Neuroradiology DOI 10.1007/s00234-014-1455-7
INTERVENTIONAL NEURORADIOLOGY
Cranial Doppler ultrasound in Vein of Galen malformation Dan Meila & Kathrin Lisseck & Collin Jacobs & Heinrich Lanfermann & Friedhelm Brassel & Axel Feldkamp
Received: 2 June 2014 / Accepted: 15 October 2014 # Springer-Verlag Berlin Heidelberg 2014
Abstract Introduction Vein of Galen malformation (VGM) is the severest paediatric neurovascular disease usually requiring multiple staged embolisations. In the high-risk group of children with high-flow arteriovenous shunts, timing of treatment is uncertain. Low Doppler resistance index (RI) is known to be associated with adverse outcome in hypoxic-ischaemic brain injury in children. In this study, we want to present our longterm results of cranial transfontanellar Doppler ultrasound in children with VGM. Methods We identified and retrospectively analysed 264 transfontanellar Doppler measurements in 19 endovasculartreated true VGM (five females, 14 males) between 2000 and 2013. The recordings were obtained from the internal carotid arteries (ICA), the anterior cerebral arteries (ACA) and the basilar arteries (BA). Maximal systolic velocity (Vs), enddiastolic velocity (Ved) and the RI were measured before and after embolisation. Results Untreated, nearly all cases showed pathologic high systolic (up to >1.0 m/s), very high diastolic velocities (up to >0.5 m/s) and low RI (30 years, DEGUM III). The measurements were obtained on routine pre-operative and immediate postoperative ultrasound diagnostic workups. We excluded results from children that were restless or crying. The investigations were analysed and evaluated by one of the first authors (K. L.). In all children, sagittal and coronal sections were performed through the open fontanelle as an acoustic window as described by Deeg [7]. The recordings were obtained from the internal carotid arteries (ICA), the anterior cerebral arteries (ACA) and the basilar arteries (BA). From the flow profile, the maximal systolic velocity (Vs), the end-diastolic velocity (Ved) and the RI were measured before and after shunt reduction by endovascular means. Vs corresponds to the peak of flow profile, whereas Ved marks the end of the pulse cycle, according to Deeg [8]. The RI by Pourcelot [6] is defined as RI ¼ Vs−Ved Vs All radiographic studies, hospital charts and follow-up outcome of each patient were retrospectively reviewed and analysed by both first authors equally (D. M. and K. L.). All children were clinically examined by our neonatologists and neuropaediatricians. The outcome score was graded on a five-point scale according to Jones et al. [9] ranging from 0 (death) to 4 (normal). Different types of VGM were classified as choroidal or mural on the basis of their angioarchitecture.
way ANOVA was used with the post hoc Tukey test. A p value 1.0 m/s) and very high diastolic (up to >0.5 m/s) velocities (mean Vs 0.85 m/s and mean Ved 0.41 m/s, data not shown). Vs and Ved were not significantly different between the three arteries, neither before nor after treatment. RI before and after treatment The RI is not dependent on the ultrasound angle and thus reflects reliable values. There were no statistically significant differences of the pre-embolisation RI between the three different measured arteries, i.e. ICA, BA and ACA (Fig. 1). There was a statistically significant difference (p=0.012) between the pre-embolisation RI and the post-embolisation RI with pathologic low RI (0.4) before and nearly normal RI (0.57) after successful shunt reduction (Fig. 2). RI and clinical outcome The children with bad outcomes (1 and 2) also had pathologic low RI (mean 0.51), while the majority of cases with good outcome (4) showed normal RI (mean 0.6) after treatment. There were no statistically significant differences between the different outcome scores with respect to the post-embolisation RI. However, a tendency towards correlation could be observed (Fig. 3).
Statistical assessment
Illustrative cases
Statistical analyses were performed using SPSS 20.0 (SPSS Inc., Chicago, IL, USA) and SAS 9.2 (SAS Institute, Cary, NC, USA). To detect any difference between the groups, one-
In the following, we highlight salient features of the typical Doppler findings in VGM with the help of some illustrative cases (Figs. 4, 5, 6 and 7).
5 days 5 months 2 days 5 days 8 days 3 days 4 days 10 days 2 months 6 months 1 day 4 days 1 day 4 months 3 months 7 days 20 days 3 months
M F M M M M M M F F F M M F M M m M
M
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
19
RI resistance index, n. a. not applicable
4 months
Age at first treatment
Patient no.Sex
Table 1 The results of the study
Choroidal
Choroidal Choroidal Choroidal Choroidal Choroidal Choroidal Choroidal Choroidal Choroidal Choroidal Choroidal Choroidal Choroidal Choroidal Choroidal Choroidal Mural Choroidal
Type of VGM
No symptoms
Congestive heart failure Seizures, stroke Congestive heart failure Congestive heart failure Congestive heart failure Congestive heart failure Congestive heart failure Congestive heart failure Mild heart failure Mild heart failure Congestive heart failure Congestive heart failure Congestive heart failure Mild heart failure Mild heart failure Congestive heart failure Congestive heart failure Congestive heart failure
Presenting symptoms
3 Mean: SEM:
2 13 4 6 2 6 3 4 4 2 3 9 3 3 4 5 2 6
Total no. of treatments
0.74 0.41 0.12
n. a. 0.47 0.2 0.41 0.4 n. a. 0.47 0.35 0.39 0.65 0.33 n. a. 0.39 0.57 0.63 0.26 0.44 0.34
RI pre-embo
n. a. 0.57 0.09
0.42 0.68 0.56 0.55 0.46 0.53 0.54 0.57 n. a. n. a. 0.51 0.56 0.57 0.65 0.71 0.44 0.72 0.49
RI post-embo
100
100 50 100 100 50 50 >90 100 75 100 50 100 100 50 100 100 100 100
Angiographic obliteration (%)
4
4 2 4 4 1 1 4 4 3 4 2 4 4 4 4 4 4 3
Outcome score
Neuroradiology
Neuroradiology
Fig. 1 RI pre-embolisation in the ICA, BA and ACA. Data are means ± standard error of the mean (SEM). The black bar in the middle of each box represents the median. The box includes all values between the 25th and 75th percentiles. Whiskers indicate values still within the 1.5 interquartile range (IQR). There were no statistically significant differences between the three different measured arteries (p > 0.05). n.s. not significant.
