Neuroimaging and the Prediction of Outcomes in Preterm Infants
http://www.100md.com
《新英格兰医药杂志》
Among the most pressing questions for the parents and caregivers of extremely preterm infants are those about their children's future. Will they survive? If so, what are their chances of leading a life that we consider normal?
In a large study of infants born before the end of the 26th week of gestation, only one in five had no neurodevelopmental limitations at six years of age, and one in five was severely disabled.1 So how should physicians counsel the parents of a child born at 25 weeks of gestation or earlier?
The determination of this prognosis poses many problems. First, the future of the child is influenced by numerous factors, many of which occur after discharge from the hospital.2 Second, recognized clinical predictors of neurodevelopmental outcome, such as the presence or absence of chronic lung disease, are not particularly useful in practice. Third, clinicians who receive the prognostic information have widely varying opinions about what constitutes a normal future — one size does not fit all. Finally, in times of stress, most parents are overwhelmed, unable to make sense of probabilities, and apt to become discouraged.3
One strategy to improve the ability of clinicians to counsel parents is to seek additional prognostic information. Cranial ultrasonography currently provides the best prognostic information available in the neonatal intensive care unit.4 Hemorrhage seen on ultrasonographic images is typically classified according to the methods of Papile et al.5 Hemorrhage classified as grade III ("intraventricular hemorrhage with ventricular enlargement") or grade IV ("intraventricular hemorrhage with parenchymal hemorrhage") is considered to confer a poor prognosis; the two grades are often grouped together to represent "severe intraventricular hemorrhage." The hemorrhage, however, conveys less prognostic information than does ultrasonographic evidence of white-matter damage. Also, these categorizations are limited by the heterogeneity of the findings — and the associated implications — within a given grade.6,7,8 The prognosis for a patient with grade III intraventricular hemorrhage depends on the extent and the cause of the ventriculomegaly.9,10 Mild ventriculomegaly conveys a more favorable prognosis than do moderate and severe ventriculomegaly (which are likely to reflect diffuse white-matter damage resulting in hydrocephalus ex vacuo). Grade IV intraventricular hemorrhage generally conveys an unfavorable prognosis, but this is attributable in most cases not to parenchymal hemorrhage but to substantial white-matter damage.
Although the Papile classification has prognostic utility, its limitations have led some commentators to suggest its elimination and the use of alternative approaches to identify children who are at high risk, on the basis of findings on ultrasonography.6,7,8 For example, the finding of an area of echolucency in the cerebral white matter (the hallmark of what is often called periventricular leukomalacia) is a very good indicator of focal white-matter damage and is probably the best predictor of cerebral palsy available in the neonatal intensive care unit. It is also probably an indicator of the existence, though not the magnitude or location, of diffuse damage.11
Still, the use of ultrasonography to predict the prognosis of the extremely preterm infant leaves much to be desired. Some children considered to be at high risk for developmental dysfunction do much better than expected,12 and some considered to be likely to do well do not.13
As compared with ultrasonography, magnetic resonance imaging (MRI) provides much higher spatial resolution. Consequently, it has the potential to offer considerably better information about the type and magnitude of damage to the brain. Yet few studies have compared ultrasonographic images and early MRI scans.4,11 The reasons for this include the high cost of MRI; the need to modify existing equipment to accommodate a very small infant; the need to have the magnet accessible to the neonatal intensive care unit; for the most fragile neonates, the need for equipment (needles, monitoring equipment, and ventilators) that will not be affected by the magnet; and finally, the need for special training and experience in performing MRI and interpreting the results.
In this issue of the Journal, Woodward et al.14 provide us with a glimpse into the future, when the best possible neuroimaging infrastructure and expertise will improve the ability of clinicians to predict the subsequent function of infants born far before term. The investigators enrolled 167 preterm infants born at or before 30 weeks of gestation and assessed the infants using cranial ultrasonography 4 to 6 weeks after birth, MRI near term equivalent (gestational age of 40 weeks), and a standardized examination protocol at 2 years of age (corrected for prematurity). They found neonatal white-matter abnormalities in about two thirds of infants and gray-matter abnormalities in about half. Moderate-to-severe white-matter abnormalities appeared to be strong predictors of motor dysfunction and cognitive dysfunction and to be more predictive than gray-matter abnormalities. White-matter abnormalities remained significant predictors of some neurodevelopmental outcomes even after findings on cranial ultrasonography were taken into account.
The results of Woodward et al. are intriguing, but they leave unanswered several questions relevant to the extrapolation of these data to clinical practice. Four of the five components of the white-matter score can be assessed on ultrasonographic images; only one component — the nature and extent of white-matter signal abnormality — requires MRI. The extent to which this one feature improves prognostic information remains uncertain and warrants further study.
