Pediatr Radiol (2015) 45 (Suppl 3):S406–S412 DOI 10.1007/s00247-015-3357-1
ADVANCES IN PEDIATRIC NEURORADIOLOGY
Neurosonography: in pursuit of an optimized examination Alan Daneman 1 & Monica Epelman 2,3,4
Received: 8 March 2015 / Accepted: 1 April 2015 # Springer-Verlag Berlin Heidelberg 2015
Abstract This article emphasizes technical factors that are helpful for optimizing sonographic examinations of the brain in preterm and term neonates. It also reviews existing data regarding the accuracy of neurosonographic examinations relative to MR. Many neurosonographic signs are subtle and can be easily overlooked, which could lead to delayed diagnosis or misdiagnosis. Therefore meticulous attention to sonographic technique is required to maximize the diagnostic utility of this modality. Although MR images of the brain often depict abnormalities more clearly than sonography, neurosonography continues to be an exceptionally valuable tool for evaluation of the neonatal brain, even in full-term neonates. Furthermore, its accuracy is greater, even in the latter age group, than many older publications suggest. Prospective studies using state-of-the-art equipment comparing findings on sonography and MR are needed to better understand how the accuracy of these modalities changes with refinements in equipment and to help us better understand the role of neurosonography relative to MR.
Keywords Brain . Neonates . Neurosonography . Optimization . Ultrasound
Introduction This article emphasizes the requirements for obtaining the maximum information from a sonographic (US) examination of the brain in neonates. The article highlights the need for individuals performing US examinations to understand what lesions they are looking for and urges them to develop an aggressive approach to optimizing the US techniques to best depict the normal structures and abnormalities. This article does not discuss the pathogenesis of abnormalities encountered in the neonatal brain and does not illustrate all of these lesions; they have been described in detail in many articles, to which the reader is referred [1–5].
Background
* Alan Daneman
[email protected] 1
Department of Medical Imaging, University of Toronto, Hospital for Sick Children, 555 University Ave., Toronto, ON, Canada M5G 1X8
2
Department of Radiology, University of Central Florida, Orlando, FL, USA
3
Florida State University, Tallahassee, FL, USA
4
Department of Medical Imaging, Nemours Children’s Hospital, Orlando, FL, USA
Sonographic examination of the neonatal brain continues to be the first-line imaging modality in preterm and fullterm neonates and is commonly performed worldwide. The well-known strengths of US are that it is noninvasive, costs less than MR and can be performed at bedside, allowing examinations to be done without moving the neonate. Despite these advantages, MR remains the gold standard for neuroimaging at all ages, and this includes neonates. Until recently the literature comparing the accuracy of US and MR examinations of the neonatal brain has suggested that US is a much less accurate modality compared to MR [6–14]. There are, however, several deficiencies in many of the earlier reports [3]. Several of
Pediatr Radiol (2015) 45 (Suppl 3):S406–S412
S407
large or gross abnormalities are depicted. However, Dinan et al. [5] have highlighted the fact that many US signs are subtle and can be easily overlooked, which could lead to delayed diagnosis or misdiagnosis. Therefore, it is essential to extract as much information from the US examination as possible. To achieve this one has to use state-of-the-art equipment and perform the examination meticulously, aggressively and flexibly. Furthermore, one has to have a thorough knowledge of normal neuroanatomy and pathology in order to understand exactly what one is searching for and to interpret the findings correctly. Finally, an understanding of the accuracy of interpretation of US images in this age group is essential.