Typical transfontanellar Doppler in a healthy child and in a VGM child Case 1 A male neonate was transferred to our hospital on the day of his birth because of his worsening congestive heart failure in consequence of a VGM. The Doppler ultrasound showed a high diastolic flow and a very low RI of 0.36 (Fig. 5a). The neonate had signs of heart failure like tachypnoea, sometimes
Fig. 3 RI post-embolisation and clinical outcome. The children with bad outcomes (1 or 2) had pathologic low RI, while the majority of cases with good outcome (4) showed normal RI. Data are means±SEM. There were no statistically significant differences of the RI between the different outcome scores (p>0.05)
even apnoea and arterial hypotension. Seven days after his birth, the first embolisation was executed in which an immediate shunt reduction was achieved (Fig. 6). We used a technique called “kissing microcatheter technique” [2] that lasted for 5 h. Primarily, the arterial microcatheter was navigated into one of the main arterial feeders of the VGM. After, a second microcatheter was navigated into the venous collector of the VGM via the transfemoral venous route. After visualising the fistulous connections, the microcatheter from the venous side was advanced into the feeding artery in a retrograde fashion. Then coiling of the fistulous connections was performed as described by Meila et al. [2]. Nevertheless, four additional embolisation sessions were necessary. The second session was performed 6 weeks later, the third and the fourth each 4 months later and the last session 6 months. After the third embolisation, the RI reached 0.57 (Fig. 5b) which is nearly within the normal range. After five embolisations in a one and a half year period, full closure of the VGM was achieved enabling cessation of medication. The boy is continuing to show a normal development without neurological reduction. Case 2
Fig. 2 RI pre- vs post-embolisation (all three arteries pooled, n=13). Data are means±SEM. There was a statistically significant difference (p=0.012) between the pre-embolisation RI and the post-embolisation RI with pathologic low RI (0.4) before and nearly normal RI (0.57) after treatment. The black bar in the middle of each box represents the median. The box includes all values between the 25th and 75th percentiles. Whiskers indicate values still within the 1.5 interquartile range (IQR). An asterisk (*) represents a significant difference between the values under the clamp
A 20-day-old boy with macrocrania was transferred to our department from another hospital. A true VGM with a mural, fistulous type of a shunt was seen in the ultrasound and MRI/ MRA. The remarkably enlarged vein compressed the sylvian aqueduct and led to a severe and symptomatic hydrocephalus. Initial Doppler measurements showed pathologic values with a Vs of 0.79 m/s and a Ved of 0.43 m/s resulting in a pathologic low RI of 0.45 (Fig. 7a). In two embolisation sessions in a 3-week period, closure of all fistulous connections was achieved. Immediate post-embolisation Doppler 1 day after treatment (Fig. 7b) showed normalisation of the
Neuroradiology
Fig. 4 Typical transfontanellar Doppler ultrasound in a healthy child (a) and in an untreated child with a VGM (b) for comparison. (a) (measurement of the right ACA) shows normal values with Vs 0.5 and Ved 0.1 resulting in a RI of 0.8. (b) (measurement of the right ICA) shows a
typical slower increase (arrow) to a high systolic velocity peak (Vs 1.2 m/ s) and a remarkable very high end-diastolic velocity (Ved 0.8 m/s) leading to a pathologic low RI of 0.33 in the VGM child
velocities with a Vs of 0.44 m/s and a Ved of 0.1 m/s. The RI was within the normal range measuring 0.77. Furthermore, the ultrasound confirmed the beginning thrombosis of the enlarged venous sac. The boy was transferred back to a completely normal neurological status.
Due to the time-related limitation of transfontanellar Doppler usage, it is unclear to which extent the further increase of RI value would have reached if one could measure after closure of fontanelles. Thrombosis processes do occur, and sometimes, important reorganisation of the hemodynamic situation may need time. We did not include patients with VGM diagnosed or treated after the first year of life. However, it is known that these children are in most of the cases less severe symptomatic than neonates [1]. At least, in most of the cases, they do not present with congestive heart failure, rather with hydrocephalus and/or with other symptoms. Eventually, Doppler measurements of velocities depend on the ultrasound angle, and thus, high intra- and inter-operator measuring variability is unavoidable. According to Deeg and Rupprecht [8], flow velocities showed a strong age as well as weight dependency. These parameters must be taken into consideration when pathologic flow velocities are analysed. Furthermore, they reported that in contrast, the RI did not show any age or weight dependency. Besides, the physiologic factors influencing the flow velocities in brain arteries especially the pCO2 and the vigilance of the child must be taken into consideration. A high pCO2 results in a dramatic increase, whereas a fall in pCO2 leads to a decrease of the flow velocities. A rise of blood pressure in crying children is followed by an increase of all flow velocities especially in diastole [8]. For these reasons, we focussed on our study on the independent RI measurements.
