Adaptive mechanisms of developing brain. The neuroradiologic assessment of the preterm infant.

Since the 1980s, cranial sonography has been routinely performed in premature infants. This has produced a wealth of information about the more dramatic central nervous system lesions of IVH, PVL, and late VM. This information has included timing and evolution of these lesions and their eventual correlation with outcome. For two reasons the advent of MR imaging scanning has produced an interest in using this modality to evaluate these same infants. First, MR imaging gives an obviously superior image, and its ability to detect lesions is far superior to that of ultrasound. Second, the ability of cranial sonography to detect all of the children with CP or low IQ is limited. In our studies of outcome in very low-birth weight infants grade 3 to 4 IVH, PVL, or VM are able to detect only about 50% of the infants who developed CP by 3 years. This condition should be highly correlated with structural brain disease; an imaging modality that was more sensitive to central nervous system lesions should offer an advantage in predicting outcome. In the only prospective assessment of the ability of these two modalities to predict outcome at 3 years, van de Bor and colleagues found MR imaging did not do better than cranial sonography. This was largely because both modalities detected the most severe lesions, and most children with milder lesions on MR imaging had normal outcome. Studies of late (age 1 to teenage years) MR imaging scans in preterm infants show that a high percentage have white matter lesions but these lesions correlate poorly with outcome. If our concern when counseling parents is to alert them when a serious adverse outcome is likely in their child, then cranial sonography is to be favored precisely because it is less able to detect subtle lesions, which the developing brain has the capacity to overcome. On the other hand, if our aim is to detect all lesions, even though these lesions do not predict serious adverse outcomes, then MR imaging is to be favored. Research aimed at discovering etiologies and mechanisms of brain injury in these high-risk infants should use the more sensitive modality MR imaging. Finally, the interesting observation that preterm infants fare as well as they do despite MR imaging-identified lesions might stimulate research studying the adaptive mechanisms of developing brain.