Introduction

Extracranial-Intracranial (EC/IC) arterial bypass surgery is a well known microsurgical procedure for augmenting distal cerebral circulation. Indications include the prevention of recurrent stroke in hemodynamic insufficiency due to occlusive cerebrovascular disease and its use as an adjunct in the treatment of complex cerebral aneurysms and skull base tumors [1, 2, 3, 4, 5, 6, 7]. The most frequent surgical modification of the technique consists in establishing a direct end-to-side anastomosis between either the frontal or parietal branch of the superficial temporal artery (STA) and a distal segment of the middle cerebral artery (MCA).

In the past various approaches were used to assure STA-MCA bypass function. Quantitative and qualitative ultrasound based techniques are applied during surgery, while conventional catheter angiography (CA) is used routinely for the assessment of bypass patency in the early postoperative course [8, 9, 10]. While intraoperative techniques can assess bypass functioning within a narrow time window only, CA allows time-independent investigation of the established anastomosis. In addition, CA enables grading of bypass function to some extent [3, 11, 12] for estimating the extent of blood supply via the EC/IC anastomosis. The information obtained by CA, however, appears to be investigator dependent and provides qualitative information only. Furthermore, this method carries some serious procedure-related risks, such as arterial embolism, arterial dissection, and groin hematoma and infection [13]. Thus, noninvasive, quantitative, and investigator-independent imaging techniques are desired for the serial assessment of EC/IC bypass patency and measurement of blood supply through the anastomosis in the postoperative course. The introduction of new magnetic resonance (MR) based imaging techniques such two-dimensional cine-phase contrast MR makes possible the noninvasive qualitative [14, 15, 16, 17] and quantitative [18, 19, 20, 21] investigation of blood flow within cerebral vessels. These techniques permit the examination of cerebral vessels without the need for intra-arterial contrast agents. Due to their technical principles [20] and their robustness these approaches allow the direct assessment of flow direction and measurement of volume flow even in small arterial vessels under physiological conditions.

The aim of this study was (a) to quantify an MR-based angiographic (MRA) grading system for the standard STA-MCA anastomosis function used at present, (b) to determine its value in measuring cerebral blood supply via the EC/IC bypass, and (c) to characterize the potential role of quantitative MR-based techniques for the investigation of bypass patency in the postoperative course.

Materials and methods

The study enrolled 41 patients (33 men, 8 women; mean age 57±9 years) with either transient or minor retinal ischemic attacks or minor stroke and confirmed occlusion of the ICA or MCA ( n =34) or stenosis of the MCA or ICA ( n =7) in CA, and without significant angiographic collateral flow via the ophthalmic artery, who fulfilled the criteria for application of EC/IC arterial bypass surgery.

Indication for EC/IC arterial bypass surgery and surgical procedure

The indication for EC/IC bypass surgery in our institution is, in addition to the necessity for symptomatic occlusive cerebrovascular disease, based upon the quantitative assessment of regional CBF (rCBF) and cerebrovascular reserve capacity (CVRC). Patients are considered candidates for surgical revascularization if CVRC is either less than 30%, or if a paradoxical decrease in rCBF (“steal phenomenon”) occurs after acetazolamide challenge (acetazolamide 15 mg/kg/body weight intravenously) [22]. Only patients with normal cranial computed tomography or evidence of border-zone infarction undergo evaluation procedures. Functional rCBF studies are performed after diagnostic CA by means of stable xenon-enhanced CT (4.5 min wash-in protocol, 30% Xe, 60% O2).

Patients meeting the inclusion criteria, i.e., in whom hemodynamic compromise is assured, undergo standard EC/IC bypass surgery. Neurosurgical procedure consists in the establishment of a direct anastomosis between either the frontal or the parietal branch of the superficial temporal artery (STA; donor vessel) and a cortical (recipient) vessel of the middle cerebral artery (MCA—M2 or M3 segment) via a standard temporal craniotomy [23, 24].

Conventional angiography

STA-MCA anastomosis were investigated by CA after the surgical procedure in all patients to assess bypass patency postoperatively. Bypass function and collateral circulation through the anastomosis were graded as follows extent [3, 11]: extensive (grade III), ante- and retrograde filling of the entire MCA system; moderate (grade II), filling of two or more MCA branches; and poor (grade I), filling of the anastomosed MCA branch only (Fig. 1). The bypass grade was assessed by two independent observers and finally rated using consensus among the observers.