Discussion The present study intends to share our long-term results of transfontanellar Doppler ultrasound measurements in children with VGM. Limitations of the study Firstly, the number of cases is low. However, the study deals with a rare disease and only a few centres in the world have a larger patient population of children with VGM. Thus, only a multicentric study would provide more data. Secondly, there are certain limitations due to the retrospective nature of the study. There were no precisely defined and standardised ultrasound investigations in children with VGM before the start of this study. Some children were referred to us in emergency situations. In some of these situations, the decision was taken not to lose time with additional investigations before endovascular treatment. On the other hand, some children were transferred back to the referring hospitals before our experienced neonatologist could perform the final Doppler examination. This is the reason why some measurements are missing. Thirdly, the use of transfontanellar investigations is limited by nature due to the closure of the fontanelles during the maturing process after the children’s first year of life. Normalisation of the cerebral vascularisation and angioarchitecture after successful shunt reduction or closure of all fistulous connections may take a certain period of time.
Flow velocities, RI and clinical outcome Our findings in VGM are similar to Deeg and Rupprecht’s results in healthy children [8] showing that Vs, Ved and RI were not significantly different in the three measured great intracranial arteries (ICA, BA and ACA). The ultrasound Doppler investigation in children, especially in newborns, can be challenging when children begin to become restless or start crying. Furthermore, there is not enough data until
Neuroradiology Fig. 5 A 7-day-old neonate presented with severe congestive heart failure. a An initial Doppler measurement of the BA. A high systolic velocity (Vs 1.4 m/s) and a very high diastolic velocity (Ved 0.9 m/s) are demonstrated leading to a pathologic low RI of 0.36. Note the typical long and flat sloping diastole in VGM. b Transfontanellar Doppler after two embolisations. Vs is now within the normal range with 0.49 m/s and Ved markedly decreased to 0.21 m/s, but still two times higher than normal. The resulting RI of 0.57 is nearly normal
now about possible thermal modifications of the brain in preterm or newborn children after Doppler ultrasound waves. Children with VGM are severely ill, and they are in most of the cases already extremely impaired. In this context, our findings may be helpful and time-saving. Based on our findings, we suggest measuring only one of the three arteries in the course of clinical investigations and routine paediatric ultrasound workups in VGM. The use of Doppler imaging in children with VGM has been reported for the first time in some case reports in the late 1980s [10–12]. However, until now, no larger study has been published showing a consecutive series of Doppler
measurements with reliable values in VGM. In our study, all children with VGM showed a typical flow profile with a slow acceleration slope to a high systolic velocity peak (Vs) and a very high diastole (Ved) with a long and flat deceleration slope. This is a very unique and typical finding that can only be found in VGM. These findings were in accordance with pathologic low RI. The changes could be observed in all three measured great intracranial arteries in untreated VGM. We assume that the reason must be a counter regulation of the brain as changes concerning all intracranial arteries can also be found under other pathologic circumstances, e.g. patent ductus arteriosus. We believe that the main and most
Neuroradiology
Fig. 6 Initial DSA of the first embolisation. Lateral view of a vertebral artery injection showing choroidal type of VGM (a). “Kissing microcatheter technique”, as described by Meila et al. [2], in a nonsubtracted lateral view, same projection (b). One microcatheter was navigated transarterial, the second transvenous, both bridging one fistulous connection of the VGM. Demonstrated coiling of the fistulous connection (c, d); non-subtracted (c) and subtracted (d) angiography, same projection. Note the visible normal posterior cerebral artery branch after embolisation (arrows)
important reasons leading to a low RI are caused by the increase of the diastole. Due to the high-flow arteriovenous malformation in VGM hypoperfusion of the surrounding, normal brain parenchyma occurs or may occur. Consequently, cerebral autoregulation follows, aiming at regulating cerebral perfusion. This pathophysiologic counter regulation is reflected by a high and long, flat sloping diastole. In this context, typical pathologic low RI in VGM can be observed. Archer et al. and Levene et al. reported that an abnormal Doppler measurement with an RI below 0.55 in the ACA of asphyxiated neonates correlates with a poor neurodevelopmental prognosis [4, 13]. The same group could also show that in asphyxiated infants, Doppler examination
predicted outcome with an accuracy of 86 % [4]. Nishimaki et al. [14] recently reported in another study on blood flow velocities in the ACA and the BA in asphyxiated infants. Children with severe asphyxia had a mean RI of 0.52 in the ACA and a mean RI of 0.53 in the BA, respectively. Similar to our findings, the difference between the RI in the ACA and BA in the severe asphyxia group was not significant. Nevertheless, the literature clearly shows a correlation between low RI and poor outcomes. In our study, 16 out of 19 untreated children with VGM had pathologic low RI with a mean RI of 0.41. Out of the three cases with normal preembolisation RI, one had no symptoms and the other two had only mild heart failure symptoms, not requiring early or urgent endovascular treatment. On the other hand, nearly all of the 16 patients with pathologic low RI had severe symptoms, i.e. in most of the cases, congestive heart failure. Furthermore, we were able to show that after treatment, the groups with bad outcome (1 and 2) had a mean pathologic low RI of 0.51, whereas the group with normal outcome had a mean RI of 0.6, being in the normal range, respectively. There were no statistically significant differences of the RI between the different outcome scores. However, our patient population size was too small to reach statistical significance in this context. One of the biggest challenges in the management