Neonatal brain MRI and cranial ultrasound (Passo Fundo)

This retrospective single-centre cohort comes from a tertiary referral hospital for high-risk pregnancy in southern Brazil, with authors from Universidade Federal da Fronteira Sul, Hospital São Vicente de Paulo and Universidade de Passo Fundo in Passo Fundo. Newborns who were born at the hospital or transferred there within their first 30 days, and who had brain MRI with a neonatal protocol while hospitalized between January 2016 and March 2019, were reviewed. The study compared the last transfontanellar ultrasound before MRI with the MRI findings and described maternal and neonatal risk factors. The images are not shared.

Composition

51 neonates (25 female, 26 male) after one exclusion. Most were preterm: 21 born before 28 weeks, 18 at 28 to under 32 weeks, 9 at 32 to under 37 weeks and 3 at term. Mean gestational age at birth was about 29 weeks, mean birth weight about 1,275 g, and 22 weighed under 1,000 g. There were nine sets of twins. The paper reports maternal age, prenatal care, Apgar scores, length of stay and gestational and neonatal complications such as premature labor, sepsis and resuscitation.

Acquisition

All infants were scanned under anesthesia, 35 at 1.5 T and 16 at 3 T. The neonatal protocol had sagittal and axial T1-weighted, axial T2-weighted, FLAIR, susceptibility-weighted and diffusion-weighted images with ADC maps. Scanner vendors are not named. Ultrasound was done through the fontanelle as routine screening; on average MRI followed 19.5 days later (range 0 to 63).

Annotations

A neuroradiologist reviewed the MRI images, and ultrasound findings were taken from the radiology reports. Both were classified for hemorrhage with Papile grade, leukoencephalomalacia, hydrocephalus and other findings. MRI showed an alteration in 30 infants and hemorrhage in 22 (germinal matrix grade I in 7, grade II in 10, grade IV in 2, intraparenchymal in 3); ultrasound showed an alteration in 16.

Known limitations

• Images are not available. The article has no data availability statement, and its figshare record holds only the article's figures and tables.
• Small, single-centre, retrospective sample scanned on a mix of 1.5 T and 3 T systems.
• The interval between ultrasound and MRI varied widely, so lesions may have changed between the two exams.
• Ultrasound findings come from reports, not from a re-read of the images.