|Year : 2017 | Volume
| Issue : 2 | Page : 75-79
Neuroimaging practices in very low birth weight neonates
Arjun Mahendran1, Matthew Engel1, Carolyn R Ahlers-Schmidt1, Debra Desilet-Dobbs2, Barry T Bloom1
1 Department of Pediatrics, The University of Kansas School of Medicine-Wichita, Wichita, KS, USA
2 Department of Radiology, The University of Kansas School of Medicine-Wichita, Wichita, KS, USA
|Date of Web Publication||13-Apr-2017|
Barry T Bloom
550 N. Hillside, Wichita, KS 67214
Source of Support: None, Conflict of Interest: None
Objective: Cranial ultrasound (CUS) and magnetic resonance imaging (MRI) are used to detect brain injury in very low birth weight (VLBW) neonates. Our study attempts to determine the abnormalities found on physician-selected MRI compared to CUS findings and whether any infant characteristic can predict new MRI-detected abnormalities. Methods: Radiology reports of neonates admitted between 2010 and 2014, with birth weights between 501 and 1500 g, were retrospectively reviewed. We excluded infants who died or were transferred and those with a major congenital malformation. Reports were individually reviewed for pertinent positive findings. In addition, various perinatal and maternal characteristics were collected from the electronic medical record. Receipt of MRI and MRI findings were compared with respect to perinatal and maternal characteristics. Results: Out of the 605 VLBW neonates evaluated for the study, 94 infants received MRI. Among the infants with MRI, 43 (46%) had significant findings and 17 (18%) had new significant findings not found by CUS. The MRI did not detect any new intraventricular hemorrhage (IVH) and ventriculomegaly or cystic periventricular leukomalacia (PVL). Infants who had lower gestational age, longer length of stay, lower birth weight, and who had severe IVH, PVL, severe retinopathy of prematurity, or bronchopulmonary dysplasia had significantly higher odds of receiving MRI. Findings from MRI (exclusive or nonexclusive) were not significantly associated with any perinatal characteristic. Conclusions: Although routine use cannot be recommended, physician-selected MRI has a relatively high detection rate of exclusive findings and can potentially increase detection rate.
Keywords: Cranial ultrasound, magnetic resonance imaging, neuroimaging
|How to cite this article:|
Mahendran A, Engel M, Ahlers-Schmidt CR, Desilet-Dobbs D, Bloom BT. Neuroimaging practices in very low birth weight neonates. J Clin Neonatol 2017;6:75-9
|How to cite this URL:|
Mahendran A, Engel M, Ahlers-Schmidt CR, Desilet-Dobbs D, Bloom BT. Neuroimaging practices in very low birth weight neonates. J Clin Neonatol [serial online] 2017 [cited 2017 Jun 25];6:75-9. Available from: http://www.jcnonweb.com/text.asp?2017/6/2/75/204505
| Introduction|| |
Due to advances in neonatal medicine, a subset of premature neonates, infants with birth weight <1500 g (very low birth weight [VLBW]), represent nearly 1.5% of all live births. Neonates, in this weight group, have higher mortality and morbidity rates than neonates with appropriate weights. VLBW neonates are especially prone to cerebral insult secondary to hemorrhage, hypoxia, and ischemia. Neuroimaging techniques, such as cranial ultrasound (CUS) and magnetic resonance imaging (MRI), have long been used to assess these injuries.
The American Academy of Neurology (AAN) published a 2002 practice parameter that advised serial CUS to be the standard neuroimaging screening tool for detecting brain injury in all neonates born <30-week gestational age. The AAN did note that brain MRI in the 1st week of life for full-term infants was more sensitive than CUS in detecting white matter abnormalities (WMA), but there were insufficient data to relate these findings to neurodevelopment prognosis. Thus, no recommendation for routine MRI was given.
