Classification of internal carotid artery aberrations based on CTA imaging: evaluating the risk of ICA injury in the transoral approach to upper cervical surgery

Abstract

Objective

Cervical internal carotid artery (ICA) aberrations increase transoral atlantoaxial surgery risk. Existing classifications ignore the ICA’s relationship with the atlanto-axial complex. This study aimed to define this relationship using CTA and assess associated injury risk.

Methods

The study included 846 patients underwent CTA from January 2019 to February 2024 in our medical center. These classifications were based on the position of aberrant ICAs relative to the odontoid processes and the lateral masses of the atlas.

Results

In a cohort of 846 patients with a total of 1,692 ICAs, 65 patients (7.68%) exhibited aberrations, including 27 males (41.54%) and 38 females (58.46%). The incidence of ICA aberrations was comparable between patients under 60 years (6.67%) and those aged 60 years or older (8.02%) (P > 0.05). Bilateral aberrations were significantly more common than unilateral ones (64.62% vs. 35.38%, P < 0.05). We identified a total of 108 abnormal ICAs (6.38%), with 55 on the left side and 53 on the right (50.93% vs. 49.07%, P > 0.05), showing no significant difference. According to our proposed classification, there were 65 cases of Type I, 32 cases of Type II, 9 cases of Type III, and 2 cases of Type IV. The prevalence of Subtype A aberrations (82 cases, 75.93%) was significantly higher than that of Subtype B (26 cases, 24.07%) (P < 0.01).

Conclusions

Aberrations in the course of the ICA at the upper cervical level are relatively common, presenting a potential risk of ICA injury during the transoral approach for upper cervical surgery. Therefore, our classification of ICA aberrations based on CTA should be considered a standard component of preoperative evaluation.

Introduction

After branching off from the common carotid artery, the internal carotid artery (ICA) ascends through the parapharyngeal space to the external opening of the carotid canal. Variations in this segment of the ICA are relatively common in European and American populations, affecting 10–40% of individuals undergoing angiography [1]. Such variations can cause the ICA to deviate laterally towards the midline of the pharynx, adhering closely to the pharyngeal mucosa. This proximity poses a significant risk of ICA injury and potentially fatal massive hemorrhage during transoral procedures. Historical reports from early 20th-century otolaryngologists describe acute hemorrhage and lethal complications following injuries to aberrant ICA segments along the pharyngeal wall. This issue has captured the attention of otolaryngologists, leading to the development of a clinicoradiological classification system based on the morphology of the anomalous ICA [2–5]. For spinal surgeons, the transoral approach to the upper cervical spine is crucial for addressing specific conditions such as irreducible atlantoaxial dislocations (IAAD), tumors, and surgeries requiring odontoid process resection [6–9]. Traditionally, due to the technical challenges and lower frequency of these surgeries, orthopedic surgeons have infrequently assessed the impact of ICA variations before this type of surgery, especially since the normal ICA trajectory is typically further from the midline at the C0 to C4 level. However, with advancements in upper cervical surgical techniques and the increasing prevalence of transoral operations, including release, decompression, and fixation, there has been a surge in reports of intraoperative ICA injuries [10]. The existing Pfeiffer classification in otolaryngology does not adequately evaluate the potential risk of ICA injury in orthopedic procedures. Therefore, we conducted a retrospective analysis of CTA images of the ICA at the upper cervical level, comparing its path and position with the atlantoaxial bony structures, and have proposed a new clinical imaging classification for use in transoral atlantoaxial surgeries.

Materials and methods

Study subjects

We conducted a retrospective analysis of imaging data from patients who underwent computed tomography angiography (CTA) of the supra-arch artery at our medical imaging center between January 2019 and February 2024. The cohort consisted of 597 males (70.57%) and 249 females (29.43%), ranging in age from 16 to 96 years, with a mean age of 64.50 ± 10.77 years. This study protocol has been approved by the Ethics Committee (2024B136). Informed consent and signed consent forms have been obtained from the patients for the relevant case materials presented in this study. The patient consented to publication of his/her image.