these publications have dealt with imaging of the preterm brain only, and fewer publications have evaluated findings in the term brain. Furthermore, in the earlier studies older US equipment was used and the images obtained cannot compare to images that are obtained with the state-of-the-art US equipment now available. Current transducers can depict findings much better than older ones because newer transducers can function at more variable and higher frequencies. Furthermore, newer transducers can be focused to depict the near, mid and far fields to define normal anatomy and pathological lesions to greatest advantage. Additionally, in several reports there has been a long time interval (even more than a day) between the time of US and MR examinations in individual patients, and it is important to remember that even in a short space of time many significant events can affect the brain and alter findings on US and MR. In older reports US examinations have often been evaluated retrospectively, and US images are either unpublished or are suboptimal in contrast to the published MR images, which are usually of high quality. These reports often provide limited or no information regarding the US techniques used in contrast to the very detailed information provided regarding the MR techniques. None of the older articles has described the need for a flexible, aggressive approach to the US technique, as we have described previously [1–5] and which we again describe below in detail. Another problem related to some neurosonography papers and textbooks is that the descriptions of the abnormalities of the neonatal brain are suboptimal or deficient. If one simply accepts the findings in these articles [6–14] that US is not as effective as MR, one may be lulled into performing US examinations in such a manner that only
One should approach the examination aggressively and use a meticulous technique [2, 3]. Use of a limited fixed protocol that provides the same number of images in every neonate may not enable the examiner to find subtle changes or evaluate all intracranial contents. The approach should, therefore, be a dynamic and flexible one that can be altered or modified at any time depending on what is required for obtaining the most appropriate and optimized US examination for each neonate. Specific attention must be given to obtaining images that optimally delineate structures in both the near and far fields. The near-field gray and white matter (and the gray–white matter differentiation) (Fig. 1) of the cerebral hemispheres is particularly important to evaluate in full-term neonates. The more superficial structures such as the extra-axial fluid compartments (Fig. 2), meninges,
Fig. 1 Imaging in a full-term 3-day-old with history of hypoxic– ischemic injury and seizures. a Coronal US image shows patchy, illdefined increased echogenicity in the right perisylvian parenchyma (arrows). b A right parasagittal US image with the transducer angled laterally optimally shows the periphery of the brain and demonstrates a
more clearly abnormal area of increased echogenicity within the subcortical white matter (arrowheads) and accentuation of the gray– white differentiation. c Axial diffusion-weighted MR image shows corresponding area of restricted diffusion (arrow), indicating excellent correlation with US findings
Technique
S408
Pediatr Radiol (2015) 45 (Suppl 3):S406–S412
Fig. 2 Imaging in a 25-day-old full-term neonate with abnormal movements and possible seizure activity following repair of complex congenital heart disease. a Coronal US image shows a small right subdural collection (asterisks). This should be suspected when the sulci do not reach the skull within the dependant portions of the skull, as in this case. A much smaller and less obvious collection on the left is also
present. b Coronal US view with transducer angled toward the right more optimally shows the periphery of the brain and extra-axial space and better depicts the presence of the small subdural collection (between calipers). c Axial unenhanced CT image correlates well with the US findings, showing bilateral small subdural hematomas, larger on the right
major sulci and fissures (Fig. 3) and dural venous sinuses (Fig. 4) must be evaluated throughout as much of the cranial cavity as possible. The deeper supratentorial structures that one must also focus on are the basal ganglia (Fig. 5), thalami (Fig. 6) and corpus callosum (Fig. 7) as well as the infratentorial structures such as the brainstem and cerebellum (Fig. 8). One should keep in mind all the factors below when attempting to optimize US examinations of the brain in neonates:
(1) Acoustic windows. Multiple acoustic windows should be used in order to achieve a complete evaluation of as much of the brain and its coverings as possible. The windows include not only the anterior and posterior fontanels but also the temporal, mastoid and occipital areas and even the foramen magnum. (2) Transducers. A variety of transducers (including vector, curved and linear-array) functioning at variable megahertz should be used in each case. With each type of transducer one must optimize resolution and depth
Fig. 3 Imaging in a 9-day-old born at 32 weeks of gestation with multiple congenital anomalies and hypotonia. a Coronal magnified US image shows asymmetry in the appearance of the Sylvian fissures. Note how the right Sylvian fissure (arrow) is thicker, less-well-defined and slightly more echogenic when compared to the left. Another sulcus above the Sylvian fissure on the right has a similar appearance and
appears different from the other visible sulci. b Corresponding coronal T2-weighted MR image reveals a small area of cortical loss and increased T2 signal within the posterior aspect of the right frontal lobe adjacent to the right Sylvian fissure, consistent with an infarct in the distribution of a branch of the right middle cerebral artery. There is good correlation of the site of pathology between the modalities
Pediatr Radiol (2015) 45 (Suppl 3):S406–S412
S409
Fig. 4 Imaging in a 12-day-old term infant with septic shock and seizures. a A magnified color Doppler US image in the coronal view obtained with a linear-array transducer shows flow through the most anterior portion of the superior sagittal sinus (arrow). b A similar image obtained slightly more posteriorly than (a) shows no flow in the expected area of the superior sagittal sinus (arrow). Slightly prominent draining cortical veins are appreciated in the area. c Sagittal T1-weighted MR
