Use of the Posterior Auricular Artery for Indirect Bypass in Moyamoya: A Pediatric Case Series

Encephaloduroarteriosynangiosis (EDAS) for indirect revascularization of the brain in moyamoya disease and syndrome is predominantly performed using a branch of the superficial temporal artery (STA), often the parietal branch, as the donor artery. We have at times used other branches of the external carotid artery as the donor artery, including the posterior auricular artery (PAA). We have used the PAA when there has been no suitable branch of the STA, e.g., when the parietal branch has been too small to use, and as an additional donor artery, e.g., when we judged that the STA branch that was available was too anterior to revascularize the peri-Rolandic cortex by itself. In this study, we review 3 cases that used the PAA for EDAS to demonstrate that this artery can be successfully used for indirect bypass in children and adolescents.

Briefly, EDAS using the PAA is performed much like it would be using the STA, except for the location of the artery. The suitability of either artery for EDAS is evaluated preoperatively on catheter angiogram. When the PAA is used, patients are positioned supine with a roll under the ipsilateral shoulder. The head is then turned such that the scalp superior to the mastoid process is facing the ceiling. A suitable PAA can usually be found with Doppler just behind the superior pinna and above the mastoid (Fig. 1). It can then usually be traced superiorly for several cm before the Doppler signal is no longer audible. Usually, the length available for EDAS is less than that available when using the parietal branch of the STA, as the PAA travels mostly superiorly, whereas the parietal branch meanders both superiorly and posteriorly. We therefore dissect as long a length of the PAA as possible when using it for indirect revascularization. It is our practice to dissect the donor artery distally until it starts to leave the surface of the galea aponeurotica and ascend into the subcutaneous fat. Once the donor artery leaves the stabilizing surface of the galea, it becomes more difficult to dissect without injuring it. The artery is then elevated with a narrow cuff of galea and a craniotomy is performed. The dura is opened in line with the donor artery. Usually, we open the arachnoid over prominent sulci. The galeal cuff and artery are inverted and laid over the brain and the cuff is sutured to the dura. Lately, we have been sewing the galea to the dura at either end of the dissected artery and to the pia in between, as a modified pial synangiosis, although this was not done for any of the 3 patients in this series. We have also been opening the dura alongside multiple meningeal artery branches as described by Sandoval-Garcia et al. [unpublished work: presented at 44th Annual Meeting of the AANS/CNS Joint Section on Pediatric Neurological Surgery, December 2015; manuscript in preparation – communication with the senior author]. Closure is routine except that there is usually no temporalis muscle to close. Patients are maintained on aspirin before and after surgery.

We reviewed a retrospective database of all moyamoya patients at Children’s Hospital Colorado who underwent revascularization surgery from 2005 through 2017. Study data were collected and managed using REDCap electronic data capture tools hosted at the University of Colorado, Denver [1]. We selected 3 patients who underwent EDAS using the PAA and reviewed their presentation, anatomy, and radiologic and clinical outcomes.

We evaluated revascularization on cerebral catheter angiograms, when available, obtained at least 3 months after surgery. When not available, we evaluated the degree of increase in the ratio of the diameter of the donor PAA over the diameter of the contralateral PAA (both measured by two board-certified pediatric neuroradiologists – IN, JM) between a preoperative magnetic resonance angiogram (MRA) and a postoperative MRA obtained at least 3 months after surgery. We have found that such an increase can be correlated with revascularization as seen on postoperative catheter angiogram (manuscript in revision). In addition, we graded the extent of cerebral surface collateral formation adjacent to the donor artery (PAA) on postoperative axial MRA images as well as MIPs (maximum intensity projections) as poor, meaning no collaterals were seen, fair, or 2 robust. Finally, we assessed whether there was a preoperative ivy sign [2] and whether it improved on postoperative FLAIR (fluid attenuated inversion recovery) images and whether there was improved perfusion on arterial spin-labeled (ASL) MR images.

A 3-year-old female presented with progressive right hemiparesis, difficulty ambulating, urinary incontinence, and speech regression. Physical examination was normal. Catheter cerebral angiogram demonstrated bilateral occlusion of the internal carotid arteries, with collateral vessels supplying bilateral anterior and middle cerebral arteries. The posterior cerebral arteries (PCAs) were also involved, with extensive collateral vessels.