of VGM is the treatment timing decision. Most of the children present in early childhood when neurologic and especially neurodevelopmental examination is very difficult. Children may become symptomatic with seizures or neurodevelopmental delay when irreversible brain damage already has occurred. In this context, further embolisations cannot rebuild or reorganise already damaged important cerebral structures and pathways. Even the extraordinary capacity and plasticity of the brain in young children may not change poor outcome in the presence of some predictive factors. In this context, Geibprasert et al. [15] presented in a study several factors that were significantly associated with a poor outcome in VGM. These factors included, amongst others, a very poor score in one (or more) of the following categories, i.e. focal parenchymal changes, calcifications, tonsillar herniation, arterial steal or more than two groups of multiple arterial feeders. These findings can be best diagnosed by MRI/MRA and digital subtraction angiography (DSA). Due to radiation and contrast material need for DSA and long-lasting anaesthesia procedures for MR, both investigations are not feasible for regularly close follow-up controls in early childhood. Moreover, the presence of one of these signs/factors leads in most of the times to irreversible changes. For the abovementioned reasons, we do not believe that elective embolisations should start at the age of 4–5 months. In some cases, this waiting period might be too long. Irreversible parenchyma changes may occur without that child’s parents or physician note to them. Hence, we propose additional regular Doppler investigation with the transfontanellar
Neuroradiology Fig. 7 A 20-day-old boy with a VGM presented with congestive heart failure and hydrocephalus. Initial Doppler (a) showing pathologic high systolic (0.79 m/ s) and diastolic (0.43 m/s) velocities and a pathologic low RI (0.45). The arrow shows the enlarged prosencephalic vein of Markowski. Immediate postembolisation Doppler (b) showing normalisation of the velocities (Vs 0.44 and Ved 0.1 m/s) and the RI (0.77). Note the subsequently thrombosed venous sac (arrow)
Conclusion
pathologic to normal do occur after successful endovascular embolisation. We propose the use of cranial Doppler ultrasound as an adjunctive technique to other screenings, and it should be taken into consideration for further investigations as well as for routine workups. The following conclusions can be drawn from this clinical investigation:
As far as we know, this is the largest study, by now, showing characteristic pathologic cranial Doppler values in a consecutive series of VGM. Furthermore, this is the first study showing that changes of cranial Doppler RI measurements from
1.) Untreated, all cases showed pathologic high systolic and very high diastolic velocities. Vs and Ved were not significantly different between the three arteries, neither before nor after treatment.
measurement of one of the great intracranial cerebral arteries. Further research is warranted to evaluate the role of the RI in the treatment timing decision.
Neuroradiology
2.) There were no statistically significant differences of the pre-embolisation RI between the three different measured arteries, i.e. ICA, BA and ACA. Hence, measuring one of the three arteries is sufficient for future clinical investigations and for future routine paediatric ultrasound workups in VGM. 3.) There were statistically significant differences between the pre-embolisation RI and the post-embolisation RI with pathologic low RI before and normal or nearly normal RI after successful shunt reduction. 4.) There were no statistically significant differences of the RI after treatment between the different outcome scores, but however, a tendency towards correlation could be observed. Children with bad outcomes also had pathologic low RI, while the majority of cases with good outcome showed normal RI values.
Ethical standards and patient consent We declare that this manuscript does not contain clinical studies or patient data. Conflict of interest We declare that we have no conflict of interest.
References 1. Lasjaunias P, Berenstein A, terBrugge K (2006) Surgical neuroangiography: clinical and interventional aspects in children. Springer, Berlin 2. Meila D, Hannak R, Feldkamp A, Schlunz-Hendann M, Mangold A, Jacobs C, Papke K, Brassel F (2012) Vein of Galen aneurysmal malformation: combined transvenous and transarterial method using a “kissing microcatheter technique”. Neuroradiology 54(1):51–59 3. Berenstein A, Fifi JT, Niimi Y et al (2012) Vein of Galen malformations in neonates: new management paradigms for
improving outcomes. Neurosurgery 70(5):1207–1213, discussion 1213–4 4. Archer LN, Levene MI, Evans DH (1986) Cerebral artery Doppler ultrasonography for prediction of outcome after perinatal asphyxia. Lancet 2(8516):1116–1118 5. Gray PH, Tudehope DI, Masel JP, Burns YR, Mohay HA, O’Callaghan MJ, Williams GM (1993) Perinatal hypoxic-ischaemic brain injury: prediction of outcome. Dev Med Child Neurol 35(11): 965–973 6. Pourcelot L (1974) Applications cliniques de l’examen Doppler transcutane. In: Peronneau P (ed) Velocimetrie ultrasonore Doppler. INSERM 34, 213–240 7. Deeg KH (1989) Color flow imaging of the great intracranial arteries in infants. Neuroradiology 31:40–43 8. Deeg KH, Rupprecht T (1989) Pulsed Doppler sonographic measurement of normal values for the flow velocities in the intracranial arteries of healthy newborns. Pediatr Radiol 19:71–78 9. Jones BV, Ball WS, Tomsick TA, Millard J, Crone KR (2002) Vein of Galen aneurysmal malformation: diagnosis and treatment of 13 children with extended clinical follow-up. AJNR Am J Neuroradiol 23: 1717–1724 10. Deeg KH, Scharf J (1990) Colour Doppler imaging of arteriovenous malformation of the vein of Galen in a newborn. Neuroradiology 32: 60–63 11. Tessler FN, Dion J, Vinuela F et al (1989) Cranial arteriovenous malformations in neonates: color Doppler imaging with angiographic correlation. AJR 150:1027–1030 12. Abbitt PL, Hurst RW, Ferguson RDG et al (1990) The role of ultrasound in the management of vein of Galen aneurysms in infancy. Neuroradiology 32:86–89 13. Levene MI, Fenton AC, Evans DH, Archer LN, Shortland DB, Gibson NA (1989) Severe birth asphyxia and abnormal cerebral blood-flow velocity. Dev Med Child Neurol 31(4):427–434 14. Nishimaki S, Iwasaki S, Minamisawa S, Seki K, Yokota S (2008) Blood flow velocities in the anterior cerebral artery and basilar artery in asphyxiated infants. J Ultrasound Med 27(6):955–960 15. Geibprasert S, Krings T, Armstrong D, terBrugge KG, Raybaud CA (2010) Predicting factors for the follow-up outcome and management decisions in vein of Galen aneurysmal malformations. Childs Nerv Syst 26:35–46