image shows extensive thrombus filling and expanding the more posterior part of the superior sagittal sinus and torcular regions. Thrombus is not seen extending anteriorly; therefore flow could be elicited on the color Doppler interrogation of the most anterior portion of the superior sagittal sinus in (a). This indicates excellent correlation between the modalities
penetration. Adjustments of the megahertz used and the focal zone of the US beam enable optimal resolution of structures at different depths in the near, mid and far fields. (3) Planes. Complete examinations include images obtained in multiple planes. The standard planes are the coronal, sagittal and parasagittal as well as axial planes. However, one should not be averse to using supplementary oblique planes if the images obtained could enhance the understanding of the abnormality. (4) Angles. The angle that the transducer makes with the acoustic window used should not always be a right angle. It must be varied and angled more steeply or obtusely
because this is particularly important in evaluating as much of the periphery of the brain and its coverings as possible. (5) Field of view and magnification. Examinations that only provide images of the whole brain in each plane are inadequate. Although such examinations are useful to define the geography of the brain as a whole they do not always provide images required to evaluate details adequately (for example the gray–white matter differentiation or very small lesions). Therefore magnification of selected areas (together with focal zone and megahertz adjustments) is essential to better depict detail.
Fig. 5 Imaging in a 2-week-old term neonate with hypertension and seizure activity and who was subsequently found to have fibromuscular dysplasia. a Coronal US image shows a focal area of patchy increase in echogenicity in the left basal ganglia. b Corresponding axial apparent
diffusion coefficient map MR image reveals restricted diffusion in the left basal ganglia (arrows), showing excellent correlation between the modalities. An additional small area of restricted diffusion is seen in the periphery of the left parieto-occipital region on the MR image
S410
Pediatr Radiol (2015) 45 (Suppl 3):S406–S412
Fig. 6 Imaging in a 4-day-old term neonate with pneumococcal meningitis. a Coronal US image shows increased echogenicity in both thalami (arrows) and regions of the caudate nuclei. b Axial unenhanced CT performed the following day demonstrates bilateral multi-territorial infarcts involving both thalami and basal ganglia areas. The findings in the two modalities correlate well
Fig. 7 Imaging in a 37-week gestation term neonate with severe hypoxic–ischemic injury. a Coronal and (b) sagittal magnified US images obtained with a linear-array transducer show a thickened, hyperechoic corpus callosum (arrows). The normal corpus callosum is iso- to hypoechoic in echogencity. Note the interhemispheric fissure is completely effaced on the coronal image (a). These abnormal findings are consistent with changes from hypoxic–ischemia
Fig. 8 Imaging in a 2-week-old born at 35 weeks of gestation, with a history of persistent nystagmus and seizure activity following a traumatic delivery. a Sagittal US image shows an hyperechoic, enlarged pons (arrows). b Sagittal T1-weighted MR image following intravenous
contrast administration shows a large infiltrative lesion involving the pons that cannot be clearly differentiated from the medulla or midbrain. The lesion was subsequently proved to be a pontine glioma. There is very good correlation of the findings between US and MR
Pediatr Radiol (2015) 45 (Suppl 3):S406–S412
(6) Focal zone. It is essential to adjust the focal zone(s) available in order to optimize the image in the region of interest. This is particularly important when obtaining magnified views of small specific areas of interest. (7) Real time. A thorough neurosonographic examination includes real-time evaluation as well as a full review of the static images. Real-time evaluation enables the appreciation of subtle changes in echogenicity more easily than static images. This applies particularly to lesions that affect the gray–white matter differentiation as well as focal non-hemorrhagic infarcts, which may also cause only mild changes of echogenicity. In a busy US imaging department it may be unrealistic for all staff radiologists to check every neurosonographic examination in real time. However, in difficult cases (e.g., history of asphyxia in a term neonate or unusual infarct or hemorrhage) it is wise for the radiologist to evaluate the images either in real time or on a video recording of the examination. Currently the acquisition of cine clips reduces the time needed to acquire images at bedside and allows radiologists to appreciate subtle abnormalities that may have been missed by the technologist performing the study. Their utilization adds no time to the study, requires minimal archival space, reduces inter-observer variability and decreases the possibility of confusion with artifacts. If one relies on another operator, such as a technologist or trainee, to do the real-time examination one should be sure that the operator is technically experienced and is knowledgeable of the types and subtle appearances of the lesions encountered in this clinical setting. (8) Doppler evaluation. Normal and abnormal intracranial arteries and veins can be easily evaluated by color, power and spectral Doppler US techniques. Spectral Doppler tracings can be obtained from the major cerebral arteries and the resistive index (RI) recorded (RI=PSV- EDV/ PSV, where PSV is peak systolic velocity and EDV is end-diastolic velocity). Any fluctuations over time of the RI, PSV and EDV at the anterior cerebral artery (ACA) should be documented. Color and power Doppler evaluation may provide information regarding hyperemia of various parts of the brain, although interpretation is subjective. Doppler evaluation can depict patency and flow in the major dural venous sinuses [3, 5, 15–17] and can be used as a problem-solving tool when MR findings provide inconclusive information regarding the dural sinuses [3].
