作者: E. Bonet-Carne , M. Palacio , T. Cobo , A. Perez-Moreno , M. Lopez
DOI: 10.1002/UOG.13441
关键词:
摘要: Objective To develop and evaluate the performance of a novel method for predicting neonatal respiratory morbidity based on quantitative analysis fetal lung by ultrasound. Methods More than 13 000 non-clinical images 900 were used to computerized texture machine learning algorithms, trained predict risk ultrasound images. The method, termed ‘quantitative maturity analysis’ (quantusFLM™), was then validated blindly in 144 neonates, delivered at 28 + 0 39 + 0 weeks' gestation. Lung DICOM format obtained within 48 h delivery ability software morbidity, defined as either distress syndrome or transient tachypnea newborn, determined. Results Mean (SD) gestational age 36 + 1 (3 + 3) weeks. Among there 29 (20.1%) cases morbidity. Quantitative predicted with sensitivity, specificity, positive predictive value negative 86.2%, 87.0%, 62.5% 96.2%, respectively. Conclusions Quantitative an accuracy comparable that current tests using amniotic fluid. Copyright © 2014 ISUOG. Published John Wiley & Sons Ltd.