Texture analysis of deep medullary veins on susceptibility-weighted imaging in infants: evaluating developmental and ischemic changes

Susceptibility-weighted imaging (SWI) can be used to evaluate deep medullary veins (DMVs). This study aimed to apply texture analysis on SWI to evaluate developmental and ischemic changes of DMV in infants. A total of 38 infants with normal brain MRI (preterm [n = 12], term-equivalent age [TEA] [n = 18], and term [n = 8]) and seven infants with ischemic injury (preterm [n = 2], TEA [n = 1], and term [n = 4]) were included. Regions of interests were manually drawn to include DMVs. First-order texture parameters including entropy, skewness, and kurtosis were derived from SWI. The parameters were compared between groups according to age and presence of ischemic injury. A regression analysis was performed to correlate postmenstrual age (PMA) and parameters. A ROC analysis was performed to differentiate ischemic infants from normal infants. Among parameters, entropy showed a significant difference between the age groups (preterm vs. TEA vs. term; 5.395 vs. 4.885 vs. 4.883, p = 0.001). There was a significant positive relationship between PMA and entropy (R square = 0.402, p < 0.001). Skewness was significantly higher in the ischemic group compared with that in the normal group (1.37 vs. 0.70, p = 0.001). The ROC on skewness resulted in an AUC of 0.87 (accuracy, 83.2%) for differentiating infants with ischemic injury. A texture analysis of DMVs on SWI showed differences according to age and presence of ischemic injury. The texture parameters can potentially be used as quantitative markers for differentiating infants with ischemic injury through DMV changes. • The DMV structure of the infant brain could be quantified on SWI with texture analysis. • Entropy from texture analysis on SWI increased as infants got older. • Normal and ischemic injured infants could be differentiated with a cutoff value of 1.025 for skewness.