Another issue relates to their comparison of the findings on MRI with the findings on ultrasonography. The authors claim that MRI is a better predictor of severe cognitive delay than is ultrasonography. This is probably true. On the other hand, because (as is done in many clinical studies) the authors combined Papile grades III and IV to form a single category (reflecting "severe intraventricular hemorrhage"), the inclusion of children who did not have white-matter damage in this single category is likely to have magnified the superiority of MRI over ultrasonography in the prediction of subsequent impairment.
Woodward et al. report the sensitivity and specificity of findings on MRI for developmental outcome. These data provide some perspective but do not answer the more relevant question for a clinician: Given that a child has, or does not have, a particular finding on ultrasonography or MRI — all other things being equal — what are his or her probabilities of having, or not having, an adverse developmental outcome? In this regard, it would be helpful to have information about the positive predictive value (the proportion of children with abnormalities on MRI who later have a given developmental abnormality) and the negative predictive value (the proportion of children without abnormalities on MRI who later do not have the developmental abnormality of interest).
At present, we consider it premature to use the results of Woodward et al. to provide support for the more routine use of MRI in risk stratification, at least for clinical purposes. On the other hand, the use of MRI to better assess central nervous system abnormalities in preterm infants at term equivalent might increase the probability that children assigned to different treatment groups in a clinical trial would have a similar risk of a given developmental outcome.
Given the limitations in our ability to predict prognosis, what are neonatologists and neurologists to do? Acknowledging this inherent uncertainty, some physicians offer the equivalent of, "Take them home, give them lots of love, and let's see how things go." This approach, which lets parents know that problems might lie ahead, but does not take away hope, still has merit.
No potential conflict of interest relevant to this article was reported.
Source Information
From the Perinatal Infectious Disease Epidemiology Unit, Departments of Gynecology and Obstetrics and Pediatric Pulmonology and Neonatology, Hannover Medical School, Hannover, Germany (O.D.); and the Neuroepidemiology Unit, Department of Neurology, Children's Hospital and Harvard Medical School, Boston (A.L.).
References
Marlow N, Wolke D, Bracewell MA, Samara M. Neurologic and developmental disability at six years of age after extremely preterm birth. N Engl J Med 2005;352:9-19.
Vohr BR, Allan WC, Westerveld M, et al. School-age outcomes of very low birth weight infants in the indomethacin intraventricular hemorrhage prevention trial. Pediatrics 2003;111:e340-e346.
Thorne S, Hislop TG, Kuo M, Armstrong E-A. Hope and probability: patient perspectives of the meaning of numerical information in cancer communication. Qual Health Res 2006;16:318-336.
O'Shea TM, Counsell SJ, Bartels DB, Dammann O. Magnetic resonance and ultrasound brain imaging in preterm infants. Early Hum Dev 2005;81:263-271.
Papile LA, Burstein J, Burstein R, Koffler H. Incidence and evolution of subependymal and intraventricular hemorrhage: a study of infants with birth weights less than 1,500 gm. J Pediatr 1978;92:529-534.
Kuban K, Teele RL. Rationale for grading intracranial hemorrhage in premature infants. Pediatrics 1984;74:358-363.
Paneth N. Classifying brain damage in preterm infants. J Pediatr 1999;134:527-529.
Vasileiadis GT. Grading intraventricular hemorrhage with no grades. Pediatrics 2004;113:930-931.
Leviton A, Gilles F. Ventriculomegaly, delayed myelination, white matter hypoplasia, and "periventricular" leukomalacia: how are they related? Pediatr Neurol 1996;15:127-136.
Vollmer B, Roth S, Riley K, et al. Neurodevelopmental outcome of preterm infants with ventricular dilatation with and without associated haemorrhage. Dev Med Child Neurol 2006;48:348-352.
Holling EE, Leviton A. Characteristics of cranial ultrasound white-matter echolucencies that predict disability: a review. Dev Med Child Neurol 1999;41:136-139.
Msall ME. Developmental vulnerability and resilience in extremely preterm infants. JAMA 2004;292:2399-2401.
Laptook AR, O'Shea TM, Shankaran S, Bhaskar B, NICHD Neonatal Network. Adverse neurodevelopmental outcomes among extremely low birth weight infants with a normal head ultrasound: prevalence and antecedents. Pediatrics 2005;115:673-680.
