If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
e Jinsong Wang and Yonghui Li contributed equally to this paper.
Jinsong Wang
Footnotes
e Jinsong Wang and Yonghui Li contributed equally to this paper.
Affiliations
Division of Vascular Surgery, National-Guangdong Joint Engineering Laboratory for Diagnosis and Treatment of Vascular Diseases, the First Affiliated Hospital of Sun Yat-sen University, 58 Zhongshan Road 2, Guangzhou 510080, ChinaDepartment of Vascular Surgery, Guangdong Provincial People’s Hospital, 106 ZhongShan Road 2, Guangzhou 510080, China
Division of Vascular Surgery, National-Guangdong Joint Engineering Laboratory for Diagnosis and Treatment of Vascular Diseases, the First Affiliated Hospital of Sun Yat-sen University, 58 Zhongshan Road 2, Guangzhou 510080, China
Division of Vascular Surgery, National-Guangdong Joint Engineering Laboratory for Diagnosis and Treatment of Vascular Diseases, the First Affiliated Hospital of Sun Yat-sen University, 58 Zhongshan Road 2, Guangzhou 510080, ChinaDepartment of Breast and Thyroid Surgery, the First Affiliated Hospital of Sun Yat-sen University, 58 Zhongshan Road 2, Guangzhou 510080, China
Division of Vascular Surgery, National-Guangdong Joint Engineering Laboratory for Diagnosis and Treatment of Vascular Diseases, the First Affiliated Hospital of Sun Yat-sen University, 58 Zhongshan Road 2, Guangzhou 510080, ChinaDepartment of Breast and Thyroid Surgery, the First Affiliated Hospital of Sun Yat-sen University, 58 Zhongshan Road 2, Guangzhou 510080, China
Division of Vascular Surgery, National-Guangdong Joint Engineering Laboratory for Diagnosis and Treatment of Vascular Diseases, the First Affiliated Hospital of Sun Yat-sen University, 58 Zhongshan Road 2, Guangzhou 510080, China
Division of Vascular Surgery, National-Guangdong Joint Engineering Laboratory for Diagnosis and Treatment of Vascular Diseases, the First Affiliated Hospital of Sun Yat-sen University, 58 Zhongshan Road 2, Guangzhou 510080, China
To analyze the management of carotid body tumors (CBTs), particularly the use of preoperative embolization (EMB) and image features in minimizing surgical complications.
Background
CBT surgery is a challenging procedure, and the role of EMB in CBT surgery has remained unclear.
Methods
A total of 200 CBTs were identified among 184 medical records involving CBT surgery. Regression analysis was used to explore the prognostic predictors of cranial nerve deficit (CND), including image features. In addition, blood loss, operation times, and complication rates were compared between patients who had surgery only versus patients who had surgery along with preoperative EMB.
Results
Overall, 96 males and 88 females were identified for inclusion in the study, with a median age of 37.0 years. Computed tomography angiography (CTA) showed the presence of a tiny gap adjacent to the encasement of carotid vessels, which could help minimize carotid arterial injury. High-lying tumors that encased the cranial nerve were usually managed with synchronous cranial nerve resection. Regression analysis revealed that the incidence of CND was positively associated with Shamblin Ⅲ, high-lying, and a maximal CBT diameter of ≥ 5cm. Among 146 EMB cases, two cases of intracranial arterial embolization occurred. No statistical difference was found between the EBM and Non-EBM groups in terms of bleeding volume, operation time, blood loss, blood transfusion requirement, stroke, and permanent CND. Subgroup analysis revealed that EMB decreased CND in Shamblin III and low-lying tumors.
Conclusions
CBT surgery should be performed with preoperative CTA to identify favorable factors for minimizing surgical complications. Shamblin Ⅲ or high-lying tumors, as well as CBT diameter, are predictors of permanent CND. EBM does not reduce blood loss or shorten operation time.
Carotid body tumors (CBTs) are the most common among the head-and-neck paragangliomas that develop at the carotid bifurcation, with an incidence < 1:30,000 in the general population.(
) Slow-developing and often asymptomatic in nature, CBTs have been placed by the World Health Organization Classification of Head and Neck Tumors on a spectrum of metastatic potential(
), with early surgical intervention recommended if the patient is of acceptable operative risk.