The patient first underwent left STA EDAS. Three months later, she underwent right EDAS. As the parietal branch of the right STA was slightly further forward than usual (Fig. 2) and we wished to ensure good revascularization of the peri-Rolandic cortex, we augmented a right STA EDAS with a PAA EDAS. The parietal STA branch and PAA were approached via two separate skin incisions. The parietal STA branch was inadvertently sharply cut during dissection and was reanastomosed. Two separate craniotomies were made. The STA EDAS was augmented with dural inversion around a large branch of the middle meningeal artery (MMA), as described by Gadgil et al. [3] (this was before we started routinely opening the dura alongside multiple meningeal artery branches during EDAS). Her symptoms all recovered, except that she occasionally had difficulty remembering or pronouncing words when tired. Due to extremely poor blood flow in the PCAs and anterior cerebral arteries (ACAs) on catheter angiogram, the patient later underwent bilateral occipital artery (OA) EDAS and bilateral midline occipital and frontal midline ribbon synangiosis [4], all without complication.

Postoperative catheter cerebral angiograms showed good revascularization from all her procedures (Fig. 2). Anteriorly, there was more revascularization via the right MMA than the right parietal STA donor artery. There was excellent revascularization through an enlarged PAA. On MRA, robust collateral formation was seen on the surface of the brain next to the patent PAA donor artery. The ratio of the diameter of the PAA donor artery over the diameter of the contralateral PAA on axial MRA was increased from prior to surgery (Fig. 2). At last follow-up, the patient remains neurologically intact on physical examination, with good balance and full, symmetrical strength in all her extremities. She has occasional headaches. She was noted to have a seizure-like episode and was placed on anti-epileptic medications, without recurrence.

A 12-year-old male with previous headaches presented after an episode of headache followed by lightheadedness, paresthesias in the hands, and slurred speech. Neurologic examination was normal. MRI/MRA showed severe left middle cerebral artery (MCA) stenosis but no infarct. On FLAIR images, there was subtly increased signal throughout many of the left cerebral sulci, or an ivy sign. ASL imaging demonstrated increased blood volume throughout the left cerebral sulci.

Catheter cerebral angiogram demonstrated proximal MCA occlusion with distal MCA flow through prominent collateral vessels. There were no substantial left STA parietal branches but was a large left PAA (Fig. 3).

The patient underwent left PAA EDAS. At 2-year follow-up, he remained neurologically intact, with no headaches or ischemic symptoms. Repeat MRI/MRA showed no new ischemic changes. The donor artery was patent and the ratio of its diameter over the diameter of the right PAA had increased from prior to surgery (Fig. 3). Fair collateral formation was seen on the surface of the brain near the donor artery. The ivy sign was subtly improved (not shown); unfortunately, no postoperative ASL sequence has been done.

A 17-year-old female who had previously undergone two resections and subsequent radiation treatment for craniopharyngioma was found to have narrow MCAs on routine follow-up MRI. There were no acute ischemic changes. On FLAIR imaging, there were bilateral ivy signs, more prominent on the left. The patient had preexisting learning delays. Six months earlier, she had had an episode of left facial droop and hemiparesis with slurred speech that resolved over a week. Neurologic examination was normal.

Catheter cerebral angiogram demonstrated occlusion of both MCAs with filling of both ACAs, as well as significant left proximal PCA stenosis. Her MCA territories were partially supplied on the left via the ACA and extensively supplied on the right by her posterior circulation and by branches from the external carotid artery at her previous craniotomy site. On the left, the only substantial STA branch supplying the parietal scalp arose from the frontal branch too far superiorly to be ideal for EDAS, but there was a large left PAA (Fig. 4).

The patient underwent left PAA EDAS. Right EDAS was not performed due to spontaneous revascularization at her previous craniotomy site. Two years later, after ASL imaging showed poor perfusion in the left posterior parietal and occipital lobes (not shown), she also underwent left OA EDAS. ASL imaging had not been performed prior to her first EDAS. At last follow-up, she had a normal neurologic exam with no recurrence of ischemic symptoms or headaches. Postoperative MRI/MRA showed fair collateral formation on the surface adjacent to the patent PAA donor artery. The ratio of the diameter of the PAA donor artery over the diameter of the contralateral PAA had increased from prior to surgery (Fig. 4). The ivy sign on FLAIR was improved on the left (not shown), although the postoperative FLAIR imaging was contrast enhanced, whereas the preoperative FLAIR was without contrast, making the improvement hard to interpret.

EDAS was developed by Matsushima and colleagues in 1979 [5]. The procedure was inspired by a 1964 report by Tsubokawa et al. [6] in which a dural graft containing the MMA was used to revascularize an area of cerebral ischemia after intracranial arterial thrombosis.

