1816 BALL AJNR: 18, November 1997 Acknowledgment 7. Seltzer SE, Holman BL, Thrall JH, et al. Academic Radiology in a networked environment. Acad Radiol 1996;3:865–872 I gratefully acknowledge the support of Janet L. Strife, 8. Hanks GE. The effects of health care reform on academic medical my chairperson, who has done much to support the con- centers: 1994 gold medal address. Int J Radiat Oncol Biol Phys cepts contained in this report, who recognizes the need for 1995;31:999–1004 change, and who is optimistic about the eventual success 9. Bragg DG, Hendee WR. Support for biomedical research and its of these proposals. impact on radiology. Radiology 1994;193:599–603 10. Matherlee KR. The outlook for clinical research: impacts of federal funding restraint and private sector reconfiguration. Acad Med 1995;70:1065–1072 References 11. Kirks DR. The critical roles of research and education in health 1. Drake DF. Managed care: a product of market dynamics. JAMA care system reform. Acad Radiol 1994;1:293–294 1997;277:560–563 12. Alazraki N. Clinical practice versus research: how can radiology 2. Hillman BJ. Whose turf is imaging? Independent practice, aca- win? Invest Radiol 1993;28:985–986 demics and research. AJR Am J Roentgenol 1991;156:443–447 13. James AE Jr, Partain CL, Heller RM, Patton JA, Price RR. Can 3. Friedman PJ. The impact of health care reform on academic we administrate academics? Preserving the biomedical re- radiology. Invest Radiol 1994;29:790–796 search enterprise in medical imaging. Invest Radiol 1989;24:815– 4. Holman BL. The changing face of academic radiology: can it 816 survive managed care. Acad Radiol 1995;2:1011–1015 14. Staab EV, Langland-Orban B, Gapenski LG. Survey of overhead 5. Friedenberg RM. Medical education and practice in a new envi- assessments of academic health center radiology departments. ronment. Radiology 1997;202:33A–36A Acad Radiol 1994;1:276–282 6. Gallin JI, Smits HL. Managing the interface between medical 15. Hillman BJ, Fajardo LL, Witzke DB, Cardenas D, Irion M, Fulginiti schools, hospitals, and clinical research. JAMA 1997;277:651– JV. Factors influencing radiologists to choose research careers. 654 Invest Radiol 1989;28:842–848 The Encephalopathic Neonate: Choosing the Proper Imaging Technique A. James Barkovich, University of California, San Francisco The central nervous system (CNS) of the neonate may Although the value of high-quality neuroimaging in the be injured by a number of different mechanisms, including assessment of the neonate who has suffered an insult to birth trauma, hypoxia-ischemia, hypoglycemia, hyperbil- the CNS has been generally well accepted, the choice of irubinemia, inborn errors of metabolism, and neonatal in- neuroimaging study has not. Some radiologists rely pri- fections. Neurologic assessment of affected neonates in- marily on transfontanelle sonography while others advo- cludes evaluation of alertness level, cranial nerve function, cate the use of computed tomography (CT) or magnetic motor function (tone, posture, motility, power, and reflex- resonance (MR) imaging techniques. The purpose of this es), presence of neonatal reflexes (Moro, palmar grasp, report is to review briefly the literature on neonatal neuro- and tonic neck response), and gross sensory examination. imaging and to propose a logical approach. However, because of the immaturity of the CNS, neonatal neurologic assessment tests only the function of the brain stem and basal ganglia. Abnormal findings will alert the Techniques clinician to the fact that the infant has suffered a CNS Sonography injury. The precise cause of injury and the severity, extent, and location of the injury to the cerebral cortex are difficult Because the anterior fontanelle of the neonate is usually to establish on clinical grounds. Neuroimaging plays an large, nearly the entire brain can be seen with transfonta- essential role in the assessment of brain injury in these nelle sonography. The ultrasound machine is portable, so patients by helping to establish the cause of injury and the it can be used at the bedside in the neonatal intensive care expected neurologic outcome. unit and obviates transporting the sick infant. Sonography Address reprint requests to A. James Barkovich, Neuroradiology Section, Room L-371, University of California, San Francisco, 505 Parnassus Ave, San Francisco, CA 94143. Index terms: Brain, diseases; Infants, newborn; Special reports AJNR 18:1816–1820, Nov 1997 0195-6108/97/1810–1816 © American Society of Neuroradiology AJNR: 18, November 1997 CONTROVERSIES 1817 is particularly useful in the imaging of premature neonates, undergoing necrosis followed by cavitation and then who have small brains and unstable circulatory systems. shrinkage of the cavity with resultant focal