CBT surgery has remained a challenging procedure due to the abundant vascularity of the tumor and its contiguity to the internal carotid artery and cranial nerves.(
) In particular, surgical expertise and comprehensive preoperative management are critical in cases where carotid arteries are completely encased. Therefore, preoperative embolization (EMB) has been developed to ensure that complications are minimized. Several early studies have suggested that this procedure could decrease blood loss, shorten operation times, and even reduce neurologic complications.
) Given the rarity of CBTs, preoperative EMB for CBT surgery remains a relatively novel approach.
In the past decade, an increasing number of CBT patients have undergone surgical resections, with a substantial proportion accepting preoperative EMB. Thus, this study was designed to review one hospital’s experiences with CBT surgery and compare the clinical outcomes between surgery alone and surgery with preoperative EMB.
Methods
Data collection
The Institutional Review Board of the First Affiliated Hospital of Sun Yat-sen University approved this retrospective study. Cases were identified from medical records spanning from May 1990 to December 2020 at the Division of Vascular Surgery, the First Affiliated Hospital of Sun Yat-sen University. CBT diagnostic criteria included clinical manifestations, preoperative imaging features, digital subtraction angiography, and histology. In addition, demographics, clinical presentations, co-morbidities, tumor characteristics, imaging reports, operative notes, pathologies, and follow-ups were tabulated.
, which assesses the relationship between CBTs and carotid vessels, was applied in this study. Tumors located above the second cervical vertebra were defined as “high-lying”; otherwise, tumors were considered to be “low-lying.” Tumor diameters were measured with preoperative CT scans. Postoperative cranial nerve deficit syndrome (CND), including transient and permanent CND, was defined as transient if resolved at a three-month follow-up. Recurrent loco-regional tumors were paraphrased as “Recurrence.” Paragangliomas originating from the vagal nerve were excluded from this study. Technical success was defined as a significant reduction in tumor blood supply (≥70%) and was independently evaluated by interventional radiologists after EMB.
Groups
Since 1996, arteriography was usually performed in CBTs with a maximal diameter ≥3 cm. All participants were divided into EMB and non-EMB groups. The criteria for the decision to perform EMB were as follows: (
Patients with clinical suspicion of CBT underwent cranial and carotid CTA. CTA was used to define the tumor’s size, location, Shamblin classification, and infiltration of neighboring tissues and local lymph nodes. A tiny space represented low-density and well-defined tumor boundary was also identified on CTA. The tiny space between the tumor and the artery by the cross-section surrounding the CBT (Fig. 1a, Illustrated) indicated that the tumor did not invade the arterial wall and thus was able to be resected without arterial compromise. On the other hand, absence of the tiny space (Fig.1b, Illustrated) indicated arterial invasion, which meant that arterial reconstruction would follow. Patency of the Circle of Willis was preoperatively evaluated using intracranial and carotid arterial CTA in all patients, especially for Shamblin Ⅲ tumors. Endocrine examinations were not routinely performed unless there was a co-existing pheochromocytoma. Positron emission tomography was reserved for malignant tumors.
Figure 1The value of preoperative CTA in evaluating whether a tumor invaded nearby arteries.Fig. 1a) The tiny space between the tumor and the artery by the cross-section along the whole tumor (arrow) indicates that the tumor did not invade the artery wall, meaning the tumor could be resected without arterial injury. Fig. 1b) If there is no space between the tumor and the artery (arrow), it means the tumor invaded the artery and that reestablishing the artery should be meticulously planned.
The recurrent laryngeal nerve was evaluated through vibration of the vocal cord on indirect laryngoscopy by an otolaryngologist. Two on-duty neurologists performed pre- and postoperative examinations to detect neurological deficits. Head CT scans were used to confirm suspected cerebral infarctions.
Preoperative embolization
EMB was generally performed under local anesthesia with transfemoral arterial access. Selective angiographic imaging of the internal, external, and common carotid arteries was used to identify the feeding arteries to ensure patency of the Circle of Willis. On angiography, signs of splaying of the carotid vessels and an intense blush in the tumor consolidated the diagnosis of CBT. Selective feeding artery angiographies aimed to delineate the anatomy of the tumor. Embolization was achieved using polyvinyl alcohol (PVA), microspheres, gel foam, and coils based on the characteristics of the feeding arteries. PVA was often used for the tumor vascular bed, and gel foam and coils were utilized for proximal larger branches of the feeding arteries. This study used PVA and gel foam preferentially.