Fig. 1
figure 1

Angiographic EC/IC arterial bypass grading (lateral projection of selective external carotid artery injection). A Grade I bypass: poor intracranial opacification. B Grade II: moderate opacification. C Grade III bypass: extensive angiographic filling

Full size image

MRI and MRA studies

MRI, MRA, and blood volume flow (BVF) studies were performed using a 1.5-T MR unit (Magnetom Vision, Siemens, Erlangen, Germany). All patients underwent identical MR protocols using a circular polarized head coil and a Helmholtz neck coil. After performing T1-weighted scout images (TR: 545 ms; TE: 15 ms slice thickness: 4 mm; slice gap: 0.6 mm) to gain anatomical reference information, BVF was assessed quantitatively within the established STA-MCA anastomosis. Sufficient angiographic and MRA studies were obtained in all patients; both investigations were carried out within 10 days after surgery.

Dynamic two-dimensional cine-phase contrast MR technique was applied using electrocardiography-triggered fast radiofrequency spoiled gradient-echo sequence with following sequence parameters: TR: 28 ms, TE: 5 ms, flip angle: 30°, FOV: 220 mm, matrix: 192×256, 1 acquisition. Velocity encoding was set between 40 and 250 cm/s depending on blood flow velocity of the measured vessel. Depending on the patient’s heart rate, 25–35 single two-dimensional phase-contrast images were acquired over the cardiac cycle (time resolution 28 ms). All BVF measurements were made perpendicular to the course of the arteries, within the straight segment of the distal STA (Fig. 2). Values for group BVF are presented as mean and standard deviation. Statistical evaluation of data used the paired t test for normally distributed data samples.

Fig. 2
figure 2

Two-dimensional cine-phase MR technique. A Coronal T2-weighted reference slice with measurement plane placed over straight segment of the superficial temporal artery. B Original registration curve obtained within the superficial temporal artery. C, D Postoperative MRA in a patient with unilateral carotid artery occlusion: coronal ( C) and transversal ( D) views

Full size image

Results

Selective external carotid artery injection investigating the EC/IC anastomosis led to the following results. According to the diagnostic criteria applied [3], bypass filling was classified as grade I in 6 cases (15%), grade II in 15 (36%), and grade III in 20 (49%; Table 1). BVF obtained by means of two-dimensional cine-phase contrast MRI ranged between 13.8 and 177.6 ml/min in the EC/IC anastomosis, with a mean of 79.4±30.4 ml/min. Mean BVF in patients with grade I bypass filling was 31.9±9.8 ml/min, ranging from 13.8 to 48.0 ml/min. Significantly increased BVF was observed in grade II and grade III patients, in whom mean BVF reached 73.6±16.7 ml/min ( P <0.001 vs. grade I; range: 48.3–111.0 ml/min) and 97.2±26.6 ml/min ( P <0.001 vs. grade I; range: 49.2–177.6 ml/min), respectively.

Table 1 Blood volume flow
Full size table

All patients with BVF values of48 ml/min or lower ( n =6) showed grade I bypass function in angiography. Patients with grade II or III bypass demonstrated higher BVF values, with only BVF higher than 111 ml/min (3/20 studies, 15%) being specific for grade III bypass function. In 17 (85%) investigations with angiographic grade III bypass BVF values were within the range obtained in patients with grade II bypass. Thus despite a statistically significant difference between the patient groups with grade II and grade III bypass function (grade II vs. grade III, P <0.001) no specific lower cutoff BVF value was found representative for a grade III bypass that would allow differentiation between grade II and III by means of BVF.

Discussion

The main findings of the present study are that (a) angiographic grading of bypass function compares well with the BVF through the anastomosis in the early postoperative course, (b) considerable brain blood supply is provided via angiographically “poor” functioning bypasses, and (c) noninvasive MRI techniques are more effective for assessing bypass function under physiological conditions.

EC/IC arterial bypass surgery was used extensively for a variety of ischemic and latent ischemic disorders of the brain before the results of the international cooperative study of EC/IC anastomosis were published in 1985 [25]. The study failed to show a reduction in stroke and stroke-related deaths in patients with stenotic or occlusive cerebrovascular disease undergoing EC/IC bypass surgery. The patient selection criteria applied in this study were criticized extensively, leading to the development of the concept of hemodynamic insufficiency in occlusive and stenotic cerebrovascular disease over the following years [26, 27, 28, 29, 30, 31, 32, 33]. Based on this theory, patients with assured hemodynamic compromise are most likely to benefit from EC/IC bypass surgery since the risk of subsequent stroke in these patients is known to be increased in the natural course [32, 34].

The efficacy of this surgical procedure in terms of stroke prevention, which has been demonstrated in various studies [2, 3, 35, 36, 37, 38, 39], depends fundamentally upon graft patency and the capability of the EC/IC bypass to deliver blood into the dependent vascular territory. CA, which is widely available, is currently used to assess graft patency. CA allows reliable detection of primary bypass failure and furthermore enables the qualitative assessment of intracranial filling via the established anastomosis.

