ORIGINAL ARTICLE Isolated mild white matter signal changes in preterm infants: a regional approach for comparison of cranial ultrasound and MRI findings M Weinstein 1,2 , D Ben Bashat 1,3,4 , V Gross-Tsur 5 , Y Leitner 3,6 , I Berger 3,7 , R Marom 3,7 , R Geva 2 , S Uliel 6 and L Ben-Sira 3,8 OBJECTIVE: To compare echogenicity detected using cranial ultrasound (cUS) and diffuse excessive high signal intensity (DEHSI) detected using magnetic resonance imaging (MRI) by identical region-based scoring criteria in preterm infants. To explore the association between these white matter (WM) signal changes with early neurobehavior. STUDY DESIGN: Forty-nine pre-selected premature infants with only echogenicity on a first routine cUS1 underwent MRI and a repeated cUS2 at term equivalent age. Echogenicity and DEHSI were graded in various brain areas and diffusivity values were calculated. Neurobehavior was assessed using the Rapid Neonatal Neurobehavioral Assessment Procedure. RESULT: WM signal changes were significantly higher on cUS1 than cUS2; and higher in MRI than cUS2 in posterior regions. Infants with DEHSI demonstrated reduced tissue integrity. Imaging findings were not correlated with early neurobehavior. CONCLUSION: Echogenicity and DEHSI likely represent the same phenomenon. Reduction of over-interpretation of WM signal changes may help define criteria for the judicious use of imaging in routine follow-up of premature infants. Journal of Perinatology advance online publication, 20 March 2014; doi:10.1038/jp.2014.33 Keywords: echogenicity; DEHSI; cranial ultrasound; MRI; prematurity; DTI INTRODUCTION Cranial ultrasound (cUS) and magnetic resonance imaging (MRI) have been shown to have significant value for the prediction of outcome in preterm infants when substantial brain pathologies, such as intraventricular hemorrhage grade IV, cerebellar hemor- rhages, periventricular leukomalacia and ventriculomegaly, are present. 1–3 However, when solitary diffuse white matter (WM) signal changes are detected, with no additional imaging findings, a dilemma exists regarding the radiological interpretation of normal and abnormal signal changes. Particular controversy surrounds the interpretation of mild WM signal changes, as the clinical significance of these signal changes is poorly understood. Echogenicity detected using cUS and diffuse excessive high sig- nal intensity (DEHSI) detected using MRI are common WM signal changes that are prevalent in preterm infants. Echogenicity is defined by 'brightness' more intense than the choroid plexus, whereas DEHSI is defined as higher signal intensity in WM than in normal unmyelinated WM on T2-weighted images. However, it is not clear whether echogenicity and DEHSI represent the same phenomenon. Understanding the relationship between these two frequent WM signal changes, defined by different modalities, may help us describe this phenomenon as it is not clear whether WM signal changes represent level of maturation or are part of a continuum of WM injury. 4 Presence of periventricular echogenicities on cUS has been shown to correlate with DEHSI on MRI; 1 however, the absence of periventricular echogenicity did not predict normal WM signal intensity on MRI. 1,5 The discrepancy may be explained by the different modalities (magnetic field vs sonar waves) and differential access to the neonatal brain (in cUS via fontanelle, limiting the angle of view; in MRI no entry point, providing multi- spatial views). Ultrasound is considered to be ‘user dependent’, however, identification of DEHSI on MRI is also somewhat subjective. 6 Few studies directly compared echogenicity and DEHSI on a regional level using the same criteria. 1 Microstructural properties underlying WM signal change in MRI can be assessed using diffusion tensor imaging (DTI). Previous studies have shown that DEHSI is associated with altered water diffusion, such as increased apparent diffusion coefficient and a decrease in fractional anisotropy (FA). 7–12 In the current study, we used DTI in order to understand the microstructural properties of the WM tissue with DEHSI and for validation of the radiological regional assessment of DEHSI. Recent MRI studies of preterm infants did not detect an association between the presence of DEHSI and neurodevelop- mental outcome at 18 and 24 months 9,12,13 and at 9 years; 14 although earlier studies reported lower overall development at 18 and 36 months in preterm infants with DEHSI. 15,16 However, most of these studies did not isolate WM signal changes, rather includ- ing preterm infants with additional brain abnormalities. Further- more, most of these studied did not correlate WM signal changes with neurodevelopmental assessment in the neonatal stage. This study aimed to explore the relationship between cUS and MRI findings by developing and proposing common criteria for regional assessment of WM signal changes, resulting in a radiological score for each infant. We used a common term ‘WM 1 Functional Brain Center, Wohl Institute for Advanced Imaging, Tel-Aviv Sourasky Medical Center (TASMC), Tel-Aviv, Israel; 2 Department of Psychology, Gonda Multidisciplinary Brain Research Center, Bar Ilan University, Ramat-Gan, Israel; 3 Sackler Faculty of Medicine, Tel-Aviv University, Tel-Aviv, Israel; 4 Sagol School of Neuroscience, Tel-Aviv University, Tel-Aviv, Israel; 5 Department of Pediatric Neurology, Shaare-Zedek Medical Center, Jerusalem, Israel; 6 Department of Pediatrics, Child Development Center, TASMC, Tel-Aviv, Israel; 7 Department of Neonatology, Lis Maternity Hospital, TASMC, Tel-Aviv, Israel and 8 Department of Radiology, TASMC, Tel-Aviv, Israel. Correspondence: Dr L Ben-Sira, Department of Radiology, TASMC, 6 Weizman Street, Tel-Aviv 64239, Israel. E-mail: liatb@tlvmc.gov.il Received 19 October 2013; revised 20 January 2014; accepted 27 January 2014 Journal of Perinatology (2014), 1 – 7 © 2014 Nature America, Inc. All rights reserved 0743-8346/14 www.nature.com/jp