Evaluation of Fetal Lung Ultrasound Images by Digital Texture Analysis Methods


YILDIZ U. G. , SEZİK M. , Cetiner H. , Aydogan T.

JOURNAL OF CLINICAL AND ANALYTICAL MEDICINE, cilt.6, ss.479-484, 2015 (ESCI İndekslerine Giren Dergi) identifier

  • Cilt numarası: 6
  • Basım Tarihi: 2015
  • Doi Numarası: 10.4328/jcam.3636
  • Dergi Adı: JOURNAL OF CLINICAL AND ANALYTICAL MEDICINE
  • Sayfa Sayıları: ss.479-484

Özet

Aim: Evaluation of fetal lung maturity in preterm pregnancies without requirement for an invasive procedure such as amniocentesis is of importance. The aim of the present study was to extract numerical features from fetal pulmonary ultrasound images, using computerized texture analysis methods. Material and Method: Twenty fetal ultrasound images from 18 pregnancies that were followed up in our department for threatened preterm delivery between 24-37 weeks of gestational age were included before corticosteroid administration. Transverse sections including well-defined visualization of bilateral fetal lungs without artifacts were evaluated. Regions of interests (ROIs) with a 64x64 pixel area and homogenous pulmonary tissue were selected. Images were analyzed with invariant moments (IM), grey level co-occurrence matrix (GLCM), and wavelet analysis (WA) using MATLAB R2014a computer software. Results: The mean gestational age was 30.9 +/- 3.2 weeks. A total of 159 features were extracted from the ROIs of each image. Therefore, fetal ultrasound images were coded into numerical values, using advanced texture analysis techniques. Discussion: Assessment of ultrasound images from fetal lungs at different gestational ages was feasible with the introduced digital tissue analysis algorithm. Non-invasive evaluation of fetal lung maturity will subsequently be investigated in line with the defined procedure.