Surgical resection
Seven surgeons performed the procedures within 72 hours after EMB. A trans-cervical approach was applied, and the incision curved away from the mandibular angle to avoid injury to the marginal mandibular nerve. The common carotid artery was dissected and exposed, with tumor-feeding branches of the external carotid artery (ECA) ligated, and cranial nerves were preserved. An ECA firmly encased by a large tumor would have to be sacrificed. Any instances of intraoperative ICA injury were repaired with direct closure, patch angioplasty, or great saphenous vein or graft reconstruction. If the carotid artery was clamped due to ICA injury, cerebral perfusion was maintained using the following measures: Systolic blood pressure was raised to 140∼150mmHg or raised by 20∼30mmHg. Heparin was given to achieve a therapeutic activated clotting time (≥200 to 250 seconds). If the Circle of Willis was incomplete or the estimated cross-clamping time was long, then the use of a shunt was compulsory (Table 2). Postoperatively, anticoagulation was initiated for cases requiring a venous graft or prosthetic reconstruction of the ICA. For high-lying CBTs, the mandible bone was disarticulated to facilitate exposure.
Table 1Demographic characteristics, clinical manifestations and lesion characteristics of the 200 cases of CBT. ( No (%) or median (range))
Outpatient visits or telephone follow-ups were performed at three months, one year, and every year with ultrasound and clinical follow-up. These evaluations assessed the recurrence and metastasis of tumors, CND, and stroke, as well as the patency of the target vessels. In addition, CTA was performed in patients with suspected tumor recurrence and at the 1-year follow-up in patients with high-lying tumors or those with ICA reconstruction.
Statistical analysis
The continuous variables were expressed as the mean or median and range. In contrast, the categorical variables were recorded as the total number and percentage of patients. Significant inter-variable differences were indicated with a P value < 0.05. Continuous data were analyzed with analysis of variance and categoric variables with χ2 testing. Univariate and multivariable logistic regression models were used to explore the risk factors of CND. Crude estimates were derived from logistic regression models containing only the specified risk factor, while adjusted values from the models included all risk factors. The multivariable model used entry selection to identify risk factors. The entry and removal criteria were set as 0.05 and 0.10, respectively. Odds ratios (ORs) >1 indicated high risks of CND. Subgroup analysis was also performed to further evaluate the role of EMB in different diameters, Shamblin Classifications, and tumor sites.
Results
Participants’ demographics and clinical manifestations
During the study period, the Division of Vascular Surgery diagnosed 202 patients with CBTs, performed 184 surgeries, and resected 200 CBTs (Figure 2).
Figure 2Flow chart of patient inclusion in the study
In total, 96 males and 88 females with a median age of 37.0 years were included in the analysis. The most common symptom was a painless neck mass. Family history was confirmed in 10 (5.0%) patients. Seven (3.5%) patients had an ipsilateral biopsy or an exploratory operation with subsequent surgical resection (Table 1).
CBT characteristics
Bilateral CBTs were identified in 52 (26.0%) cases, and multicentric CBTs were found in 2 cases. High-lying CBTs were seen in 49 (24.5%) cases. Thirty-four (17.0%), 89 (44.5%), and 77 (38.5%) cases were classified as Shamblin Group I, Group Ⅱ, and Group Ⅲ, respectively (Table 1). Co-existing pheochromocytoma was found in 1 patient. In addition, 1 patient was associated with vagal paraganglioma originating from the vagus nerve at the level of the cranial base.
For 10 malignant CBTs, preoperative CTA revealed invasion of neighboring structures and enlarged regional lymph nodes. CBT resection combined with lymph node dissection and histology confirmed regional lymph node metastasis in one case. Three malignant CBTs presented with lung or bone metastasis preoperatively. Five patients who had a total of six CBTs were diagnosed with malignant CBT at follow-up, including three with distant metastasis and three recurrences, which resulted in follow-up radiotherapy.