Accuracy In 2003, Epelman et al. [18] presented a prospective study at the Radiological Society of North America assembly that had
S411
evaluated US and MR imaging of the neonatal brain; the study was later published in Pediatric Radiology [3]. These authors compared the accuracy of US and MR imaging of the brain in 58 term neonates using state-of-the-art equipment at that time for both modalities and using the US technique described above. All MR examinations were performed on equipment with a 1.5-T magnet. The purpose of that study was to determine whether US imaging could compete with MR imaging. All US studies were performed within 2 h of the MRI. The radiologists performing and interpreting the US studies were blinded to the MR and clinical findings. The radiologists evaluated the US studies prospectively and a report was provided following the evaluation of the real-time and static images. The US findings were then compared to the results of the MR examinations, which had been evaluated by pediatric neuroradiologists. The results of this prospective study showed that all full-term neonates who showed a positive abnormality on MR images (on diffusion-weighted imaging or other sequences, or lactate peaks) were found to have positive abnormalities on US images. In a second study Daneman et al. [2] retrospectively evaluated the US and MR examinations in 43 other neonates who had been evaluated before the time period of the 2003 study [3, 18]. The US examination technique that had been utilized during this period was not as detailed as that used in the prospective 2003 study. The accuracy, sensitivity and specificity of US imaging were 91%, 100% and 33% in the prospective 2003 study [3, 18] and 53%, 74% and 19% in the retrospective study [2]. These studies emphasize that when comparing the utility of US in depicting lesions documented on MR imaging, the examinations must be done as closely together as possible and state-of-the-art US equipment and techniques must be used. In term neonates with hypoxic–ischemic injury, special features that must be evaluated on every US examination include the gray–white matter differentiation and focal abnormal echogenicity (which can be extremely subtle) in the cerebral hemispheres and the deeper structures. Spectral Doppler US evaluation of resistive index and fluctuations in RI, and color or power Doppler evaluation of hyperemia of various parts of the brain as well as presence or absence of flow in the dural venous sinuses can provide complementary information.
Summary Based on our experience, US remains an essential modality for evaluation of the preterm and term brain. If utilized optimally, US appears to be a more accurate modality than the older literature suggests. However, prospective studies are always necessary to compare present and future state-of-the-art US equipment with that of MRI in order to document how the accuracy of US changes with the introduction of newer
S412
Pediatr Radiol (2015) 45 (Suppl 3):S406–S412
equipment. Such studies would help us better define the role of US relative to clinical findings and relative to progress made in MR imaging (e.g., 3-T equipment). This would continue to guide us in selecting which patients require MR imaging. With the development of diagnostic imaging into divisions based on an organ system approach it seems appropriate that US of the brain be performed by radiologists involved in using other imaging modalities for the evaluation of the brain, i.e. trained neuroradiologists or radiologists with general training but who have a specific interest and detailed knowledge of brain anatomy and pathology. However this does not often occur. Many of the individuals performing US of the brain are pediatric radiologists whose major interests are in body imaging as well as pediatricians trained in neonatology and who often have had no formal training in US imaging. This may be part of the cause for the deficiencies in many earlier studies. As pediatric radiologists we are the ones responsible for providing detailed descriptions of US techniques and accurate descriptions of the abnormalities depicted by US. It is our responsibility not only to report this in peer-reviewed journals, but also to ensure that these are appropriately described in textbooks as well. We need to teach trainees in clear language how to optimize the techniques they are to utilize and what they should be searching for and we also have to ensure that we provide clinicians with the most appropriate and accurate information in a common language everyone can understand.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14. Conflicts of interest None 15.
