Management of Multiple Head and Neck Paragangliomas With Assistance of a 3-D Model

Abstract

Introduction:

Extirpation of multiple head and neck paragangliomas carries challenge due to close anatomic relationships with critical neurovascular bundles.

Objectives:

This study aims to assess whether the application of 3-D models can assist with surgical planning and treatment of these paragangliomas, decrease surgically related morbidity and mortality.

Methods:

Fourteen patients undergoing surgical resection of multiple head and neck paragangliomas were enrolled in this study. A preoperative 3-D model was created based on radiologic data, and relevant critical anatomic relationships were preoperatively assessed and intraoperatively validated.

Results:

All 14 patients presented with multiple head and neck paragangliomas, including bilateral carotid body tumors (CBT, n = 9), concurrent CBT with glomus jugulare tumors (GJT, n = 4), and multiple vagal paragangliomas (n = 1). Ten patients underwent genomic analysis and all harbored succinate dehydrogenase complex subunit D (SDHD) mutations. Under guidance of the 3-D model, the internal carotid artery (ICA) was circumferentially encased by tumor on 5 of the operated sides, in 4 (80%) of which the tumor was successfully dissected out from the ICA, whereas ICA reconstruction was required on one side (20%). Following removal of CBT, anterior rerouting of the facial nerve was avoided in 3 (75%) of 4 patients during the extirpation of GJT with assistance of a 3-D model. Two patients developed permanent postoperative vocal cord paralysis. There was no vessel rupture or mortality in this study cohort.

Conclusion:

The 3-D model is beneficial for establishment of a preoperative strategy, as well as planning and guiding the intraoperative procedure for resection of multiple head and neck paragangliomas.

Introduction

Paragangliomas arising in the head and neck region are uncommon and include carotid body tumors (CBT), glomus jugulare tumor (GJT), and paragangliomas originating from the vagus or the sympathetic nerve chain. ¹ Some uncommonly reported sites include the nasal cavity, orbit, oropharynx, and larynx. ² Given the complex neurovascular relationships surrounding these tumors at the skull base and the neck, surgical treatment remains challenging, especially for tumors that are intimately involved and surround the great vessels and multiple cranial nerves (CNs). ³ Complications from injury to these critical neurovascular structures are not infrequent, and potentially functionally devastating when dealing with multiple tumors. Hence, surgical extirpation of these tumors needs to be carefully considered as a treatment option for appropriately selected head and neck paragangliomas. ⁴

Sporadic paragangliomas typically present as solitary tumors, whereas patients with family history, tend to have bilateral or multiple tumors. ⁵ Mutations of the genes for the succinate dehydrogenase, SDHx, have been identified as causing familial head and neck paragangliomas. ⁶ Therefore, for patients with an inheritable predisposition with SDHx mutations, they can harbor multiple paragangliomas including concurrent neck paragangliomas (CBT or vagal paragangliomas) and GJT.^7,8

Compared to a solitary paraganglioma, the management of multiple paragangliomas is more challenging and technically demanding given the complex neurovascular structures at risk for damage and dysfunction. ⁴ Adequate preoperative assessment of the relationships between the tumor and the internal carotid artery (ICA) and function of CNs (VII to XII) are necessary prior to choose a safe and appropriate surgical treatment strategy.^9-11 However, in patients with multiple occurrences of paragangliomas, the ICA and CNs may intimately adhere to the tumor. Surgical resection of these multiple tumors at a time or in a staged fashion may pose great risk of injury to the great vessels and related CNs. Therefore, careful preoperative evaluation and planning is critical and representative anatomic surgical models may be helpful.

In addition to routine radiographic examinations such as contrast-enhanced computed tomography (CT) or magnetic resonance imaging (MRI) scans for assessment of patients with paragangliomas, digital subtraction angiography (DSA) is valuable for identification of the corresponding feeding vessels, completion of preoperative tumor embolization, assessment of the patency of anterior and posterior communications, and balloon occlusion test. ¹² Moreover, the use of a preoperative 3-D model has been previously shown to help with complex oral and maxillofacial reconstruction and resection of tumors arising from the parapharyngeal space.^13,14 However, the potential value of these 3-D models in resection of multiple head and neck paragangliomas has not been sufficiently explored.