INTERVENTIONAL NEURORADIOLOGY
Cranial Doppler ultrasound in Vein of Galen malformation Dan Meila & Kathrin Lisseck & Collin Jacobs & Heinrich Lanfermann & Friedhelm Brassel & Axel Feldkamp
Received: 2 June 2014 / Accepted: 15 October 2014 # Springer-Verlag Berlin Heidelberg 2014
Abstract Introduction Vein of Galen malformation (VGM) is the severest paediatric neurovascular disease usually requiring multiple staged embolisations. In the high-risk group of children with high-flow arteriovenous shunts, timing of treatment is uncertain. Low Doppler resistance index (RI) is known to be associated with adverse outcome in hypoxic-ischaemic brain injury in children. In this study, we want to present our longterm results of cranial transfontanellar Doppler ultrasound in children with VGM. Methods We identified and retrospectively analysed 264 transfontanellar Doppler measurements in 19 endovasculartreated true VGM (five females, 14 males) between 2000 and 2013. The recordings were obtained from the internal carotid arteries (ICA), the anterior cerebral arteries (ACA) and the basilar arteries (BA). Maximal systolic velocity (Vs), enddiastolic velocity (Ved) and the RI were measured before and after embolisation. Results Untreated, nearly all cases showed pathologic high systolic (up to >1.0 m/s), very high diastolic velocities (up to >0.5 m/s) and low RI (30 years, DEGUM III). The measurements were obtained on routine pre-operative and immediate postoperative ultrasound diagnostic workups. We excluded results from children that were restless or crying. The investigations were analysed and evaluated by one of the first authors (K. L.). In all children, sagittal and coronal sections were performed through the open fontanelle as an acoustic window as described by Deeg [7]. The recordings were obtained from the internal carotid arteries (ICA), the anterior cerebral arteries (ACA) and the basilar arteries (BA). From the flow profile, the maximal systolic velocity (Vs), the end-diastolic velocity (Ved) and the RI were measured before and after shunt reduction by endovascular means. Vs corresponds to the peak of flow profile, whereas Ved marks the end of the pulse cycle, according to Deeg [8]. The RI by Pourcelot [6] is defined as RI ¼ Vs−Ved Vs All radiographic studies, hospital charts and follow-up outcome of each patient were retrospectively reviewed and analysed by both first authors equally (D. M. and K. L.). All children were clinically examined by our neonatologists and neuropaediatricians. The outcome score was graded on a five-point scale according to Jones et al. [9] ranging from 0 (death) to 4 (normal). Different types of VGM were classified as choroidal or mural on the basis of their angioarchitecture.
way ANOVA was used with the post hoc Tukey test. A p value 1.0 m/s) and very high diastolic (up to >0.5 m/s) velocities (mean Vs 0.85 m/s and mean Ved 0.41 m/s, data not shown). Vs and Ved were not significantly different between the three arteries, neither before nor after treatment. RI before and after treatment The RI is not dependent on the ultrasound angle and thus reflects reliable values. There were no statistically significant differences of the pre-embolisation RI between the three different measured arteries, i.e. ICA, BA and ACA (Fig. 1). There was a statistically significant difference (p=0.012) between the pre-embolisation RI and the post-embolisation RI with pathologic low RI (0.4) before and nearly normal RI (0.57) after successful shunt reduction (Fig. 2). RI and clinical outcome The children with bad outcomes (1 and 2) also had pathologic low RI (mean 0.51), while the majority of cases with good outcome (4) showed normal RI (mean 0.6) after treatment. There were no statistically significant differences between the different outcome scores with respect to the post-embolisation RI. However, a tendency towards correlation could be observed (Fig. 3).
Statistical assessment
Illustrative cases
Statistical analyses were performed using SPSS 20.0 (SPSS Inc., Chicago, IL, USA) and SAS 9.2 (SAS Institute, Cary, NC, USA). To detect any difference between the groups, one-
In the following, we highlight salient features of the typical Doppler findings in VGM with the help of some illustrative cases (Figs. 4, 5, 6 and 7).
5 days 5 months 2 days 5 days 8 days 3 days 4 days 10 days 2 months 6 months 1 day 4 days 1 day 4 months 3 months 7 days 20 days 3 months
M F M M M M M M F F F M M F M M m M
M
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
19
RI resistance index, n. a. not applicable
4 months
Age at first treatment
Patient no.Sex
Table 1 The results of the study
Choroidal
Choroidal Choroidal Choroidal Choroidal Choroidal Choroidal Choroidal Choroidal Choroidal Choroidal Choroidal Choroidal Choroidal Choroidal Choroidal Choroidal Mural Choroidal
Type of VGM
No symptoms
Congestive heart failure Seizures, stroke Congestive heart failure Congestive heart failure Congestive heart failure Congestive heart failure Congestive heart failure Congestive heart failure Mild heart failure Mild heart failure Congestive heart failure Congestive heart failure Congestive heart failure Mild heart failure Mild heart failure Congestive heart failure Congestive heart failure Congestive heart failure
Presenting symptoms
3 Mean: SEM:
2 13 4 6 2 6 3 4 4 2 3 9 3 3 4 5 2 6
Total no. of treatments
0.74 0.41 0.12
n. a. 0.47 0.2 0.41 0.4 n. a. 0.47 0.35 0.39 0.65 0.33 n. a. 0.39 0.57 0.63 0.26 0.44 0.34
RI pre-embo
n. a. 0.57 0.09
0.42 0.68 0.56 0.55 0.46 0.53 0.54 0.57 n. a. n. a. 0.51 0.56 0.57 0.65 0.71 0.44 0.72 0.49
RI post-embo
100
100 50 100 100 50 50 >90 100 75 100 50 100 100 50 100 100 100 100
Angiographic obliteration (%)
4
4 2 4 4 1 1 4 4 3 4 2 4 4 4 4 4 4 3
Outcome score
Neuroradiology
Neuroradiology
Fig. 1 RI pre-embolisation in the ICA, BA and ACA. Data are means ± standard error of the mean (SEM). The black bar in the middle of each box represents the median. The box includes all values between the 25th and 75th percentiles. Whiskers indicate values still within the 1.5 interquartile range (IQR). There were no statistically significant differences between the three different measured arteries (p > 0.05). n.s. not significant.
