Introduction

The craniovertebral junction is anatomically complicated and the location of several common congenital or acquired osseous lesions. Some types of spinal tumor also occur in this area relatively frequently. Tokuda and colleagues [1] reviewed selective vertebral artery (VA) angiograms of 300 patients and reported three types of VA variation at the C1–2 level: (1) VA running in the spinal canal below C1, indicative of persistent first intersegmental artery (FIA); (2) fenestration of the VA above and below C1; and (3) posterior inferior cerebellar artery (PICA) originating from the C1/2 level. Using computed tomographic angiography (CTA), Yamazaki’s group [2, 3] reported increased frequency of VA variations at the C1–2 level in patients with osseous anomalies at the craniovertebral junction or patients with Down syndrome. CTA is useful for evaluating both arterial variations and osseous lesions, but it is more invasive than magnetic resonance angiography (MRA). Conventional angiography may only be useful when an interventional procedure is planned. To avoid complication during surgery, it is important to identify VA variations at the C1–2 level. We describe the prevalence of these VA variations on MRA in what we believe is the first report of VA variations at the C1–2 level studied in a large MRA series.

Materials and methods

In accordance with the policies for exemption set by our internal institutional review board, we retrospectively reviewed MRA images of the supra-aortic region from 5,319 such examinations performed in 3,626 patients in our institution from April 1, 2007 through December 31, 2009. The majority of the examined patients had or were suspected of having ischemic cerebrovascular diseases. Among these patients, 625 did not undergo examination of the extracranial region. Of the remaining 3,001 patients, we excluded 262 with unilateral or bilateral VA occlusion or suboptimal image quality, ultimately reviewing and analyzing MRA images of 2,739 patients (1,615 men, 1,124 women; aged 6 months to 98 years, mean age, 64.0 years). We did not obtain informed consent because of the retrospective nature of the clinical study.

All patients were examined on one of two 1.5-T MR systems (Achieva Nova Dual, Philips Medical Systems, Best, The Netherlands and Magnetom Avanto, Siemens Medical Systems, Erlangen, Germany). The standard three-dimensional time-of-flight MRA protocol was repetition time (TR), 20 ms; echo time (TE), 6.0 ms; flip angle, 16°; field of view (FOV), 20 × 20 cm; and slab thickness, 0.65 mm for Achieva Nova Dual (acquisition time; 5 min 30 s) and for Magnetom Avanto, TR, 22 ms; TE, 7.0 ms; flip angle, 18°; FOV, 22 × 22 cm; and slab thickness, 0.7 mm, with acquisition time of 7 min 34 s. Contrast-enhanced MRA images were not included.

The first author, an experienced neuroradiologist, retrospectively reviewed all MRA images with special attention to VA course and branching at the level of the C1–2 vertebral bodies. Only maximum intensity projection (MIP) images with 180° horizontal rotation (20 images) were reviewed using viewers. We used the SYNAPSE (Fujifilm Medical Company, Tokyo, Japan) picture archiving and communication system to evaluate images. Additionally, we checked MR angiographic source images and made partial MIP images in case it was necessary to identify and demonstrate the variation clearly. The third author performed statistical analysis using chi-square test.

Results

We observed three types of VA variation at the C1–2 level, as reported previously by Tokuda’s group [1]: (1) persistent FIA without normal VA branch, (2) VA fenestration, and (3) PICA of C1/2 origin. We diagnosed at least one of the three types of VA variation in 138 of 2,739 patients, a prevalence of 5.0%, and detected the variations bilaterally in 15 of the 138. There was no laterality in frequency, but we found female predominance (P < 0.05, chi-square test). We most frequently observed persistent FIA (87 patients, 3.2%, Fig. 1), 11 of which were bilateral (Fig. 2). We also observed two VA fenestrations and two PICAs of C1/2 origin on the contralateral side. We found almost equal prevalence of VA fenestration (Fig. 3; 0.9%) and PICAs of C1/2 origin (1.1%) and observed only one PICA of C1/2 origin bilaterally (Fig. 4). Two PICAs of C1/2 origin had no normal VA branch (Fig. 4). Table 1 summarizes results. We observed no aneurysm in this region. An occipital artery arose from the left VA at the craniovertebral junction, and there was a case of type 2 proatlantal intersegmental artery on the left side.

Fig. 1
figure 1

A 69-year-old woman with a syncopal attack. The anteroposterior (a) and lateral (b) projections of partial maximum intensity projection magnetic resonance angiographic images of the vertebrobasilar system show that the left vertebral artery (VA) takes an abnormally low course at the C1/2 level (arrows). The left posterior inferior cerebellar artery (PICA) is not identified, but the right PICA is arising at the level of the foramen magnum. Computed tomography angiography (CTA) from the aortic arch to the intracranial region was added to evaluate posterior circulation. The lateral projection of the CTA of the cervical level (c) shows the left VA entering the spinal canal through the C1/2 intervertebral space (arrow), indicative of a persistent first intersegmental artery

Full size image
Fig. 2
figure 2

An 83-year-old woman with subarachnoid hemorrhage. The AP projection of a partial MIP MRA image of the vertebrobasilar system shows bilateral persistent FIAs (arrows). The terminal segment of the right VA is hypoplastic or occluded