Image processing methods

The TOSHIBA Aquilion ONE + 320-slice spiral CT scanner was used for these examinations. Patients were positioned supine and underwent a preliminary non-contrast scan followed by a contrast-enhanced scan. Post-scan, the original images were processed through thin-slice reconstruction to generate detailed CTA images. Both a radiologist and a spine surgeon jointly reviewed and analyzed the images of the ICAs. ICAs that deviated from their typical paths were classified as abnormal, and the relationship between the aberrant ICA and the atlanto-axial complex was carefully assessed.

Classification method

Confirmation of axial lines on CTA coronal images

For transoral atlantoaxial surgery, recognizing that greater deviation of the ICA towards the cephalic and midline correlates with increased risk of vascular injury, we defined six vertical and one horizontal axis on the CTA coronal images pertinent to the anatomical landmarks for anterior atlantoaxial surgery. The horizontal axis (H line) connects the lower ends of the bilateral lateral masses of the atlantoaxial joint. The vertical axes include: the α line passing through the midpoint of the lateral mass of the atlas perpendicular to the H line (αR and αL for right and left, respectively); the β line passing through the lateral edges of the odontoid processes perpendicular to the H line (βR and βL for right and left, respectively); and the γ line passing through the outer edges of the lateral mass of the atlas perpendicular to the H line (γR and γL for right and left, respectively). According to standard anatomical texts, the cervical ICA normally ascends straight to the skull base without branching and is positioned posterolateral to the pharyngeal wall [3, 11]. An ICA running within the γ line is considered abnormal.

Classification of aberrant ICAs

Aberrations are classified into four types based on the farthest projection of the ICA towards the contralateral side in the coronal CTA images at the atlantoaxial level. Type I: Located outside the α line (midline of the atlas lateral masses) on the same side; Type II: Located between the α and β lines (midline of the lateral masses and the outer edge of the odontoid process) on the same side; Type III: Located between the bilateral β lines (lateral edges of the odontoid processes); Type IV: Exceeds the contralateral β line. Additionally, based on the position of the upper edge of the most protruding portion of the anomalous internal carotid artery (ICA) relative to the H line within the projection area of the atlantoaxial region, each classification is further divided into subtypes A and B: Subtype A indicates that the upper edge of the most protruding portion of the anomalous ICA is above the H line, while Subtype B indicates that it is below. Detailed classification methods and schematic illustrations are presented in Tables 1, 2, and Fig. 1, respectively.

Table 1.

Classification of aberrant ICA at the upper cervical level on coronal CTA images

Classification	The course of ICA deviating toward the midline on the projection of the atlantoaxial region
Type I	Between the α and γ line
Type II	Between the α and β lines
Type III	Between the bilateral β lines
Type IV	Exceeds the contralateral β line

Table 2.

Subtypes of aberrant ICAs at the upper cervical level on coronal CTA images

Subtype	Position of the upper edge of the most protruding portion of the aberrant ICAs
Subtype A	Above the H line
Subtype B	Below the H line

Fig. 1.

Schematic Illustrations of the classification of aberrant ICAs. A: Normal ICA, coursing outside the γ line; B: Type I: aberrant ICA located between the α and γ line on the same side; C: Type II: aberrant ICA located between the α line and β line on the same side; D: Type III: aberrant ICA located between the bilateral β lines; E: Type IVB: aberrant ICA exceeding the contralateral β line and beneath the H line; F: Type IVA: aberrant ICA exceeding the contralateral β line and above the H line

Statistical analysis

All data were processed using SPSS 25.0 software, and chi-square tests were employed to evaluate whether there were statistical differences in ICA aberrations among general data such as gender and laterality. The significance level was set at P = 0.05.

Results

Results of ICA aberration classification

Among 846 patients, 65 (7.68%) presented with ICA aberrations, and 108 out of 1692 ICAs (6.38%) exhibited aberrations. Among the 65 patients with ICA aberrations, there were 27 males and 38 females (15.26% vs. 4.52%, P < 0.05), suggesting that the incidence of ICA aberrations was significantly higher in females than in males. The incidence of ICA aberrations showed no significant difference between patients under 60 years (6.67%, 14/210) and those aged 60 years or older (8.02%, 51/636) (P > 0.05). Typical images of each type of aberrant ICAs are depicted in Fig. 2. The incidence of unilateral aberrations (23 cases, 35.38%) was significantly lower than that of bilateral aberrations (42 cases, 64.62%, P < 0.05). There were 55 left-sided aberrations and 53 right-sided aberrations (50.93% vs. 49.07%, P > 0.05), demonstrating no statistical significance. Subtype A aberrations accounted for 82 cases, while Subtype B accounted for 26 (75.93% vs. 24.07%, P < 0.05), indicating that Subtype A aberrations were significantly more prevalent than Subtype B. The statistical results are presented in Tables 3, 4 and 5.