Preoperative embolization
In total, 135 patients who collectively had 148 CBTs underwent EMB. Among them, 142 cases achieved significantly reduced tumor vascularity. There were six treatment failures due to feeding arteries originating from the internal carotid artery or carotid bifurcation. In two different cases, embolization was complicated by hemiplegia and hemianopsia, respectively. Both cases involved Shamblin Ⅱ tumors embolized with polyvinyl alcohol and gel foam. Some symptoms experienced by patients persisted at follow-up. Two patients experienced allergic reactions, 6 reported headache or dizziness, and 3 contracted a fever that cleared after medical treatment.
Surgical treatment
Shamblin Ⅰ and Ⅱ tumors were dissected from the ICA using bipolar forceps. Altogether, 10 instances of intraoperative ICA injuries occurred in these dissection cases. Shamblin Ⅲ tumor resections were more vulnerable to ICA injuries (P<0.001), with 33 cases of injury documented. All injured ICAs were successfully repaired except for one injured ICA, which was ligated due to extreme adhesion from a previous operation. Systolic stump pressure of the carotid artery in this patient was 65mmHg. The patient recovered without a stroke.
A shunt was applied in 21 cases of patients undergoing graft replacement for ICA rupture, which accounted for 48.8% of cases of ICA reconstruction. Seven patients with high-lying CBTs involving the base of the skull underwent resection by mandibular subluxation or disarticulation approach. No complications were encountered using this approach. In 15 cases undergoing both CBT and cranial nerve resections (Table 2), high-lying CBTs were identified in 10 cases (20.4% vs. 3.3%, P<0.001).
Genetic screening
In total, 58 patients consented to undergoing genetic tests between 2011 and 2017. SDH mutations were identified in 22 patients, including 4 SDHB, 14 SDHD, and 4 SDHAF2.
In-hospital and follow-up outcomes
Postoperative neurologic complications included four strokes, one transient ischemic attack, and 62 CNDs. All four of the patients who had 2 Shamblin Ⅱ and 2 Shamblin Ⅲ CBTs underwent EBM. Strokes manifested as hemiplegia, with one as glossolalia. At follow-ups after rehabilitation, all patients returned to normal activity with mild weakness.
Eight patients did not complete the follow up, and one patient died during the 28-month follow-up of postoperative CND. There were 34 transient CNDs and 28 permanent CNDs. There were also 5 ICA occlusions for reconstruction, including two ICA reconstructions with great saphenous veins and three with artificial grafts. However, none of these patients developed a stroke (Table 3).
Table 3In-hospital and follow-up results of CBTs resection. (No. (%) or median (range))
In the univariate logistic model, CBTs classified as high-lying (OR=5.185, 95% confidence interval (CI): 2.606-10.318, P<0.001) or Shamblin Ⅲ (OR=3.276, 95% CI: 1.757- 6.106, P<0.001), as well as tumors with a diameter ≥5 cm (OR=2.769, 95% CI: 1.495- 5.128, P=0.001), carried higher risks of postoperative CND.
In multivariable regression analysis, CBTs classified as high-lying (OR=3.398, 95% CI: 1.563-7.383, P=0.002) were associated with higher risks of postoperative CND (Table 4).
Table 4Risk factors of cranial nerve deficit following surgical resection
Risk factors
Crude
Adjusted
OR
95%CI
P value
OR
95%CI
P value
Postoperative cranial nerve deficit
Bilateral
1.112
(0.565, 2.188)
0.759
1.127
(0.510, 2.492)
0.767
Malignant
0.952
(0.238, 3.809)
0.944
0.759
(0.163, 3.522)
0.724
Diameter≥5cm
2.769
(1.495, 5.128)
0.001
1.868
(0.913, 3.820)
0.087
High-lying
5.185
(2.606, 10.318)
<0.001
3.398
(1.563, 7.383)
0.002
Shamblin III
3.276
(1.757,6.106)
<0.001
1.851
(0.915, 3.745)
0.087
Permanent cranial nerve deficit
Bilateral
1.720
(0.737, 4.014)
0.210
2.355
(0.781, 7.097)
0.128
Malignant
1.577
(0.317, 8.840)
0.578
1.238
(0.196, 7.828)
0.821
Diameter≥5cm
4.435
(1.845, 10.662)
0.001
3.586
(1.204, 10.687)
0.022
High-lying
9.993
(4.122, 22.224)
<0.001
4.773
(1.808, 12.595)
0.002
Shamblin III
6.214
(2.494, 15.483)
<0.001
2.893
(1.045, 8.006)
0.041
Note: Crude estimates are from logistic regression models containing only the specified risk factor. Adjusted values are from logistic regression models including all risk factors.