Full size image
Fig. 3
figure 3

A 51-year-old woman with right VA dissection. The AP projection of a partial MIP MRA image of the vertebrobasilar system shows fenestrated left VA and right persistent FIA (arrows). There is fusiform dilatation of the terminal segment of the right VA

Full size image
Fig. 4
figure 4

A 78-year-old man with cerebral infarctions. The AP (a) and right anterior oblique (b) projections of partial MIP MRA images of the vertebrobasilar system show that the left PICA is arising from the VA at the level of C1/2 (long arrow). The right PICA is also arising from the same level, but there is no normal branch of the right VA (short arrow)

Full size image
Table 1 VA variations at the C1–2 level diagnosed by MRA among 2,739 patients (male, 1,615; female, 1,124)
Full size table

Discussion

According to Padget [4], at 5 to 6 weeks’ gestation, the VAs develop from plexiform anastomoses between the seven embryonic cervical intersegmental arteries. The seventh intersegmental arteries enlarge and become the subclavian arteries. The VAs arise from the enlarging seventh cervical intersegmental arteries and supply the caudal end of the longitudinal neural arteries that develop to the basilar artery. If the FIA persists without persistence of the normal VA branch, the VA takes an anomalous course that enters the spinal canal at the level between the C1 and C2 vertebral bodies. Figure 5 presents a schematic illustration of this variation. Called persistent FIA, this is the variation we most frequently observed; its prevalence on MRA was approximately 3%, which was five times higher than that of reported angiographic prevalence [1, 5]. Without careful observation, persistent FIA may easily be overlooked because associated anomalous branches are not seen in this variation. Bilateral persistent FIAs are not so rare and may be more easily overlooked because the variation is symmetrical. Routine review of MR angiographic source images might permit easy identification of this variation that passes through the C1/2 intervertebral space.

Fig. 5
figure 5

Schematic illustration of arterial variations of the VA at the C1/2 level (left lateral projection). a Persistent FIA, b fenestration of the VA above and below C1, c PICA originating from the C1/2 level

Full size image

Extracranial fenestration is a well-known variation of the VA [6] and is most frequently seen at the C1–2 level. This variation represents persistence of the normal VA branch together with persistence of the FIA. Its reported angiographic prevalence is about 1% [1], similar to its prevalence in our MRA series (0.9%). Double fenestration of the extracranial VA on the same side is rarely reported [7]. Bilateral extracranial VA fenestrations are also rarely seen [8]. However, in our large MRA series, we observed neither double nor bilateral VA fenestrations. If the variation occurs independently on either side, the prevalence of bilateral extracranial VA fenestrations is suggested to be 0.01%.

If the FIA persists and continues to the PICA without fusion with the VA, it forms the extracranial C1/2 origin of the PICA. An extracranial origin of the PICA is seen in 5% to 20% of cases [9]. However, in the majority of cases, the PICA arises from the VA at the level of the craniovertebral junction (Fig. 1a). Sato and associates [5] reported angiographic prevalence of this variation to be only 0.36%. If contrast medium is injected unilaterally, variations of the contralateral VA cannot be detected. In our large MRA series, the prevalence of the PICA of C1/2 origin was 1.1%. Because the PICA is a relatively small vessel, it may not be identified on MRA in some patients. We observed one patient with bilateral PICAs of C1/2 origin. We also found two cases of PICA of C1/2 origin without a normal VA branch; to our knowledge, this variation has not been reported. Rarely, an aneurysm in the spinal canal can be seen on this anomalously originating PICA at its proximal segment [10, 11]. In our series, there was no such case, but we experienced a case of rare variation − type 2 proatlantal intersegmental artery.

Most variations of the VA are asymptomatic and found incidentally. Thus, the clinical importance of these variations is limited. Nevertheless, Takahashi and colleagues [12] reported a case of bilateral persistent FIAs that caused cord compression myelopathy. Other clinical presentations can include cervical pain, occipital neuralgia, and accessory nerve palsy [12]. If craniovertebral junction surgery is planned, preoperative identification of these variations is very important to avoid complications, such as arterial injury, during surgical intervention. In addition, these VA variations, especially the persistent FIA, are dangerous if lateral puncture of the C1/2 is attempted for myelography because the needle direction is superimposed on the large anomalous artery that runs in the C1/2 intervertebral space [13].

Yamazaki’s group [2, 3] reported the usefulness of CTA in diagnosing these variations because both arterial and osseous images can be obtained simultaneously (Fig. 1c). Although osseous clear images cannot be obtained, MRA can be more easily and noninvasively performed than CTA. To our knowledge, this is the first report of VA variations at the C1–2 level studied in a large MRA series. CTA may be required in cases suggestive of osseous lesions. Conventional angiography may have limited indication in cases with planned interventional procedures.

Our investigation is limited because it is a retrospective clinical study, and only one neuroradiologist analyzed the MRA images. In addition, there may have been some selection bias in our study materials and results because the majority of the examined patients had or were suspected of having ischemic cerebrovascular diseases.

Conclusion

We observed three types of VA variation at the C1–2 level, an overall prevalence of 5.0%. There was female predominance; the persistent FIA was most frequently observed, sometimes bilaterally; and PICAs of C1/2 origin rarely arose without a normal VA branch. To avoid complications during surgery at the craniovertebral junction, these VA variations should be correctly identified preoperatively.