Fig. 2.

Typical CTA Images of different ICA Aberration Classification (The vessel indicated by the red arrow is ICAs). A: Bilaterally Normal; B: Bilateral Type ⅠA; C: Right Type IB, Left Normal; D: Right Type IIA, Left Type IA; E: Right Type IIB, Left Normal; F: Right Type IIB, Left Type IIIA; G: Right Type IIIB, Left Type IA; H: Right Type IA, Left Type IVA; I: Right Type IIIA, Left Type IVB

Table 3.

Gender differences in patients with ICA aberrations

	Number of patient with aberrations	Proportion (%)	Proportion in the total study population (%)
Male	27	41.54	4.52
Female	38	58.46	15.26
Difference (P Value)	P > 0.05	P > 0.05	P < 0.05

Table 4.

Laterality differences in patients with ICA aberrations

	Number of aberrations	Proportion (%)	Difference(P Value)
Left side	55	50.93	P > 0.05
Right side	53	49.07
Unilateral	23	35.38	P < 0.05
Bilateral	42	64.62
Type A	82	75.93	P < 0.01
Type B	26	24.07

Table 5.

Specific classification results of aberrant ICAs

Classification	Left side	Right side	Total	Proportion (accounting for the total number of abnormal ICAs)
ⅠA	35	25	60	55.56%
ⅠB	1	4	5	4.63%
ⅡA	10	9	19	18.52%
ⅡB	3	10	13	12.04%
ⅢA	1	1	2	1.85%
ⅢB	3	4	7	6.48%
ⅣA	1	0	1	0.93%
ⅣB	1	0	1	0.93%

2.2 illustrative case (see Fig. 3)

Fig. 3.

A case of IAAD undergoing combined anterior and posterior cervical surgery. A: Preoperative lateral cervical X-ray shows atlantoaxial dislocation; B: Preoperative X-ray after skull traction shows no significant relief of dislocation; C: Preoperative CT shows atlantoaxial dislocation with mild basilar invagination; D: Preoperative MRI shows spinal cord compression due to atlantoaxial dislocation; E: CTA shows a Type IIIB ICA aberration on the right side, with the left side normal (The vessel indicated by the red arrow is ICAs); F: Preoperative open-mouth position for transoral atlantoaxial surgery; G: Intraoperatively lateral mass joint release using a curved bone chisel (red circle); H-I: Postoperative lateral cervical X-ray and CT show complete reduction of atlantoaxial dislocation; J: Postoperative MRI shows decompression of cervical spinal cord; K: Postoperative CT shows the cage (red circle) implantation between the lateral mass joints; L: Lateral cervical X-ray shows desirable maintenance of reduction at 24 weeks follow-up ; M: CT shows satisfied bony fusion of the atlantoaxial lateral mass joints at 24 weeks follow-up (red circle)

A 65-year-old female patient with IAAD underwent transoral joint release, posterior reduction and fixation, followed by cage implantation and bone graft fusion between the posterior atlantoaxial lateral mass joints. Preoperative CTA assessment demonstrated a Type IIIB ICA aberration on the right side, while the left side was normal, classified as a moderate risk aberration.