EDAS and pial synangiosis for anterior circulation moyamoya are predominantly performed using the STA [7, 8]. Although direct bypass for moyamoya using the PAA has been reported [9, 13], to our knowledge only recently has there been a publication focusing on the suitability of the PAA for EDAS, in pediatric or adult patients. Lee et al. [14] evaluated anatomic variations of the STA and PAA in a series of catheter angiograms in pediatric moyamoya patients and quantified variations in which the PAA could be used instead of the STA. They found that the STA was unsuitable for use in EDAS in 6.8% of hemispheres, for reasons such as absence, hypoplasia, or anterior positioning of the parietal branch of the STA. Interestingly, they found that 49% of PAAs did not reach the top of the pinna, which would make them unsuitable for EDAS. They defined a PAA as being adequate for EDAS if it was visible on angiogram halfway between the top of the pinna and the vertex. Ten hemispheres in this study underwent PAA EDAS, and the authors found similar Matsushima grades [15] in patients who underwent PAA versus STA EDAS. In addition, Hersh et al. [16] reported use of the PAA for indirect revascularization in 2 pediatric patients with radiation-induced cerebral vasculopathy as part of a larger study in which various external carotid branches, including the STA and OA, were used; they reported robust collateralization formation at 1 year post-surgery, with no post-surgical infarctions.

In our study, the PAA alone was used for EDAS in 2 subjects in whom there was no parietal STA branch (cases 2, 3), and both the parietal STA branch and the ipsilateral PAA were used in 1 subject in whom the parietal STA branch was a little more anterior than we thought was ideal for EDAS (case 1). Of the PAA donor arteries, one (case 1) extended less than halfway between the top of the pinna and the vertex, one (case 2) extended greater than halfway between the top of the pinna and the vertex, and one (case 3) extended all the way to the vertex. All 3 subjects – including the case in which the PAA did not reach the halfway point between the top of the pinna and the vertex had good evidence of cerebral revascularization from their PAA donor artery by catheter angiogram and/or magnetic resonance angiography. All 3 had improvement in their symptoms post-EDAS. None has had a post-EDAS stroke. These results, combined with the results of Hersh et al. [16], support consideration of the PAA for indirect revascularization procedures in children and adolescents when the STA is not a suitable candidate or when an additional donor artery is desired.

There are several disadvantages and advantages in using the PAA versus the STA for indirect revascularization (Table 1). The caliber of the most proximal portion of the PAA available for use as a donor artery is generally less than that of the parietal branch of the STA, making it slightly more challenging to dissect. In a series from our center of 44 STA parietal branch donor arteries, the average diameter measured just after the takeoff from the STA trunk was 1.62 mm, whereas in the 3 PAAs in this paper, the average diameter measured at approximately the level of the top of the pinna was 1.13 mm (unpublished data). In addition, the overall length of the PAA available for use as a donor artery is less than that of the parietal branch of the STA due to the former’s more vertical course. On the other hand, at the most, only a very short incision in the temporalis muscle – if any at all – is necessary to use the PAA as a donor artery. Also, when the PAA is used, the entire length that is laid on the brain is actually in contact with the brain, whereas a portion of the parietal STA branch donor artery is generally in contact with the Sylvian fissure and not the brain, possibly making ingrowth from it less likely. Also, according to Lee et al. [14], the PAA is often not suitable for EDAS, with nearly half not extending beyond the top of the pinna.

Two patients in this series also underwent OA EDAS [15] and one underwent bilateral OA EDAS as well as contralateral STA EDAS and bilateral frontal and occipital ribbon synangiosis. The decision to perform additional revascularization procedures after a primary surgery or to perform multiple concurrent revascularization procedures is multifactorial. We may add an OA procedure, usually indirect, if there is severe or progressive PCA stenosis, if there are ischemic signs or symptoms referable to the PCA territory, or if there is radiologic evidence of poor flow in the PCA – e.g., the ivy sign on FLAIR MRI images [2] or poor perfusion on ASL MRI images. Ribbon synangiosis [4] may be performed in analogous circumstances to revascularize the midline cortex.