More recent studies have suggested MRI around 36–40-week postmenstrual age to be more sensitive than CUS in detecting WMA  and cerebellar lesions  as well as predicting neurodevelopmental outcomes such as cerebral palsy and other neurodevelopmental impairments.,, Near-term MRI has become a popular routine test for all preterm neonates, especially in the advanced level (Level III or higher) Neonatal Intensive Care Unit (NICU), to give prognostic information to parents. Although MRI is a powerful, nonradiating tool to evaluate the preterm brain, it is an expensive and time-consuming technique. The question arises whether every extremely preterm neonate needs a brain MRI at least once during the hospital stay. Plaisier et al. attempted to answer this question by comparing sequential CUS with routine MRI at term equivalent age and showed that while CUS detected most common abnormalities, MRI outperformed in detecting white matter injury and cerebellar hemorrhages. The study shows routine MRI-detected brain abnormalities in 38% infants and CUS in 42% infants.
Our study focuses on determining whether physician-selected MRI affects the detection rate of abnormalities and whether any infant characteristic can predict new MRI abnormalities. Our objective was to evaluate data regarding neuroimaging practices in our regional Level III NICU, which routinely gets serial CUS and selectively obtains MRIs at 36–40-week postmenstrual age. Our study attempts to: (1) identify certain perinatal/neonatal characteristics that are associated with receiving MRI; (2) identify the findings exclusively found in brain MRI and not in serial CUS; and (3) identify perinatal characteristics that would be predictive of exclusive MRI findings.
| Methods|| |
Our NICU is a 62-bed Level III facility that receives approximately 750 admissions/year (60% inborn, 20% outborn, and 20% maternal transfers). VLBW infants admitted to the NICU at this institution between January 2010 and December 2014, were identified using a standardized database. Infants were excluded if they died or were transferred before discharge or if they were diagnosed with any congenital malformation listed by the Vermont Oxford Network. De-identified serial CUS and brain MRI radiological reports that occurred during the ICU stay for each infant were downloaded from the electronic medical record. Relevant perinatal characteristics were extracted from a NICU database (NeoData, Isoprime Corporation; Lisle, IL, USA). The study was approved by the Institutional Review Boards at The University of Kansas School of Medicine-Wichita and the Wichita Medical Research and Education Foundation.
Information extracted included infant demographics (e.g., birth weight, gestational age, sex), clinical characteristics (e.g., Apgar scores, perinatal medications), and presence of comorbidities (e.g., bronchopulmonary dysplasia [BPD], retinopathy of prematurity [ROP]). Maternal characteristics available on infant's medical records, such as mode of delivery, parity, pregnancy complications, and medications, were also extracted. The radiology reports of serial CUS and brain MRI were individually reviewed by a physician, and findings listed under the impression section were collected. Radiological reports were created using a standardized protocol and frequently comment on all aspects of the neonatal brain including ventricle size, presence/absence of intraventricular hemorrhages (IVHs), cerebellar hemorrhage, extra-axial fluid, and hypoxic ischemic injury. MRI data were obtained using a 1.5 Tesla GE Signa Scanner (General Electric Healthcare Technologies, Waukesha, Wisconsin, USA) using a standard pediatric brain protocol. CUS data were obtained using routine neonatal CUS protocol including mastoid views specifically looking at the cerebellum.
Exclusive MRI findings were defined as those only seen on the brain MRI and not on any of the serial CUS. Significant findings were defined as Grade III IVH, periventricular/intraventricular hemorrhages (PIVHs), cystic periventricular leukomalacia (PVL), moderate to severe ventriculomegaly secondary to posthemorrhagic sequela, porencephalic cyst, cerebellar lesions, WMA, and migrational abnormalities. Cerebellar lesions include cerebellar hemorrhages, atrophy, and encephalomalacia. WMA include noncystic PVL, hemorrhages, atrophy, and encephalomalacia.
