Sun Kyoung You MD, Presenter: Nothing to Disclose

Young-Hun Choi, Abstract Co-Author: Nothing to Disclose

Sang-Joon Park BA, Abstract Co-Author: Nothing to Disclose

Jung-Eun Cheon MD, Abstract Co-Author: Nothing to Disclose

Woo Sun Kim MD, Abstract Co-Author: Nothing to Disclose

In-One Kim MD, Abstract Co-Author: Nothing to Disclose

Periventricular leukomalacia(PVL) is the major cause of neurodevelopmental problems encountered in survivors of premature birth. The aim of this study was to evaluate the ultrasound texture analysis as a potential imaging tool for quantitative assessment of periventricular white matter (PVWM) injury in preterm infants.

73 preterm infants (median gestational age; 28wks, median birth weight; 905.2g) who were treated in the neonatal intensive care unit and had serial cranial US and brain MR obtained at near term period (38.0 ±3.7wks) were included in our study. Periventricular echogenicity (PVE) on serial cranial US were evaluated qualitatively by visual grading as grade 1, 2, and 3 compared to echogenicity of the choroid plexus and quantitatively using Gray Level Co-ocurrence Matrix (GLCM) method; a second order statistical method of texture analysis. These features were selected based on empirical observation that the normal PVWM exhibits homogeneous echotexture, whereas the ischemic PVWM often exhibits heterogeneous echotexture. Four GLCM textural features which represent homogeneity of an image was measured: (1) angular second moment (ASM), (2) inverse differential moment (IDM), (3)contrast, and (4) entropy. Quantitative and qualitative US features of PVE were compared between two groups (group 1: PVL on MRI(10), group 2; no PVWM abnormality on MRI(63)).

All GLCM features did not show statistically significant difference between two groups, although mean value of ASM and IDM were higher in group 1 than those of group 2 (ASM p-value ; 0.176, IDM p-value ; 0.52). Contrast (degree of heterogeneity) and entropy (degree of randomness) were lower in group 1 than those of group 2 (contrast p-value; 0.68, entropy p-value; 0.221). IDM and entropy values of PVE grade 1 (less than choroid plexus) showed statistically significant difference between two groups. (IDM p-value ; 0.006, entropy p-value; <0.05)

Texture analysis using GLCM matrix method may serve as a complementary tool for quantitative assessment of PVE in selected cased of increased PVE. It is still questionable whether US screening of PVWM in premature infants can be a sensitive predictor of the PVWM injury in preterm infants.

Off-line texture analysis of PVE utilizing GLCM matrix might be a complementary tool for quantitative approach to vulnerable brain structures in preterm infants

You, S, Choi, Y, Park, S, Cheon, J, Kim, W, Kim, I, Does Ultrasound Texture Analysis of Periventricular White Matter Predict the Periventricular White Matter Injury in Preterm Infants?.  Radiological Society of North America 2013 Scientific Assembly and Annual Meeting, December 1 - December 6, 2013 ,Chicago IL. http://archive.rsna.org/2013/13014790.html