Typical transfontanellar Doppler in a healthy child and in a VGM child Case 1 A male neonate was transferred to our hospital on the day of his birth because of his worsening congestive heart failure in consequence of a VGM. The Doppler ultrasound showed a high diastolic flow and a very low RI of 0.36 (Fig. 5a). The neonate had signs of heart failure like tachypnoea, sometimes
Fig. 3 RI post-embolisation and clinical outcome. The children with bad outcomes (1 or 2) had pathologic low RI, while the majority of cases with good outcome (4) showed normal RI. Data are means±SEM. There were no statistically significant differences of the RI between the different outcome scores (p>0.05)
even apnoea and arterial hypotension. Seven days after his birth, the first embolisation was executed in which an immediate shunt reduction was achieved (Fig. 6). We used a technique called “kissing microcatheter technique” [2] that lasted for 5 h. Primarily, the arterial microcatheter was navigated into one of the main arterial feeders of the VGM. After, a second microcatheter was navigated into the venous collector of the VGM via the transfemoral venous route. After visualising the fistulous connections, the microcatheter from the venous side was advanced into the feeding artery in a retrograde fashion. Then coiling of the fistulous connections was performed as described by Meila et al. [2]. Nevertheless, four additional embolisation sessions were necessary. The second session was performed 6 weeks later, the third and the fourth each 4 months later and the last session 6 months. After the third embolisation, the RI reached 0.57 (Fig. 5b) which is nearly within the normal range. After five embolisations in a one and a half year period, full closure of the VGM was achieved enabling cessation of medication. The boy is continuing to show a normal development without neurological reduction. Case 2
Fig. 2 RI pre- vs post-embolisation (all three arteries pooled, n=13). Data are means±SEM. There was a statistically significant difference (p=0.012) between the pre-embolisation RI and the post-embolisation RI with pathologic low RI (0.4) before and nearly normal RI (0.57) after treatment. The black bar in the middle of each box represents the median. The box includes all values between the 25th and 75th percentiles. Whiskers indicate values still within the 1.5 interquartile range (IQR). An asterisk (*) represents a significant difference between the values under the clamp
A 20-day-old boy with macrocrania was transferred to our department from another hospital. A true VGM with a mural, fistulous type of a shunt was seen in the ultrasound and MRI/ MRA. The remarkably enlarged vein compressed the sylvian aqueduct and led to a severe and symptomatic hydrocephalus. Initial Doppler measurements showed pathologic values with a Vs of 0.79 m/s and a Ved of 0.43 m/s resulting in a pathologic low RI of 0.45 (Fig. 7a). In two embolisation sessions in a 3-week period, closure of all fistulous connections was achieved. Immediate post-embolisation Doppler 1 day after treatment (Fig. 7b) showed normalisation of the
Neuroradiology
Fig. 4 Typical transfontanellar Doppler ultrasound in a healthy child (a) and in an untreated child with a VGM (b) for comparison. (a) (measurement of the right ACA) shows normal values with Vs 0.5 and Ved 0.1 resulting in a RI of 0.8. (b) (measurement of the right ICA) shows a
typical slower increase (arrow) to a high systolic velocity peak (Vs 1.2 m/ s) and a remarkable very high end-diastolic velocity (Ved 0.8 m/s) leading to a pathologic low RI of 0.33 in the VGM child
velocities with a Vs of 0.44 m/s and a Ved of 0.1 m/s. The RI was within the normal range measuring 0.77. Furthermore, the ultrasound confirmed the beginning thrombosis of the enlarged venous sac. The boy was transferred back to a completely normal neurological status.
Due to the time-related limitation of transfontanellar Doppler usage, it is unclear to which extent the further increase of RI value would have reached if one could measure after closure of fontanelles. Thrombosis processes do occur, and sometimes, important reorganisation of the hemodynamic situation may need time. We did not include patients with VGM diagnosed or treated after the first year of life. However, it is known that these children are in most of the cases less severe symptomatic than neonates [1]. At least, in most of the cases, they do not present with congestive heart failure, rather with hydrocephalus and/or with other symptoms. Eventually, Doppler measurements of velocities depend on the ultrasound angle, and thus, high intra- and inter-operator measuring variability is unavoidable. According to Deeg and Rupprecht [8], flow velocities showed a strong age as well as weight dependency. These parameters must be taken into consideration when pathologic flow velocities are analysed. Furthermore, they reported that in contrast, the RI did not show any age or weight dependency. Besides, the physiologic factors influencing the flow velocities in brain arteries especially the pCO2 and the vigilance of the child must be taken into consideration. A high pCO2 results in a dramatic increase, whereas a fall in pCO2 leads to a decrease of the flow velocities. A rise of blood pressure in crying children is followed by an increase of all flow velocities especially in diastole [8]. For these reasons, we focussed on our study on the independent RI measurements.