All data were described using frequencies and percentages for categorical data and means with standard deviations for continuous data. Bivariate analyses were used to determine if any of the infant characteristics were associated with receiving MRI. Distributions of exclusive MRI findings were calculated. Finally, odds ratios were calculated for exclusive MRI findings in relation to certain perinatal characteristics. All bivariate analyses were independent sample t-tests for continuous variables and Chi-square tests for categorical variables. When sample sizes were small, Fisher's exact test was used to test categorical relationships. All analyses were conducted using SPSS version 23 (IBM Corporation, Armonk, NY, USA).
| Results|| |
A dataset of 4533 infants admitted to the NICU between January 2010 and December 2014, was queried to identify 828 infants whose birth weights were between 501 and 1500 g. After exclusion criteria were applied, the remaining 605 infants that had CUS image reports were evaluated for this study [Figure 1]. Of those infants, 94 (14%) had an additional brain MRI report.
Baseline infant characteristics are summarized in [Table 1]. Infants with a report of MRI typically had lower birth weights, lower Apgar scores at 5 min, lower gestational age at birth, and had longer length of stay [Table 2]. The presence of severe IVH or PIVH and PVL was associated with receiving MRI as was the presence of comorbidities such as ROP, BPD, necrotizing enterocolitis, late onset coagulase-negative staphylococcus infections, and fungal infections. Birth weight <1000 g, gestational age at birth <30 weeks, apnea past 35 weeks, presence of seizures, and intubation within 24 h of delivery were also associated with receiving an MRI [Table 3]. However, there was no significant perinatal characteristic that was associated with exclusive or nonexclusive MRI findings.
|Table 3: Odds ratio of receiving or of having new findings detected with magnetic resonance imaging|
Click here to view
Among the 94 infants with reported MRI, 43 (46%) had any significant finding revealed by MRI while 36 (38%) had findings detected by serial CUS. There were 17 (18%) MRIs which revealed exclusive significant findings. CUS detected all grades of IVH as well as all documented episodes of PVL and ventriculomegaly. Exclusive MRI findings included WMA, migration abnormalities, and cerebellar lesions [Figure 2]. Our data also reveal that the near-term MRI detects all the abnormalities normally detected in the late CUS [Table 4].
|Figure 2: Comparison images of cranial ultrasound and magnetic resonance imaging in one infant. (a) Cranial ultrasound performed on a 50-day-old infant using a mastoid view that shows no abnormalities in the cerebellum and (b) an axial T2-weighted gradient brain magnetic resonance imaging performed on the same infant at 90 days showing a small punctuate hemorrhage within the cerebellum (white arrow)|
Click here to view
| Discussion|| |
Our data suggest infants who were more premature, had lower birth weights, and had comorbid conditions were more likely to receive MRI. Furthermore, abnormal neurologic signs (presence of seizures and prolonged apnea) and abnormal CUS findings appear to drive our providers to obtain a brain MRI. Although serial CUSs detected a substantial number of brain lesions, selective brain MRI outperformed CUS in detecting cerebellar lesions, WMA, and migration abnormalities; this is consistent with previous studies.,,,
Our finding that brain MRI can detect all findings normally detected in the near-term CUS suggests that the MRI can replace the near-term CUS to reduce costs. In our institution, despite only 14% of VLBW infants receiving MRI, there was an 18% detection rate of exclusive, significant findings when compared to the standard serial CUS findings. This suggests even selective MRI can detect significant rate of brain abnormalities and can direct future care. In fact, compared to the Plaisier et al. study, in which CUS detects abnormalities at a higher percentage than routine MRIs, our study shows that detection rate is higher in selective MRIs.
Apart from small sample size, other limitations include this study's single site and retrospective nature. Further, at our site, there was no defined protocol for ordering brain MRI; orders were based solely on individual physician discretion. Due to small number of brain MRIs, there was not enough sample size to perform a stratified analysis to isolate each of perinatal characteristics and reduce confounding bias.