In 2 of our 3 cases, no routine 1-year postoperative catheter angiogram was obtained. For several years, all routine postoperative neurovascular imaging studies after revascularization surgery for moyamoya at our institution have been MRAs. We find that, although revascularization cannot be seen as well on MRA as on catheter angiogram, it can be seen well enough in patients who are doing well. We look for an increased ratio of the diameter of the donor vessel to the diameter of the contralateral vessel on postoperative versus preoperative axial MRA as well as surface collaterals on the brain adjacent to the donor artery as reliable indicators of revascularization [unpublished work: presented at 42nd Annual Meeting of the American Society of Pediatric Neurosurgeons, January 2019; manuscript in revision]. The contralateral vessel serves as an internal control as measured diameters in the same blood vessel may be affected by differences in technique and blood flow in different scans. In our 2 cases with no postoperative catheter angiogram, there was fair collateralization. We find that the diameters of the donor artery and corresponding contralateral vessel can best be seen on axial MRA images. New surface collaterals are also seen well on axial MRA and on MIPs, although their true extent is typically revealed to be significantly greater on catheter angiograms. Previous studies have evaluated revascularization after moyamoya surgery using MRA MIP images [17, 18] or ASL imaging [19, 22], or have compared the postoperative diameters of the STA and ipsilateral MMA to their preoperative diameters on MRA and/or catheter angiography [18, 20, 23]. No previous studies have used contralateral vessel diameters as internal controls. We still perform routine preoperative catheter angiograms for surgical planning and postoperative catheter angiograms when there is radiologic evidence of moyamoya progression on MRA or clinical progression on history or examination.

Although all 3 patients had good long-term outcomes, 2 of the 3 had more than one revascularization procedure, making it hard to attribute their outcomes to their PAA EDAS. This paper demonstrates that the PAA is an option for EDAS in children and adolescents when the STA has no suitable branches. It is not, however, meant to provide criteria for what is “suitable” and what is not. This will be different for different surgeons and in different situations. Cases 2 and 3 in our series illustrate situations in which we thought that a large PAA was more suitable for EDAS than was the STA, and case 1 illustrates a situation in which we augmented revascularization through the STA with a PAA EDAS – even though the PAA was relatively small – when the parietal branch was usable but was judged to be a little too far forward to be ideal.

Finally, only one of our cases had a postoperative catheter angiogram. Unfortunately, neither of the other two had both a preoperative and postoperative ASL image or other perfusion study. One of the patients had only preoperative ASL MRI images and the other had only postoperative ASL images. ASL imaging is routinely included in MRI for pediatric stroke and moyamoya at our institution. Unfortunately, it previously was sometimes inadvertently omitted. Both patients without postoperative angiograms had preoperative ivy signs on FLAIR imaging. However, we did not include FLAIR images from either patient because one of them was only subtly improved after surgery whereas the other had FLAIR imaging without contrast before surgery and FLAIR with contrast after surgery. Although the ivy sign seemed to have improved after surgery, how contrast affects the ivy sign on FLAIR seems to be variable. A paper by Yoon et al. [16] found that contrast increased the ivy sign in 7 of 28 hemispheres of children with moyamoya and decreased it in 4; it was unchanged in the rest. Our latter patient did have later postoperative FLAIR imaging without contrast on which the ivy sign also appeared to have improved, but these were performed after the patient had an ipsilateral OA EDAS, another confounding factor. While we believe that the change in ratio of the donor vessel over the contralateral vessel between preoperative and postoperative MRA scans is a reliable indicator of revascularization, as of the date of submission of this paper, those data are not yet published.

The PAA is a viable option for EDAS in children and adolescents and can be successfully used when the STA is not suitable as a donor artery or when an additional donor artery is desired.

This study protocol was reviewed and approved by the Colorado Multiple Institutional Review Board, approval number 18-0890. Written informed consent for publication of this case report and any accompanying images was obtained from patients over 18 and from parents/legal guardians of patients under 18.

The authors have no conflicts of interest to declare.

This research received no specific funding.

J. Chris Hawking and C. Corbett Wilkinson made substantial contributions to the design and conception of this project. J. Chris Hawkins, Megan Ryan, Sarah Graber, Ilana Neuberger, Jodi Slade, Michael Young, John Maloney, and C. Corbett Wilkinson were involved in the acquisition and interpretation of data and critically revised the manuscript and have approved the manuscript as it is written. J. Chris Hawkins, Michael Young, and C. Corbett Wilkinson compiled the primary manuscript. J. Chris Hawkins, Sarah Graber, Ilana Neuberger, Jodi Slade, Megan Ryan, and John Maloney compiled the images. Jodi Slade is the artist who painted Figure 1. Megan Ryan and C Corbett Wilkinson compiled Table 1.

The data that support the findings of this study are not publicly available due their containing information that could compromise the privacy of research participants but are available from the corresponding author (CCW) upon reasonable request.