In the past, several angiographic grading systems have been introduced to measure blood supply through the bypass [3, 11, 12, 40]. These systems analyze the extent of intracranial vascular filling after selective external carotid artery injection using either a three-point [3, 11] or eight-point [12, 40] grading scale. The systems in use at present distinguish only between “low,” “medium,” and “high” or “poor,” “moderate,” and, “extensive” bypass function. Single studies have demonstrated a correlation between angiographic bypass filling and clinical course [40, 41] and intraoperative changes in rCBF [42]. However, criticism has been directed at the procedure itself. Since the manual injection of a contrast dye under variable pressure is necessary for angiographic evaluation, nonphysiological conditions are assumed to be present during CA. Consequently, alternative imaging techniques are required which allow the noninvasive assessment of graft patency and its function under physiological conditions in the postoperative course.

MRA, which permit the noninvasive investigation of the cerebral circulation, has therefore been to study EC/IC bypass patency [14, 15, 16, 17]. These investigations, however, have focused primarily on the capability of MRA technique to confirm bypass graft patency when shown with CA. The results demonstrate good agreement between the two techniques, with MRA providing additional information concerning flow direction within the bypass [16]. Since the introduction of two-dimensional cine-phase contrast MRI technique, which allows the quantitative measurement of blood volume flow within blood vessels, the methodology has gained wide acceptance for various indications [18, 19, 20, 21]. This technique, which provides intravascular BVF rates in absolute values, blood flow velocity, and blood flow characteristics and hence permits the quantitative evaluation of EC/IC arterial bypass function.

The patency of a EC/IC anastomosis depends on factors related both to the patient and to the surgery. Patient-associated factors include size and compatibility of donor and recipient vessel, extent of atherosclerotic changes in the donor vessel, and the patient’s blood requirement distal to the arterial stenosis or occlusion. In addition, the microsurgical technique applied directly affects graft patency. All patients studied here suffered from hemodynamic compromise, i.e., exhausted CVRC in functional rCBF studies. The current requirement for additional blood allowing the restoration of cerebral perfusion was thus assured prior to operation for every patient. The angiographic investigation of bypass patency therefore reflects structural anastomosis properties rather than hemodynamic aspects within the revascularized territory after EC/IC bypass surgery. Nevertheless, there are certainly individual local differences in blood requirement. Their effect on both BVF and vessel opacification with CA is uncertain.

The present study shows the practicability of a three-point angiographic bypass grading scale for estimating blood flow via an EC/IC anastomosis in the early postoperative course. The obtained group BVF values, measured by both two-dimensional cine-phase contrast MRI compared well with those using the angiographic scale. Thus the extent of intracranial blood vessel filling after selective external carotid artery injection, as qualitative measure for bypass function, reflects the structural properties of the EC/IC anastomosis to some extent. This holds true especially for the grade I bypass group, where the lowest BVF values were obtained. The measured volume flow through the bypass nevertheless reached as much as 48 ml/min. Thus despite only “poor” intracranial opacification there is considerable volume flow via the anastomosis, which appears to be higher than that described earlier [43, 44]. Patients with grade II and grade III bypass patency showed, in addition to a significant mean difference between groups, a wide range of BVF values within the groups. Thus, in contrast to patients with grade I graft patency, direct prediction of angiographic bypass grade II and III was not possible by MRA. The wide range of BVF values within the groups clearly demonstrates that angiographic, i.e., qualitative, assessment of graft patency does not necessarily reflect blood flow through the EC/IC bypass. These findings could be related either to the limitation of angiography, i.e., the necessity of a contrast media and variable application pressure, the subjective nature of the angiographic grading scale, or differences in blood requirement within the affected vascular territory. The BVF through the EC/IC bypass in these patients was not related to the degree of CVRC impairment in the preoperative rCBF studies. Follow-up investigations are required to determine whether bypasses with relatively low BVF but “good” and “extensive” intracranial filling tend to “mature,” i.e., to deliver more blood to the brain, over time, or whether a secondary bypass failure is likely to occur in the latter course.

Since two-dimensional cine-phase MRI has been validated in both phantom as human studies, it is extensively used to determine BVF in cerebral vessels. In the present study this technique allowed reliable quantitative assessment of BVF in a established EC/IC anastomosis under physiological conditions. Based on its apparent advantages, our results demonstrate the potential of MRA-based imaging techniques to study even very small arterial blood vessels. In addition, the technique is superior to CA for physiological investigation of the bypass. However, as for angiography, the contribution of the additional BVF via the anastomosis to restoring rCBF within the revascularized territory cannot be determined by this method.

In conclusion, use of an angiographic three-point grading system to assess EC/IC bypass patency allows estimation of BVF reliably, while quantitative BVF measurements with MRA provide more exact, investigator-independent information. This warrants the application of MRA to study the effects of EC/IC bypass surgery on cerebral hemodynamics in both the short and the long term.