The purpose of this study was to assess the application of a 3-D models to assist in preoperative surgical strategy and intraoperative guidance for resection of multiple head and neck paragangliomas, in order to decrease the potential operative morbidity.

Materials and Methods

Clinical Data

Seventeen patients with multiple head and neck paragangliomas treated from January 1, 2010, to December 31, 2019, at Beijing Tongren Hospital, Capital Medical University, China, were identified and reviewed. Fourteen of these patients underwent surgical resection of the paragangliomas and were enrolled in this study. Three patients with multiple paragangliomas chose observation or radiotherapy rather than surgery and were excluded from this study. Informed consent was obtained from these 14 surgical patients, and the study protocol was approved by the Ethics Committee of Beijing Tongren Hospital, Capital Medical University, China.

Preoperative contrast-enhanced CT and gadolinium-contrast MRI scans were performed in all 14 patients. Preoperative DSA examination was routinely employed to assess the blood supply of the tumors, patency of the anterior and posterior communications, and to complete any necessary preoperative embolization and balloon occlusion test.

3-D Printing Model

The 3-D model for all 14 patients with multiple paragangliomas was established prior to the surgery. The image data incorporating axial, coronal, and sagittal images of both the CT and MRI were obtained in Digital Imaging and Communications in Medicine format. This was imported into an IVSP Image software (IVSPlan) for segmentation and reconstruction (Video 1).

On each image plane, the arteries, veins, tumors, and bony facets (maxillary, mandibular, and temporal bone) were identified and segmented. To improve the accuracy of reconstruction, a Head Neck surgeon (H.B.X. or L.F.L.) working together with the technician of the IVSPlan company during the reconstruction process. Following the image segmentation, a 3-D model including the tumor and the great vessels (common carotid artery, ICA, and the internal jugular vein [IJV]), and bony facet of the skull base, was reconstructed with each portion created separately and then assembled together (Figure 1A). Different colors of the photosensitive resin were used to facilitate the printing process, generating an entire printed model which had a similar dimension with the actual situation (Figure 1B). The maximum printing resolution (accuracy) for these models was 0.1 mm. In addition, the computer-generated models will be available to the surgeons to help better appreciate the complex relationships whenever necessary.

Figure 1.

(A) 3-D reconstruction of multicentric head and neck paragangliomas; (B) The printed 3-D model.

Results

Demographics, hybrid tumor types, number of tumors, maximal tumor size, coexistence of pheochromocytoma, bilateral resection, necessity for anterior rerouting of the facial nerve, and postoperative morbidities are all summarized in Table 1. Among these 14 patients, 10 patients underwent genomic sequencing analysis of succinate dehydrogenase complex subunit D (SDHD), and all 10 patients demonstrated an SDHD mutation.

Table 1.

Patient Characteristics.^a

No.	Sex	Age	Tumor type	Tumor N	Max TS	Pheo	Bi resection	FN rerouted	Comorbidities
1	Male	42	Bilateral CBT	2	5	No	No
2	Male	28	Bilateral CBT	2	4	No	No
3	Female	34	Bilateral CBT	2	3.5	No	Yes		TVCP
4	Female	36	Bilateral CBT	2	3.2	No	Yes
5	Female	44	Bilateral CBT	2	4	No	No
6	Female	32	CBT and GJT	3	3.6	No	No	No	Dysphagia
7	Female	49	CBT and GJT	2	4.3	No	No	Yes
8	Female	29	Bilateral CBT	2	3.8	No	Yes		FP and PVCP
9	Male	59	CBT and GJT	3	3.2	No	No	No	PVCP
10	Female	38	Bilateral CBT	3	4.3	No	Yes		Dysphagia
11	Male	35	Bilateral CBT	2	5	No	No
12	Male	42	Bilateral CBT	2	4.6	No	No
13	Male	38	Vagal PG	2	3.8	No	No
14	Female	33	CBT and GJT	5	4	Yes	No	No	TVCP

Abbreviations: CBT, carotid body tumor; FP, facial paresis; GJT, glomus jugulare tumor; PG, paraganglioma; PVCP, permanent vocal cord paresis; TVCP, temporary vocal cord paresis.