enlargement of Moreover, the availability of Doppler sonography has the adjacent ventricle (20, 21). PVL should be suspected added a new dimension to the arsenal of the sonographer, when increased echogenicity is present in the periventricu- who can detect altered resistive indexes in neonates who lar regions on sonographic studies; however, edema also have suffered hypoxic-ischemic injury (1–3). causes increased echogenicity and it can resolve without any subsequent brain damage (22). Moreover, increased echogenicity can be seen in this region in the absence of CT PVL or edema (23) and normal sonograms have been CT is more useful in older children than in neonates. reported in infants subsequently proved to have PVL at The main reason for the limited utility in neonates is the autopsy (24, 25). Therefore, the appearance of hyperecho- high water content of the neonatal brain, which reduces genicity in itself is not enough to make a diagnosis of PVL. contrast between normal and injured tissue. CT is least The best early sonographic sign of periventricular white useful in premature neonates, in whom white matter injury matter injury is the periventricular “flare,” a globular area is most common; it is more useful in term neonates, who that has echogenicity equal to or greater than that of the are more likely to have suffered gray matter injury (4–6). choroid plexus. If prolonged periventricular flares are seen, In terms of the difficulty of examination, CT is intermediate the prevalence of spastic diplegia or tetraplegia can be as between sonography and MR imaging. The patient must be high as 50% (26). A definitive diagnosis of periventricular moved to the CT suite, but is easily monitored during the leukomalacia by sonography requires demonstration of examination and standard life-support equipment is easily cavitation and the subsequent formation of periventricular accommodated in the modern CT suite. cysts (27–31). Cavitation occurs 2 to 6 weeks (usually less than 3 weeks) after injury (31). Although it has been reported that patients with large periventricular cysts MR (more than 5 mm in diameter) have a poorer motor out- MR imaging and MR spectroscopy are probably the come than those with smaller cysts (26), it is wise to most sensitive and specific imaging techniques in the ex- remember that the size of the cysts varies over time; they amination of neonates with suspected brain injury. White enlarge as cavitation develops, and then rapidly shrink matter and gray matter injuries can be detected with MR (21). As a result of the rapid and continuous change in cyst imaging in both term and preterm neonates (5, 7–11). In size, it is probably unwise to rely too heavily on this mea- addition, abnormalities detected on MR studies have pre- surement in predicting outcome. Moreover, it is important dictive value for neurodevelopmental outcome (8, 12, 13). to remember that mildly or moderately injured tissue may Proton MR spectroscopy (14, 15) and diffusion MR imag- not cavitate yet can still cause neurologic deficit (32). ing (16) are useful adjuncts to routine MR imaging and can Other causes of neurologic impairment in premature be acquired with only minimal additional imaging time. neonates, such as infection, infarction, and malformation, However, MR imaging requires transporting infants who may be diagnosed with variable confidence at sonogra- are often hemodynamically unstable. Moreover, special phy. Congenital infections are suggested by the presence life-support and monitoring equipment are needed to per- of lenticulostriate vasculopathy (33, 34); however, this form MR imaging safely on unstable neonates (17). finding is nonspecific (33). Large cortical infarctions are easily diagnosed as triangular regions of hyperechogenic- ity; smaller, more posterior infarctions are more difficult to Recommendations detect. Midline malformations are more easily diagnosed than those occurring more laterally. In spite of its limita- Premature Infants tions, the bedside availability of sonography makes it an Premature infants are often hemodynamically unstable; invaluable tool in the diagnosis of conditions requiring therefore, transporting them is somewhat risky. Transfon- rapid intervention, such as hydrocephalus and hemato- tanelle sonography can be performed in the neonatal in- mas. tensive care unit without moving the neonate and is there- Owing to the low sensitivity of sonography in the de- fore the initial imaging study of choice in all premature tection of nonhemorrhagic, noncavitary parenchymal in- neonates with definite or suspected neurologic impair- jury, another imaging study is usually necessary if the ment. Such techniques as color Doppler sonography and neurologic status of a child is worse than can be explained