Discussion

The transoral approach is a standard procedure for addressing complicated atlantoaxial dislocations and upper cervical disorders. It enables the direct resection of scarring, bone spurs, and contracted ligaments in front of the atlanto-axial complex, facilitating subsequent release of the lateral mass joint and reconstruction of the upper cervical spine via internal fixation. The anatomical layers of this approach are simple; after incising the posterior pharyngeal wall, slight dissection leads directly to bony structures, and typically, no important artery traverses the surgical area under normal anatomy. However, if the ICA, which normally runs through the surgical field, shows aberrations, it can be directly severed during the incision of the posterior pharyngeal wall, or during the use of a bone chisel to release the upper cervical joint, remove the dens, and the process of implanting screws in the axis, leading to fatal consequences. Previous reports on vascular injury during upper cervical surgery have primarily focused on vertebral artery injuries [12, 13], and there are only a few scattered reports of cases where ICA aberrations resulted in surgical failures in upper cervical procedures. Most of these cases focus on ICA injuries resulting from posterior cervical screw placement [14, 15]. Nevertheless, ENT surgeons using the same transoral approach have reported instances of ICA injuries occurring during procedures. Mckenzie reported two cases of ICA injury during adenoidectomy that led to fatalities, one of which was a complete transection of the ICA [16], highlighting that orthopedic surgeons also face a potential risk of ICA injury during transoral upper cervical procedures. However, perhaps due to the historically lower prevalence of this approach in spinal surgery, reports have been scarce.

For surgical procedures on the posterior pharyngeal wall, Pfeiffer established an imaging classification system based on two criteria: the plane of ICA anomalies and the shortest distance between the anomalous ICA and the pharyngeal mucosa [17]. He proposed a risk assessment method for ICA injury during transoral otolaryngological surgeries. For transoral atlantoaxial surgery, as long as a longitudinal incision can be safely completed near the midline, after subperiosteal dissection, the superficial tissues of the posterior pharyngeal wall can be moderately dissected to both sides. However, for the subsequent bony structure operation, strong prying with an osteochisel is required, even high-speed electric tools like a grinding drill are usually used in a relatively large area within the outer edges of the bilateral lateral mass joints. The course and position of the ICA has a greater impact on the safety of the ICA during operations than the depth of its course level. Our analysis delineates a clear epidemiological and morphological profile of ICA aberrations pertinent to transoral surgery. The significant predilection for female patients underscores the influence of demographic factors. While a higher incidence was noted in patients aged ≥ 60 years, this difference was not statistically significant in our cohort; nevertheless, this trend aligns with the existing literature suggesting that age-related degenerative changes can be a driver of aberrant vasculature [18]. We concur that the underlying mechanism is likely rooted in age-related cervical degeneration. As insightfully suggested, degenerative processes such as disc space narrowing and spondylotic shortening lead to a reduction in vertical height of the cervical spine. This, in turn, can induce a relative redundancy and consequent tortuosity of the ICA, predisposing it to medial and anterior displacement toward the midline—a morphological change that our classification system is specifically designed to capture. Therefore, in clinical practice, careful evaluation of the ICA trajectory should be incorporated into the preoperative management of older patients to mitigate injury risk. Morphologically, the high prevalence of bilateral involvement highlights the necessity for comprehensive bilateral preoperative assessment. Furthermore, the predominance of Subtype A aberrations provides critical anatomical reassurance, suggesting that in the majority of cases, the highest risk segment of the ICA may be situated above the level of the typical atlantoaxial working corridor. However, Pfeiffer’s classification relies solely on the distance between the pharyngeal mucosa and the ICA as its fundamental criterion, lacking a detailed examination of the relationship between the atlanto-axial complex, and the ICA, which significantly limits its relevance for spinal surgeons.

A review of orthopedic literature shows that only Wang et al. explicitly state, in their research on the treatment for “sandwich” atlantoaxial dislocation, the necessity of preventing ICA damage, and they conducted a zonal classification of the ICA position on the coronal plane at the upper cervical level [19]. However, this classification focuses only on assessing the relationship between posterior screw fixation techniques and the ICA position, without a detailed discussion of anterior surgical approaches via the oral pharynx, which are more directly related to the ICA. Furthermore, their classification method does not provide descriptions or classifications regarding the horizontal position of the ICA in relation to the atlanto-axial complex. Although non-ionizing Magnetic Resonance Angiography (MRA) can demonstrate the course of the ICA, it primarily visualizes the vascular lumen while providing minimal to no imaging of surrounding bony structures. This limits the direct assessment of the anatomical relationship between the ICA and upper cervical structures such as the atlas and axis, requiring indirect inference from soft-tissue imaging of Magnetic Resonance Imaging(MRI). In contrast, three-dimensional reconstruction of CTA clearly visualizes the ICA and its surrounding structures, establishing it as a critical diagnostic modality for identifying aberrant ICA courses [16]. Our study conducted a retrospective review of CTA images of 1,692 ICAs from 846 patients over a span of 60 months at a single medical center, based on the practical requirements for performing upper cervical release, decompression and fixation surgeries through the transoral approach, assessing the abnormal course of the ICA in two dimensions: vertical and horizontal positions relative to the atlanto-axial complex, and proposing a classification method for ICA aberrations, while also providing a preliminary discussion of the surgical risks associated with different types of ICA aberrations via the transoral approach, as shown below. (An overview of the proposed classification and risk assessment associated with different surgical techniques is given in Table 6)

