Three- versus two-dimensional ultrasound for nuchal translucency thickness measurements: comparison of feasibility and levels of agreement

We have determined feasibility and levels of agreement for visualizing the nuchal translucency within a time limit with three-dimensional (3D) and two-dimensional (2D) transvaginal ultrasound. A total of 213 women of median gestational age 13 weeks (range 11–14 weeks) undergoing first-trimester nuchal translucency (NT) thickness measurements with 2D transabdominal ultrasound were included in the study. Additionally all women underwent 2D and 3D transvaginal NT examination within a time limit of 90 s. These two methods were compared with each other and with our standard method (2D transabdominal ultrasound) with respect to visualization of fetal nuchal fold and to the levels of agreement. The nuchal fold was visualized by 3D and 2D transvaginal ultrasound in 85.9% (95% CI: 80.5%, 90.3%) and 25.8% (95% CI: 20.1%, 32.2%), respectively (p<0.001). There was a statistically significant underestimation of the NT by 2D transabdominal as compared with 3D transvaginal ultrasound of 0.1 mm (p<0.001), and by 3D transvaginal as compared with 2D transvaginal ultrasound of even 0.1 mm (p<0.001). 3D transvaginal ultrasound of the nuchal fold has increased feasibility to 2D transvaginal ultrasound within a short examination time and with minimal, but significant, measurement differences. Copyright © 2001 John Wiley & Sons, Ltd.     

Accuracy of Down syndrome risks produced in a first-trimester screening programme incorporating fetal nuchal translucency thickness and maternal serum biochemistry

Over the past three years approximately 12 000 women have been screened in the first trimester through our OSCAR programme, which utilizes fetal NT and maternal serum free β-hCG and PAPP-A. During this time 30 cases of Down syndrome were identified either prenatally or postnatally. Using an established procedure the accuracy of predicted risk for Down syndrome was assessed in a population of 30 cases of Down syndrome and 11 758 unaffected pregnancies. The correlation between predicted risk and prevalence of Down syndrome was very high (r=0.9995). It is concluded that risks produced by the Fetal Medicine Foundation combined risk algorithm agree very closely with Down syndrome prevalence and can be used with confidence when counselling women of their risk. Copyright © 2002 John Wiley & Sons, Ltd.     

Prenatal diagnosis of an isochromosome 5p in a fetus with increased nuchal translucency thickness and pulmonary atresia with hypoplastic right heart at 14 weeks

We report on a fetus presenting with increased nuchal translucency at 11 weeks' gestation, suggesting cystic hygroma. Chorion villous sampling was performed, and cytogenetic analysis revealed a supernumerary isochromosome 5p leading to tetrasomy 5p: 47,XX,+ i(5p)[7]/46,XX[5] after short-term culture and 47,XX,+ i(5p)[20] after long-term culture. Subsequent targeted sonographic follow-up at 12 and 14 weeks revealed further increase of the NT to 6.4 mm and the additional presence of a congenital heart defect (pulmonary atresia with intact ventricular septum). Termination of pregnancy was performed, and the heart defect was confirmed. Isochromosome 5p was found in varying proportions in all examined organs. Only a few cases of mosaic tetrasomy 5p have been reported in the literature, and recent reports on prenatally detected isochromosome 5p showed a possible relationship to increased nuchal translucency in some cases and also a possible role of confined mosaicism in others. Whereas cases with confined mosaicism did not show suspicious signs on ultrasound, true mosaicism conversely showed increased nuchal thickness as well as structural abnormalities. This is the first report on the association of a cardiac defect with this chromosome aberration. Copyright © 2004 John Wiley & Sons, Ltd.     

Neurobiology of learning and memory: The honeybee as a model system

Conditioning experiments with bees demonstrate surprising similarities between the well studied laboratory mammals and honeybees. These similarities appear both on the behavioral level and the neural organization of memory. The paper speculates on the basis of these similarities, and the idea of common neural elements is favored.     

融合SSAE深度特征学习和LSTM网络的PM2.5小时浓度预测

精准的PM_2.5小时浓度短期预测,可以有效地提高空气污染的预报预警能力.针对传统的PM_2.5预测模型中存在的影响因素考虑不全面且影响因素选择方法适用性不强等问题,本文提出一种融合栈式稀疏自编码器（Stack Sparse Auto-Encoder,SSAE）和长短期记忆神经网络（Long-Short Term Memory,LSTM）的PM_2.5小时浓度预测模型.SSAE-LSTM模型综合考虑了时间因素、空间因素、气象因素和空气污染物因素等多种因素对PM_2.5的影响,采用SSAE以无监督方式自动提取PM_2.5抽象影响特征,实现特征的压缩和降维;然后以提取的抽象特征作为LSTM模型的输入,建立PM_2.5时间序列预测模型,挖掘PM_2.5历史序列中的长期依赖特征.为了验证方法的有效性,本文基于2016—2018年京津冀城市群71个空气监测站点的空气数据和气象数据,建立SSAE-LSTM模型对各个站点的PM_2.5浓度进行离线训练和预测实验.预测结果表明,SSAE-LSTM模型预测精度高于其它预测模型,在所有测试集上的一致性指数（IA）高达0.99,均方根误差RMSE与平均绝对误差MAE降到了13.98和7.90.此外,分析了SSAE-LSTM模型在不同季节的适用性,71个空气监测站点在春、夏、秋、冬4个季节测试集的预测值和实测值均有很好的线性关系,决定系数分别是0.86、0.92、0.96、0.93.对北京市万寿西宫站点的预测结果表明,SSAE-LSTM模型可以用于不同空气质量情况下的PM_2.5小时浓度预报,且具有应用上的可行性和可靠性.