Woodward LJ, Anderson PJ, Austin NC, Howard K, Inder TE. Neonatal MRI to predict neurodevelopmental outcomes in preterm infants. N Engl J Med 2006;355:685-694.(Olaf Dammann, M.D., and A)
In a large study of infants born before the end of the 26th week of gestation, only one in five had no neurodevelopmental limitations at six years of age, and one in five was severely disabled.1 So how should physicians counsel the parents of a child born at 25 weeks of gestation or earlier?
The determination of this prognosis poses many problems. First, the future of the child is influenced by numerous factors, many of which occur after discharge from the hospital.2 Second, recognized clinical predictors of neurodevelopmental outcome, such as the presence or absence of chronic lung disease, are not particularly useful in practice. Third, clinicians who receive the prognostic information have widely varying opinions about what constitutes a normal future — one size does not fit all. Finally, in times of stress, most parents are overwhelmed, unable to make sense of probabilities, and apt to become discouraged.3
One strategy to improve the ability of clinicians to counsel parents is to seek additional prognostic information. Cranial ultrasonography currently provides the best prognostic information available in the neonatal intensive care unit.4 Hemorrhage seen on ultrasonographic images is typically classified according to the methods of Papile et al.5 Hemorrhage classified as grade III ("intraventricular hemorrhage with ventricular enlargement") or grade IV ("intraventricular hemorrhage with parenchymal hemorrhage") is considered to confer a poor prognosis; the two grades are often grouped together to represent "severe intraventricular hemorrhage." The hemorrhage, however, conveys less prognostic information than does ultrasonographic evidence of white-matter damage. Also, these categorizations are limited by the heterogeneity of the findings — and the associated implications — within a given grade.6,7,8 The prognosis for a patient with grade III intraventricular hemorrhage depends on the extent and the cause of the ventriculomegaly.9,10 Mild ventriculomegaly conveys a more favorable prognosis than do moderate and severe ventriculomegaly (which are likely to reflect diffuse white-matter damage resulting in hydrocephalus ex vacuo). Grade IV intraventricular hemorrhage generally conveys an unfavorable prognosis, but this is attributable in most cases not to parenchymal hemorrhage but to substantial white-matter damage.
Although the Papile classification has prognostic utility, its limitations have led some commentators to suggest its elimination and the use of alternative approaches to identify children who are at high risk, on the basis of findings on ultrasonography.6,7,8 For example, the finding of an area of echolucency in the cerebral white matter (the hallmark of what is often called periventricular leukomalacia) is a very good indicator of focal white-matter damage and is probably the best predictor of cerebral palsy available in the neonatal intensive care unit. It is also probably an indicator of the existence, though not the magnitude or location, of diffuse damage.11
Still, the use of ultrasonography to predict the prognosis of the extremely preterm infant leaves much to be desired. Some children considered to be at high risk for developmental dysfunction do much better than expected,12 and some considered to be likely to do well do not.13
As compared with ultrasonography, magnetic resonance imaging (MRI) provides much higher spatial resolution. Consequently, it has the potential to offer considerably better information about the type and magnitude of damage to the brain. Yet few studies have compared ultrasonographic images and early MRI scans.4,11 The reasons for this include the high cost of MRI; the need to modify existing equipment to accommodate a very small infant; the need to have the magnet accessible to the neonatal intensive care unit; for the most fragile neonates, the need for equipment (needles, monitoring equipment, and ventilators) that will not be affected by the magnet; and finally, the need for special training and experience in performing MRI and interpreting the results.
In this issue of the Journal, Woodward et al.14 provide us with a glimpse into the future, when the best possible neuroimaging infrastructure and expertise will improve the ability of clinicians to predict the subsequent function of infants born far before term. The investigators enrolled 167 preterm infants born at or before 30 weeks of gestation and assessed the infants using cranial ultrasonography 4 to 6 weeks after birth, MRI near term equivalent (gestational age of 40 weeks), and a standardized examination protocol at 2 years of age (corrected for prematurity). They found neonatal white-matter abnormalities in about two thirds of infants and gray-matter abnormalities in about half. Moderate-to-severe white-matter abnormalities appeared to be strong predictors of motor dysfunction and cognitive dysfunction and to be more predictive than gray-matter abnormalities. White-matter abnormalities remained significant predictors of some neurodevelopmental outcomes even after findings on cranial ultrasonography were taken into account.
The results of Woodward et al. are intriguing, but they leave unanswered several questions relevant to the extrapolation of these data to clinical practice. Four of the five components of the white-matter score can be assessed on ultrasonographic images; only one component — the nature and extent of white-matter signal abnormality — requires MRI. The extent to which this one feature improves prognostic information remains uncertain and warrants further study.
