An investigation into the feasibility of fetal lung maturity prediction using statistical textural features

Fetal lung and liver tissues were examined by ultrasound in 240 subjects du ring 24 to 38 weeks of gestational age in order to investigate the feasibil ity of predicting the maturity of the lung from the textural features of so nograms. A region of interest of 64 X 64 pixels is used for extracting text ural features. Since the histological properties of the liver are claimed t o remain constant with respect to gestational age, features obtained from t he lung region are compared with those from liver. Though the mean values o f some of the features show a specific trend with respect to gestation age, the variance is too high to guarantee definite prediction of the gestation al age. Thus, we restricted our purview to an investigation into the feasib ility of fetal lung maturity prediction using statistical textural features . Out of 64 features extracted, those features that are correlated with ges tation age and less computationally intensive are selected. The results of our study show that the sonographic features hold some promise in determini ng whether the fetal lung is mature or immature.