In univariate logistic regression analysis, tumors with a diameter ≥5 cm (OR=4.435, 95% CI: 1.845-10.662, P=0.001), as well as those classified as high-lying (OR =9.993, 95% CI: 4.122-22.224, P<0.001) or Shamblin Ⅲ (OR =6.214, 95% CI: 2.494 -15.483, P<0.001), were associated with higher risks of permanent CND.
In multivariable logistic regression analysis, a maximal diameter≥5 cm (OR=3.586, 95% CI: 1.204-10.687, P=0.022), as well as tumors classified as high-lying (OR=4.773, 95% CI: 1.808-12.595, P=0.002) or Shamblin Ⅲ (OR=2.893, 95% CI: 1.045-8.006, P=0.041), were effective predictors of permanent CND (Table 4).
Comparison of outcomes
More CBTs with EMB presented as high-lying (30.8% vs 7.4%, P =0.001) and Shamblin Ⅲ (43.2% vs 25.9%, P =0.026) tumors compared to tumors without EMB.
EMB was not found to decrease rates of postoperative CND (28.8% vs 37.0%, P = 0.262), permanent CND (13.7 %vs 14.8%, P =0.840), or stroke (2.7 % vs 0%, P =0.576). Similarly, EMB did not significantly decrease blood loss ( 100 ml vs 100.0 ml, P =0.320), blood transfusion, or operation time (180 min vs 170.0 min, P =0.280) (Table 5).
Table 5Comparison outcome between the preoperative embolism group and the non-preoperative embolism group. (No. (%) or median (range))
In subgroup analysis, EMB was still not found to decrease blood loss, blood transfusion, or operation time regardless of tumor diameter (“ ≥ ” or “ ＜” 5 cm), Shamblin classification, or tumor site. However, more instances of CND were found in the EMB group when only Shamblin Ⅲ or low-lying tumors were considered (Table 6).
Table 6Comparison outcome of blood loss, blood transfusion, operation time and cranial nerve defict in different subgroups (No. (%) or median (range)).
CBTs, which are the most common head and neck paraganglioma, are predominantly unilateral tumors. In one meta-analysis involving 4,743 patients, bilateral tumors were only reported in 9.6% of patients. The current study found that 26% of CBTs from the study population were bilateral, while 5.0% of patients had a family history consistent with prior studies.(
and only two cases in the current cohort were confirmed intraoperatively as vagus paragangliomas asscociated with CBTs. This study suggested the importance of proactively investigating the possible presence of bilateral CBTs when unilateral CBTs are diagnosed.
CTA can be used to identify the anatomy and vascularity of CBT, thus confirming nearby structural relationships.
As shown in Figure 1, CTA with a cross-sectional view could quantitate the tiny space between the CBT and the carotid artery, which could direct CBT dissection without arterial injury (Figure 1). Furthermore, CTA could differentiate other tumors, identify recurrent tumors at follow-up, and assess the patency of the Circle of Willis, which is crucial to predicting stroke risks.
In this study, CBTs were surgically resected with bilateral CBTs, which required staged removal. Smaller CBTs were resected first to build patient confidence for further resection. For high-lying tumors involving the base of the skull, subluxation or disarticulation of the mandible helped to obtain a clear operation field, especially with regard to the ICA.
Strokes occurred in 2.0% of patients in this cohort, which is consistent with other studies that reported between 0-8%.
The cohort also showed three cases of unexpected ICA injury needing urgent clamping for ICA repair. Two strokes were attributed to heparinization failure and pre-procedure hypertension. One stroke was due to an abrupt ICA rupture that occurred postoperatively, which may have resulted from hypertension and the disruption of insufficient arterial intimal suturing. All reconstruction of carotid arteries was free from strokes, highlighting the importance of preoperative planning to reduce ICA injuries and stroke risks, especially for tumors classified as Shamblin III. (
) Minimal ICA defects, usually occurring at the bifurcation, could be repaired by direct suture with 6-0 propylene, while more significant defects would need to be patched or reconstructed.(
) The current study reported 2 of 14 autogenous vein occlusions and 3 of 8 graft occlusions at follow-up. The autogenous vein was recommended for young patients and selective reconstruction, while a graft made by polytetrafluoroethylene was applied if a great saphenous vein was unavailable. Anticoagulation or antiplatelet treatment was necessary for long-term patency.