Without any study that compares selective MRI with routine MRI, selective MRI should not be considered inappropriate. Our hope is that this data will help create a protocol that directs providers in obtaining selective MRIs. A follow-up study with increased sample size will be needed to study the effects of such a protocol.
| Conclusions|| |
Studies continue to show that brain MRI not only detects significant findings but also predicts neurodevelopmental deficits in the future. Our data show that even selective brain MRI process detects significant brain abnormalities. Furthermore, our data suggest that perinatal characteristics such as gestational age, birth weight, and comorbid conditions can potentially be predictive of new significant MRI findings. Selective MRI could potentially increase the detection rate for significant brain injury not captured by CUS without generating excessive use of screening MRI. However, further studies with larger sample sizes and direct comparison between selective and routine MRI need to be done to confirm this.
We would like to thank Jared Shaw and Paula Delmore for their assistance in data gathering and extraction. We would also like to thank Wichita Medical Research and Education Foundation for their approval of our protocol.
Financial support and sponsorship
This work was funded by a grant from the Wichita Medical Research and Education Foundation.
Conflicts of interest
There are no conflicts of interest.
| References|| |
U.S. Department of Health and Human Services, Health Resources and Services Administration, Maternal and Child Health Bureau. Child Health USA 2014. Rockville, MD, U.S. Department of Health and Human Services; 2014.
Eichenwald EC, Stark AR. Management and outcomes of very low birth weight. N Engl J Med 2008;358:1700-11.
Ment LR, Bada HS, Barnes P, Grant PE, Hirtz D, Papile LA, et al
. Practice parameter: Neuroimaging of the neonate: Report of the Quality Standards Subcommittee of the American Academy of Neurology and the Practice Committee of the Child Neurology Society. Neurology 2002;58:1726-38.
Inder TE, Wells SJ, Mogridge NB, Spencer C, Volpe JJ. Defining the nature of the cerebral abnormalities in the premature infant: A qualitative magnetic resonance imaging study. J Pediatr 2003;143:171-9.
Tam EW, Rosenbluth G, Rogers EE, Ferriero DM, Glidden D, Goldstein RB, et al.
Cerebellar hemorrhage on magnetic resonance imaging in preterm newborns associated with abnormal neurologic outcome. J Pediatr 2011;158:245-50.
Mirmiran M, Barnes PD, Keller K, Constantinou JC, Fleisher BE, Hintz SR, et al.
Neonatal brain magnetic resonance imaging before discharge is better than serial cranial ultrasound in predicting cerebral palsy in very low birth weight preterm infants. Pediatrics 2004;114:992-8.
Spittle AJ, Cheong J, Doyle LW, Roberts G, Lee KJ, Lim J, et al.
Neonatal white matter abnormality predicts childhood motor impairment in very preterm children. Dev Med Child Neurol 2011;53:1000-6.
Woodward LJ, Anderson PJ, Austin NC, Howard K, Inder TE. Neonatal MRI to predict neurodevelopmental outcomes in preterm infants. N Engl J Med 2006;355:685-94.
Smyser CD, Kidokoro H, Inder TE. Magnetic resonance imaging of the brain at term equivalent age in extremely premature neonates: To scan or not to scan? J Paediatr Child Health 2012;48:794-800.
Plaisier A, Raets MM, Ecury-Goossen GM, Govaert P, Feijen-Roon M, Reiss IK, et al.
Serial cranial ultrasonography or early MRI for detecting preterm brain injury? Arch Dis Child Fetal Neonatal Ed 2015;100:F293-300.
Vermont Oxford Network. Manual of Operations: Part 2, Data Definitions & Infant Data Forms. Burlington, VA: Vermont Oxford Network; 2015.
Hintz SR, Barnes PD, Bulas D, Slovis TL, Finer NN, Wrage LA, et al.
Neuroimaging and neurodevelopmental outcome in extremely preterm infants. Pediatrics 2015;135:e32-42.
Miller SP, Cozzio CC, Goldstein RB, Ferriero DM, Partridge JC, Vigneron DB, et al.
Comparing the diagnosis of white matter injury in premature newborns with serial MR imaging and transfontanel ultrasonography findings. AJNR Am J Neuroradiol 2003;24:1661-9.
[Figure 1], [Figure 2]
[Table 1], [Table 2], [Table 3], [Table 4]