Discussion The present study intends to share our long-term results of transfontanellar Doppler ultrasound measurements in children with VGM. Limitations of the study Firstly, the number of cases is low. However, the study deals with a rare disease and only a few centres in the world have a larger patient population of children with VGM. Thus, only a multicentric study would provide more data. Secondly, there are certain limitations due to the retrospective nature of the study. There were no precisely defined and standardised ultrasound investigations in children with VGM before the start of this study. Some children were referred to us in emergency situations. In some of these situations, the decision was taken not to lose time with additional investigations before endovascular treatment. On the other hand, some children were transferred back to the referring hospitals before our experienced neonatologist could perform the final Doppler examination. This is the reason why some measurements are missing. Thirdly, the use of transfontanellar investigations is limited by nature due to the closure of the fontanelles during the maturing process after the children’s first year of life. Normalisation of the cerebral vascularisation and angioarchitecture after successful shunt reduction or closure of all fistulous connections may take a certain period of time.
Flow velocities, RI and clinical outcome Our findings in VGM are similar to Deeg and Rupprecht’s results in healthy children [8] showing that Vs, Ved and RI were not significantly different in the three measured great intracranial arteries (ICA, BA and ACA). The ultrasound Doppler investigation in children, especially in newborns, can be challenging when children begin to become restless or start crying. Furthermore, there is not enough data until
Neuroradiology Fig. 5 A 7-day-old neonate presented with severe congestive heart failure. a An initial Doppler measurement of the BA. A high systolic velocity (Vs 1.4 m/s) and a very high diastolic velocity (Ved 0.9 m/s) are demonstrated leading to a pathologic low RI of 0.36. Note the typical long and flat sloping diastole in VGM. b Transfontanellar Doppler after two embolisations. Vs is now within the normal range with 0.49 m/s and Ved markedly decreased to 0.21 m/s, but still two times higher than normal. The resulting RI of 0.57 is nearly normal
now about possible thermal modifications of the brain in preterm or newborn children after Doppler ultrasound waves. Children with VGM are severely ill, and they are in most of the cases already extremely impaired. In this context, our findings may be helpful and time-saving. Based on our findings, we suggest measuring only one of the three arteries in the course of clinical investigations and routine paediatric ultrasound workups in VGM. The use of Doppler imaging in children with VGM has been reported for the first time in some case reports in the late 1980s [10–12]. However, until now, no larger study has been published showing a consecutive series of Doppler
measurements with reliable values in VGM. In our study, all children with VGM showed a typical flow profile with a slow acceleration slope to a high systolic velocity peak (Vs) and a very high diastole (Ved) with a long and flat deceleration slope. This is a very unique and typical finding that can only be found in VGM. These findings were in accordance with pathologic low RI. The changes could be observed in all three measured great intracranial arteries in untreated VGM. We assume that the reason must be a counter regulation of the brain as changes concerning all intracranial arteries can also be found under other pathologic circumstances, e.g. patent ductus arteriosus. We believe that the main and most
Neuroradiology
Fig. 6 Initial DSA of the first embolisation. Lateral view of a vertebral artery injection showing choroidal type of VGM (a). “Kissing microcatheter technique”, as described by Meila et al. [2], in a nonsubtracted lateral view, same projection (b). One microcatheter was navigated transarterial, the second transvenous, both bridging one fistulous connection of the VGM. Demonstrated coiling of the fistulous connection (c, d); non-subtracted (c) and subtracted (d) angiography, same projection. Note the visible normal posterior cerebral artery branch after embolisation (arrows)
important reasons leading to a low RI are caused by the increase of the diastole. Due to the high-flow arteriovenous malformation in VGM hypoperfusion of the surrounding, normal brain parenchyma occurs or may occur. Consequently, cerebral autoregulation follows, aiming at regulating cerebral perfusion. This pathophysiologic counter regulation is reflected by a high and long, flat sloping diastole. In this context, typical pathologic low RI in VGM can be observed. Archer et al. and Levene et al. reported that an abnormal Doppler measurement with an RI below 0.55 in the ACA of asphyxiated neonates correlates with a poor neurodevelopmental prognosis [4, 13]. The same group could also show that in asphyxiated infants, Doppler examination
predicted outcome with an accuracy of 86 % [4]. Nishimaki et al. [14] recently reported in another study on blood flow velocities in the ACA and the BA in asphyxiated infants. Children with severe asphyxia had a mean RI of 0.52 in the ACA and a mean RI of 0.53 in the BA, respectively. Similar to our findings, the difference between the RI in the ACA and BA in the severe asphyxia group was not significant. Nevertheless, the literature clearly shows a correlation between low RI and poor outcomes. In our study, 16 out of 19 untreated children with VGM had pathologic low RI with a mean RI of 0.41. Out of the three cases with normal preembolisation RI, one had no symptoms and the other two had only mild heart failure symptoms, not requiring early or urgent endovascular treatment. On the other hand, nearly all of the 16 patients with pathologic low RI had severe symptoms, i.e. in most of the cases, congestive heart failure. Furthermore, we were able to show that after treatment, the groups with bad outcome (1 and 2) had a mean pathologic low RI of 0.51, whereas the group with normal outcome had a mean RI of 0.6, being in the normal range, respectively. There were no statistically significant differences of the RI between the different outcome scores. However, our patient population size was too small to reach statistical significance in this context. One of the biggest challenges in the management of VGM is the treatment timing decision. Most of the children present in early childhood when neurologic and especially neurodevelopmental examination is very difficult. Children may become symptomatic with seizures or neurodevelopmental delay when irreversible brain damage already has occurred. In this context, further embolisations cannot rebuild or reorganise already damaged important cerebral structures and pathways. Even the extraordinary capacity and plasticity of the brain in young children may not change poor outcome in the presence of some predictive factors. In this context, Geibprasert et al. [15] presented in a study several factors that were significantly associated with a poor outcome in VGM. These factors included, amongst others, a very poor score in one (or more) of the following categories, i.e. focal parenchymal changes, calcifications, tonsillar herniation, arterial steal or more than two groups of multiple arterial feeders. These findings can be best diagnosed by MRI/MRA and digital subtraction angiography (DSA). Due to radiation and contrast material need for DSA and long-lasting anaesthesia procedures for MR, both investigations are not feasible for regularly close follow-up controls in early childhood. Moreover, the presence of one of these signs/factors leads in most of the times to irreversible changes. For the abovementioned reasons, we do not believe that elective embolisations should start at the age of 4–5 months. In some cases, this waiting period might be too long. Irreversible parenchyma changes may occur without that child’s parents or physician note to them. Hence, we propose additional regular Doppler investigation with the transfontanellar
Neuroradiology Fig. 7 A 20-day-old boy with a VGM presented with congestive heart failure and hydrocephalus. Initial Doppler (a) showing pathologic high systolic (0.79 m/ s) and diastolic (0.43 m/s) velocities and a pathologic low RI (0.45). The arrow shows the enlarged prosencephalic vein of Markowski. Immediate postembolisation Doppler (b) showing normalisation of the velocities (Vs 0.44 and Ved 0.1 m/s) and the RI (0.77). Note the subsequently thrombosed venous sac (arrow)
Conclusion
pathologic to normal do occur after successful endovascular embolisation. We propose the use of cranial Doppler ultrasound as an adjunctive technique to other screenings, and it should be taken into consideration for further investigations as well as for routine workups. The following conclusions can be drawn from this clinical investigation:
As far as we know, this is the largest study, by now, showing characteristic pathologic cranial Doppler values in a consecutive series of VGM. Furthermore, this is the first study showing that changes of cranial Doppler RI measurements from
1.) Untreated, all cases showed pathologic high systolic and very high diastolic velocities. Vs and Ved were not significantly different between the three arteries, neither before nor after treatment.
