentation algorithm to detect one of the most common fetal anomalies: a thickened nuchal translucency (NT), which is a marker for genetic and structural anomalies. METHODS: Standardized mid-sagittal ultrasound images of the fetal head and chest were collected for 560 fetuses between 11 and 13 weeks and 6 days of gestation, 88 (15.7%) of whom had an NT thicker than 3.5 mm. Image quality was graded as high or low by two fetal medicine experts. Images were divided into a training-set (n = 451, 55 thick NT) and a test-set (n = 109, 33 thick NT). We then trained a U-Net convolutional neural network to segment the fetus and the NT region and computed the NT:fetus ratio of these regions. The ability of this ratio to separate thick (anomalous) NT regions from healthy, typical NT regions was first evaluated in ground-truth segmentation to validate the metric and then with predicted segmentation to validate our algorithm, both using the area under the receiver operator curve (AUROC). RESULTS: The ground-truth NT:fetus ratio detected thick NTs with 0.97 AUROC in both the training and test sets. The fetus and NT regions were detected with a Dice score of 0.94 in the test set. The NT:fetus ratio based on model segmentation detected thick NTs with an AUROC of 0.96 relative to clinician labels. At a 91% specificity, 94% of thick NT cases were detected (sensitivity) in the test set. The detection rate was statistically higher (p = 0.003) in high versus low-quality images (AUROC 0.98 vs. 0.90, respectively). CONCLUSION: Our model provides an explainable deep-learning method for detecting increased NT. This technique can be used to screen for other fetal anomalies in the first trimester of pregnancy.
引用
- Kasera B, Shinar S, Edke P, et al. Deep-learning computer vision can identify increased nuchal translucency in the first trimester of pregnancy. Prenat Diagn. 2024. PMID: 38558081. doi:10.1002/pd.6559
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。