Table 6.

Risk assessment of aberrant ICA injury based on different types of transoral upper cervical surgeries

Types of aberrant ICA		Potential risk for ICA injury
		Surgical techniques
One side	Contralateral side	Atlantoaxial release	Atlantoaxial release combined with anterior fixtion
Normal/ⅠA/ⅠB	ⅠA/ⅠB	Low	Low
ⅡB	Normal/ⅠA/ⅠB/ⅡB	Low	Moderate
Normal/ⅠA/ⅠB	ⅡA/ⅢA/ⅢB	Moderate	Moderate
ⅡA	ⅡA/ⅡB/ⅢB	Moderate	Moderate
ⅢA	ⅡB/ⅢB	Moderate	Moderate
ⅢB	ⅡB/ⅢB	Moderate	High
ⅣB	Normal/ⅠA/ⅠB/ⅡB/ⅢB	Moderate	High
ⅢA	ⅡA/ⅢA/ⅣA/ⅣB	High	High
ⅣA	Normal/ⅠA/ⅡA/ⅡB/ⅢB/ⅣA/ⅣB	High	High
ⅣB	ⅡA/ⅣB	High	High

Type I (see Fig. 2A, B)

Types IA and IB: For patients undergoing atlantoaxial reduction or combined anterior fixation with odontoidectomy, although the ICA aberration is cohesive unilaterally or bilaterally, they are located at a greater distance from the midline, and thus, the risk of ICA injury is lower. A midline incision remains feasible, but care should be taken to protect the aberrant ICA during lateral mass joint release.

Type II (see Fig. 2C, D)

Type IIA: For unilateral IIA, although it is closer to the midline, the contralateral ICA is situated further from the midline, providing a relatively large avascular zone. A longitudinal incision can be made on the side that is closer to the normal side, following dissection, the tissue of the posterior pharyngeal wall involving the aberrant ICA is capable of being retracted. In bilateral IIA, the distance from the midline is very close, and the safe avascular region along the midline is minimal. The risk of ICA injury is considered moderate. Caution should be exercised when making a midline incision that directly reaches the anterior arch of the atlas. Subsequently, perform subperiosteal dissection to allow the release and retraction of the ICA before proceeding with further manipulation.

Type IIB: Since the cohesive ICA is situated below the level of the atlantoaxial joint, the risk of ICA injury during atlantoaxial release operation is relatively low. Both unilateral and bilateral IIB can approach the joint space level from a midline incision. However, when dealing with the atlantoaxial joint, care must be taken to safeguard the aberrant ICA located beneath it. If the surgical scope requires extension below the level of the atlantoaxial joint, (for example, when anterior fixation of the axis is necessary), collaboration with ENT specialists to release the ICA before proceeding with atlantoaxial surgery is advisable.

Type III (see Figs. 2E, F)

Type IIIA: The risk of ICA injury for unilateral IIIA is high. Incision on the side biased towards normal can be used to perform subperiosteal dissection and release the ICA prior to further procedures. Bilateral IIIA (this study did not encounter this type, but patients’ imaging data from the literature regarding atlantoaxial surgeries can be classified under this type [7]) poses a high risk of vascular injury. If the surgical field needs to extend below the level of the atlantoaxial joint, (for example, when anterior fixation of the axis is necessary), it is advisable to collaborate with ENT specialists to free the ICA before proceeding with atlantoaxial surgery, or to consider using an alternative surgical approach.