Shunts were used for reducing stroke risks in seriously damaged ICAs. In the case of abrupt ICA rupture, ICA shunts could shorten brain ischemia time, allowing for complication-free ICA reconstruction. Evidence showed that shunts used in carotid endarterectomy did not reduce stroke risk, owing to plaque shedding during shunt implantation.(
In the current study, postoperative CND occurred in 31.0% of patients, with 14.0% resulting in permanent CND. Amato et al. reviewed 19 studies between 2004 to 2014, including 625 CBT surgeries, and CND was observed in 302 (48.3%) cases, including 194 (31.0%) instances of transient and 108 (17.3%) instances of permanent CND.(
) A multi-center study by Kim et al. demonstrated that tumor proximity to the skull base, as well as tumor volume in combination with the Shamblin classification, could better predict CND risk. (
) In the current study, high-lying tumors were found to be an independent predictor of permanent CND. High-lying CBTs are subject to limited access, problematic exposure, and extensive incisions, including the mandible, which thus increases operation complexity, blood loss, and operation times. In addition, the facial nerve’s hypoglossal and marginal mandibular branches are compromised with high-lying CBTs, which predisposes patients to CND. In the current cohort, injury of the hypoglossal nerve by high-lying CBTs was the leading cause of permanent CND, which was found to be a better risk predictor of CND than the Shamblin classification system.
Compared with the non-EMB group, the EMB group did not exhibit a decreased rate of stroke in this cohort, which is consistent with a previous study.(
) On the contrary, there are inherent risks of stroke, although low, to angiography with and without intervention. In the current study, 2 of 148 cases developed stroke due to embolization in unexpected areas. To prevent strokes, EBM was contraindicated in cases with feeding arteries originating from the ICA or carotid bifurcation, or in cases of having anastomosis with intracranial circulation.
Consistent with another study, EMB was not found to decrease the rate of CND.27 However, the EMB group tended to have more Shamblin Ⅲ and large CBTs, which may influence the robustness of this conclusion. In subgroup analysis, more CND was found in the non-EMB group when only Shamblin Ⅲ and low-lying tumors were considered, implying that EMB might be suitable for Shamblin Ⅲ and low-lying tumors. That may be explained by Shamblin Ⅰand Ⅱ tumors generally being able to be safely resected, while high-lying tumors, as mentioned above, carry an extremely high risk of CND regardless of EMB.
Concerning blood loss and operative time, a recent meta-analysis (
) including 1,326 patients reported that EMB patients had significantly lower blood loss and shortened operation times. However, EMB does not decrease blood transfusion requirements or hospital stays. In the current study, EMB neither decreased blood loss nor shortened operative times. Nevertheless, less blood loss was observed when separating the tumors after EMB, consistent with a remarkable decrease in tumor blood supply after EMB. In the subgroup analysis, EMB decreased the incidence of postoperative cranial nerve deficit in Shamblin Ⅲ and low-lying tumors. Thus, we suggest that EMB be planned for Shamblin Ⅲ and low-lying tumors.
Conclusions
In conclusion, CBT tumors could be successfully resected with full examinations, preoperative planning, and improved surgical techniques. EMB was not found to decrease blood loss or shorten operation times. However, EMB might be suitable for Shamblin Ⅲ or low–lying tumors. .
Conflicts of interest
None.
Acknowledgements
This study was partially supported by grants from the National Natural Science Foundation of China (No. 81873813) and the Natural Science Foundation of Guangdong Province (No. 2020A1515011412).
References
Sajid MS
Hamilton G
Baker DM
Joint Vascular Research Group
A multicenter review of carotid body tumor management.
Grantsupport: This study was partially supported by grants from the National Natural Science Foundation of China (No. 81873813) and the Natural Science Foundation of Guangdong Province (No. 2020A1515011412).
00122-X&id=gr1.jpg){kind=link}
00122-X&id=gr2.jpg){kind=link}