Another issue relates to their comparison of the findings on MRI with the findings on ultrasonography. The authors claim that MRI is a better predictor of severe cognitive delay than is ultrasonography. This is probably true. On the other hand, because (as is done in many clinical studies) the authors combined Papile grades III and IV to form a single category (reflecting "severe intraventricular hemorrhage"), the inclusion of children who did not have white-matter damage in this single category is likely to have magnified the superiority of MRI over ultrasonography in the prediction of subsequent impairment.
Woodward et al. report the sensitivity and specificity of findings on MRI for developmental outcome. These data provide some perspective but do not answer the more relevant question for a clinician: Given that a child has, or does not have, a particular finding on ultrasonography or MRI — all other things being equal — what are his or her probabilities of having, or not having, an adverse developmental outcome? In this regard, it would be helpful to have information about the positive predictive value (the proportion of children with abnormalities on MRI who later have a given developmental abnormality) and the negative predictive value (the proportion of children without abnormalities on MRI who later do not have the developmental abnormality of interest).
At present, we consider it premature to use the results of Woodward et al. to provide support for the more routine use of MRI in risk stratification, at least for clinical purposes. On the other hand, the use of MRI to better assess central nervous system abnormalities in preterm infants at term equivalent might increase the probability that children assigned to different treatment groups in a clinical trial would have a similar risk of a given developmental outcome.
Given the limitations in our ability to predict prognosis, what are neonatologists and neurologists to do? Acknowledging this inherent uncertainty, some physicians offer the equivalent of, "Take them home, give them lots of love, and let's see how things go." This approach, which lets parents know that problems might lie ahead, but does not take away hope, still has merit.
No potential conflict of interest relevant to this article was reported.
Source Information
From the Perinatal Infectious Disease Epidemiology Unit, Departments of Gynecology and Obstetrics and Pediatric Pulmonology and Neonatology, Hannover Medical School, Hannover, Germany (O.D.); and the Neuroepidemiology Unit, Department of Neurology, Children's Hospital and Harvard Medical School, Boston (A.L.).
References
Marlow N, Wolke D, Bracewell MA, Samara M. Neurologic and developmental disability at six years of age after extremely preterm birth. N Engl J Med 2005;352:9-19.
Vohr BR, Allan WC, Westerveld M, et al. School-age outcomes of very low birth weight infants in the indomethacin intraventricular hemorrhage prevention trial. Pediatrics 2003;111:e340-e346.
Thorne S, Hislop TG, Kuo M, Armstrong E-A. Hope and probability: patient perspectives of the meaning of numerical information in cancer communication. Qual Health Res 2006;16:318-336.
O'Shea TM, Counsell SJ, Bartels DB, Dammann O. Magnetic resonance and ultrasound brain imaging in preterm infants. Early Hum Dev 2005;81:263-271.
Papile LA, Burstein J, Burstein R, Koffler H. Incidence and evolution of subependymal and intraventricular hemorrhage: a study of infants with birth weights less than 1,500 gm. J Pediatr 1978;92:529-534.
Kuban K, Teele RL. Rationale for grading intracranial hemorrhage in premature infants. Pediatrics 1984;74:358-363.
Paneth N. Classifying brain damage in preterm infants. J Pediatr 1999;134:527-529.
Vasileiadis GT. Grading intraventricular hemorrhage with no grades. Pediatrics 2004;113:930-931.
Leviton A, Gilles F. Ventriculomegaly, delayed myelination, white matter hypoplasia, and "periventricular" leukomalacia: how are they related? Pediatr Neurol 1996;15:127-136.
Vollmer B, Roth S, Riley K, et al. Neurodevelopmental outcome of preterm infants with ventricular dilatation with and without associated haemorrhage. Dev Med Child Neurol 2006;48:348-352.
Holling EE, Leviton A. Characteristics of cranial ultrasound white-matter echolucencies that predict disability: a review. Dev Med Child Neurol 1999;41:136-139.
Msall ME. Developmental vulnerability and resilience in extremely preterm infants. JAMA 2004;292:2399-2401.
Laptook AR, O'Shea TM, Shankaran S, Bhaskar B, NICHD Neonatal Network. Adverse neurodevelopmental outcomes among extremely low birth weight infants with a normal head ultrasound: prevalence and antecedents. Pediatrics 2005;115:673-680.
Woodward LJ, Anderson PJ, Austin NC, Howard K, Inder TE. Neonatal MRI to predict neurodevelopmental outcomes in preterm infants. N Engl J Med 2006;355:685-694.(Olaf Dammann, M.D., and A)