References 16. 1.
2.
3.
4.
Epelman M, Daneman A, Kellenberger CJ et al (2004) Head US: evaluation and development of a better practice. Pediatr Radiol 34: S61 Daneman A, Epelman M, Blaser S et al (2006) Imaging of the brain in full-term neonates: does sonography still play a role? Pediatr Radiol 36:636–646 Epelman M, Daneman A, Kellenberger CJ et al (2010) Neonatal encephalopathy: a prospective comparison of head US and MRI. Pediatr Radiol 40:1640–1650 Epelman M, Daneman A, Chauvin N et al (2012) Head ultrasound and MR imaging in the evaluation of neonatal encephalopathy:
17.
18.
competitive or complementary imaging studies? Magn Reson Imaging Clin N Am 20:93–115 Dinan D, Daneman A, Guimaraes CG et al (2014) Easily overlooked sonographic findings in the evaluation of neonatal encephalopathy: lessons learned from magnetic resonance imaging. Semin Ultrasound 35:625–651 Sie LT, van der Knaap MS, van Wezel-Meijler G et al (2000) Early MR features of hypoxic–ischemic brain injury in neonates with periventricular densities on sonograms. AJNR Am J Neuroradiol 21:852–861 Blankenberg FG, Loh NN, Bracci P et al (2000) Sonography, CT, and MR imaging: a prospective comparison of neonates with suspected intracranial ischemia and hemorrhage. AJNR Am J Neuroradiol 21:213–219 Childs AM, Cornette L, Ramenghi LA et al (2001) Magnetic resonance and cranial ultrasound characteristics of periventricular white matter abnormalities in newborn infants. Clin Radiol 56:647–655 Miller SP, Cozzio CC, Goldstein RB et al (2003) Comparing the diagnosis of white matter injury in premature newborns with serial MR imaging and transfontanel ultrasonography findings. AJNR Am J Neuroradiol 24:1661–1669 Debillon T, N’Guyen S, Muet A et al (2003) Limitations of ultrasonography for diagnosing white matter damage in preterm infants. Arch Dis Child 88:275–279 Inder TE, Anderson NJ, Spencer C et al (2003) White matter injury in the premature infant: a comparison between serial cranial sonographic and MR findings at term. AJNR Am J Neuroradiol 24:805– 809 Maalouf EF, Duggan PJ, Counsell SJ et al (2001) Comparison of findings on cranial ultrasound and magnetic resonance imaging in preterm infants. Pediatrics 107:719–727 Blankenberg FG, Norbash AM, Lane B et al (1996) Neonatal intracranial ischemia and hemorrhage: diagnosis with US, CT and MR imaging. Radiology 199:253–259 Roelants-van Rijn AM, Groenendaal F, Beck FJ et al (2001) Parenchymal brain injury in the preterm infant: comparison of cranial ultrasound, MRI and neurodevelopmental outcome. Neuropediatrics 32:80–89 Miller E, Daneman A, Doria AS et al (2012) Color Doppler US of normal cerebral venous sinuses in neonates: a comparison with MR venography. Pediatr Radiol 42:1070–1079 Taylor G (2005) Doppler of the neonatal and infant brain. In: Rumack CM, Wilson SR, Charboneau JW et al (eds) Diagnostic ultrasound, 3rd edn. Elsevier Mosby, St. Louis, pp 1703–1722 Dean LM, Taylor GA (1995) The intracranial venous system in infants: normal and abnormal findings on duplex and color Doppler sonography. AJR Am J Roentgenol 164:151–156 Epelman M, Daneman A, Konen O et al (2003) Perinatal brain injury. A prospective comparison of state-of-the-art ultrasound (US) and magnetic resonance imaging (MRI): can US compete with MRI? Presented at the Radiological Society of North America 89th Scientific Assembly, 30 Nov to 5 Dec. Chicago, IL