measurement of one of the great intracranial cerebral arteries. Further research is warranted to evaluate the role of the RI in the treatment timing decision.
Neuroradiology
2.) There were no statistically significant differences of the pre-embolisation RI between the three different measured arteries, i.e. ICA, BA and ACA. Hence, measuring one of the three arteries is sufficient for future clinical investigations and for future routine paediatric ultrasound workups in VGM. 3.) There were statistically significant differences between the pre-embolisation RI and the post-embolisation RI with pathologic low RI before and normal or nearly normal RI after successful shunt reduction. 4.) There were no statistically significant differences of the RI after treatment between the different outcome scores, but however, a tendency towards correlation could be observed. Children with bad outcomes also had pathologic low RI, while the majority of cases with good outcome showed normal RI values.
Ethical standards and patient consent We declare that this manuscript does not contain clinical studies or patient data. Conflict of interest We declare that we have no conflict of interest.
References 1. Lasjaunias P, Berenstein A, terBrugge K (2006) Surgical neuroangiography: clinical and interventional aspects in children. Springer, Berlin 2. Meila D, Hannak R, Feldkamp A, Schlunz-Hendann M, Mangold A, Jacobs C, Papke K, Brassel F (2012) Vein of Galen aneurysmal malformation: combined transvenous and transarterial method using a “kissing microcatheter technique”. Neuroradiology 54(1):51–59 3. Berenstein A, Fifi JT, Niimi Y et al (2012) Vein of Galen malformations in neonates: new management paradigms for
improving outcomes. Neurosurgery 70(5):1207–1213, discussion 1213–4 4. Archer LN, Levene MI, Evans DH (1986) Cerebral artery Doppler ultrasonography for prediction of outcome after perinatal asphyxia. Lancet 2(8516):1116–1118 5. Gray PH, Tudehope DI, Masel JP, Burns YR, Mohay HA, O’Callaghan MJ, Williams GM (1993) Perinatal hypoxic-ischaemic brain injury: prediction of outcome. Dev Med Child Neurol 35(11): 965–973 6. Pourcelot L (1974) Applications cliniques de l’examen Doppler transcutane. In: Peronneau P (ed) Velocimetrie ultrasonore Doppler. INSERM 34, 213–240 7. Deeg KH (1989) Color flow imaging of the great intracranial arteries in infants. Neuroradiology 31:40–43 8. Deeg KH, Rupprecht T (1989) Pulsed Doppler sonographic measurement of normal values for the flow velocities in the intracranial arteries of healthy newborns. Pediatr Radiol 19:71–78 9. Jones BV, Ball WS, Tomsick TA, Millard J, Crone KR (2002) Vein of Galen aneurysmal malformation: diagnosis and treatment of 13 children with extended clinical follow-up. AJNR Am J Neuroradiol 23: 1717–1724 10. Deeg KH, Scharf J (1990) Colour Doppler imaging of arteriovenous malformation of the vein of Galen in a newborn. Neuroradiology 32: 60–63 11. Tessler FN, Dion J, Vinuela F et al (1989) Cranial arteriovenous malformations in neonates: color Doppler imaging with angiographic correlation. AJR 150:1027–1030 12. Abbitt PL, Hurst RW, Ferguson RDG et al (1990) The role of ultrasound in the management of vein of Galen aneurysms in infancy. Neuroradiology 32:86–89 13. Levene MI, Fenton AC, Evans DH, Archer LN, Shortland DB, Gibson NA (1989) Severe birth asphyxia and abnormal cerebral blood-flow velocity. Dev Med Child Neurol 31(4):427–434 14. Nishimaki S, Iwasaki S, Minamisawa S, Seki K, Yokota S (2008) Blood flow velocities in the anterior cerebral artery and basilar artery in asphyxiated infants. J Ultrasound Med 27(6):955–960 15. Geibprasert S, Krings T, Armstrong D, terBrugge KG, Raybaud CA (2010) Predicting factors for the follow-up outcome and management decisions in vein of Galen aneurysmal malformations. Childs Nerv Syst 26:35–46