Type IIIB: If only atlantoaxial joint release is done through the transoral approach, there is a moderate risk of vascular injury (see Figur 3E of the illustrative Case). Attention should be given to safeguard the aberrant ICA located below when dealing with the lateral mass joint. If the surgical field needs to extend below the level of the atlantoaxial joint, (for example, when anterior fixation of the axis is necessary), collaboration with ENT specialists to free the ICA before continuing with atlantoaxial surgery is required, or consider alternative surgical approaches.

Type IV (see Figs. 2G, H)

Type IVA: There is a high risk of ICA injury. It is advisable to collaborate with ENT specialists to release and protect the ICA and then proceed with atlantoaxial surgery, or consider alternative surgical approaches.

Type IVB: If only atlantoaxial joint release is done through the transoral approach, there is a moderate risk of ICA injury. A longitudinal incision can be made on the side that is closer to the friendlier side, but when addressing the atlantoaxial lateral mass joint, careful subperiosteal dissection should be done, taking care to protect the aberrated ICA below. If surgical procedures need to be conducted at the C2 vertebral level (such as axis fixation), collaboration with ENT specialists to release and protect the ICA is necessary before proceeding with surgery, or consider alternative surgical approaches.

Limitations

This study only covers a preselected population; all cases included were sourced from a single medical center over 60 months, including Neurology, Neurosurgery, and Orthopedic inpatients and outpatients. Related literature similarly selects subjects from disciplines that have no clear correlation with ICA aberrations and do not result in an elevated incidence of such aberrations. Therefore, although this study aims to estimate the prevalence of cervical ICA aberrations in the general population, it cannot completely reflect the characteristics of the normal population either. Secondly, the assessment of the potential ICA injury risk related to our ICA aberration classification has a limited number of practical cases. Hence, the classification method we have proposed requires the accumulation of additional surgical cases for validation and refinement.

In addition, the CTA evaluations were performed exclusively in the neutral neck position, but transoral surgeries are conducted with the neck extended, a posture that shortens the anterior cervical column and may displace the ICA anteriorly and medially, particularly in already tortuous vessels [10, 11]. Future studies employing dynamic CTA in the flexed and extended positions are warranted to quantitatively assess these positional changes and validate or refine our risk classification under physiologically relevant conditions.

Conclusion

In summary, the classification of ICA aberrations based on CTA imaging is of crucial importance in assessing the risk of potential ICA injury during transoral upper cervical surgeries [20–22]. This retrospective study based on CTA imaging reveals that ICA aberrations are not uncommon (occurring rate 7.68%), with a higher prevalence in female patients and a predominance of bilateral aberrations. During the preparation for the transoral approach to atlantoaxial surgery, it is essential to consider ICA aberrations, similar to the evaluation of vertebral arteries in the posterior approach for atlantoaxial procedures. The aberrant ICA classification derived from CTA imaging should be regarded as a standard preoperative examination for evaluating the risk of potential ICA injury during surgery.

Abbreviations

ICA

Cervical Internal Carotid Artery

IAAD

Irreducible Atlantoaxial Dislocations

CTA

Computed Tomography Angiography

MRA

Magnetic Resonance Angiography

MRI

Magnetic Resonance Imaging

Author contributions

QS conceived the study, curated the data, drafted the original manuscript, supervised the research, designed the methodology, and provided resources. YX secured funding, contributed to conceptualization, and reviewed/edited the manuscript. SM curated data, developed methodology, and implemented software. HW curated data, contributed to methodology, and edited the manuscript. HQ curated data, assisted in methodology, and implemented software. JG conducted formal analysis and investigation, and curated data. HS performed formal analysis and investigation. XT curated data and conducted formal analysis. XX contributed to data reprocessing, formal analysis, literature review, and assisted in drafting the response to reviewers. All authors read and approved the final manuscript.

Funding

This study was supported by Tianjin Key Medical Discipline (Specialty) Construction Project (TJYXZDXK-064B).

Data availability

The datasets used and analyzed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

The Declaration of Helsinki was strictly followed throughout the study. This study was approved by the ethics committee of Tianjin Union Medical Center (2024B136). The voluntary subjects provided written informed consent.

Consent for publication

Consent for publication of identifiable data was obtained from all individual participants included in the study.

Competing interests

The authors declare no competing interests.