^a Sex, age (years), tumor type, Tumor numbers (N), Maximal tumor size (Max TS, cm), coexistence of Pheochromocytoma (Pheo), Bilateral (Bi) resection, facial nerve (FN) rerouting, and comorbidities.

Of the 14 patients included in the present study, 6 were male and 8 were female, with ages ranging from 28 to 59 years (mean = 38.5 ± 8.3 years). All patients had a familial predisposition of paraganglioma and demonstrated multicentric tumors (≥2 tumors). Nine patients demonstrated bilateral CBT (Figure 2A); 4 patients had simultaneous presence of a CBT and a GJT (Figure 2B and C); and the remaining 1 patient (patient # 13) demonstrated 2 distinct vagal paragangliomas (Figure 2D).

Figure 2.

(A) Occurrence of bilateral carotid body tumor (CBT; blue arrows); (B) Presence of a CBT (red arrow); (C) Presence of a glomus jugulare tumor (GJT) in the ipsilateral side (green arrow); (D) Presence of multiple vagal paragangliomas (red arrows).

For 9 patients with bilateral CBT, 4 patients underwent and completed resection of bilateral lesions in a 2-stage fashion. Five patients underwent resection of only one side of bilateral CBT, the side of lesion with lower Shamblin classification stage was extirpated first. Three of these patients (3/5 = 60%) experienced postoperative vocal cord paralysis (1 permanent and 2 temporary) and declined further surgical treatment of the contralateral CBT and chose continued observation. The remaining 2 patients (2/5 = 40%) had no postoperative complications but refused second-stage resection and preferred observation (n = 1) or radiotherapy (n = 1).

The 3-D models including the tumor and surrounding vascular bundles and bony structures were reconstructed and printed for all 14 patients. Of all 18 operated sides, the ICA was found to be circumferentially encased by the tumor (n = 5) based on the preoperative 3D model (Figure 3A). Vascular reconstruction was prepared preoperatively for all 5 operated tumor sides that had the circumferential ICA encasement by the tumor. However, the tumors were able to be completely separated from the ICA in 4 sides (4/5 = 80%); in the remaining tumor (1/5 = 20%), temporary ICA occlusion and subsequent vascular reconstruction with a saphenous vein were employed (Figure 3B).

Figure 3.

(A) Internal carotid artery (ICA) was circumferentially enclosed by the tumor on a 3-D model (green arrow); (B) Reconstruction of the ICA with a saphenous vein (green arrow).

For 4 patients with concurrent ipsilateral GJT and CBT, the preoperative 3-D model was beneficial for determining the possibility of resection of 2 tumors simultaneously and likelihood and necessity of facial nerve rerouting (Figure 4A). Following extirpation of the CBT, a retrofacial nerve corridor was created to facilitate the complete resection of the tumor (Figure 4B) for localized lesions surrounding the jugular bulb and tympanic cavity (n = 3); and as such, the anterior rerouting of the facial nerve was avoided (Figure 4C). For a larger sized tumor with spread extensive into adjacent structures (n = 1), however, anterior rerouting of the facial nerve was necessary and employed to facilitate safe total tumor resection.

Figure 4.

(A) Demonstration of the relationship between the glomus jugulare tumor (GJT; green arrow) and the tympanic cavity on a 3-D model, pink arrow: mastoid tip; (B) Establishment of a retrofacial corridor for resection of GJT; (C) Tumor (T) was resected via the retrofacial corridor (green arrow) without rerouting of the facial nerve. FC indicates fallopian canal; JB, jugular bulb; MEC, middle ear cavity; SS, sigmoid sinus.

Establishment of a 3-D model was also beneficial for assessment of the relationship between the tumor and the IJV, especially for tumors located close to the skull base. The IJV was partially or completely compressed by the tumor (Figure 5A). In all 18 operated tumor sides, the IJV (cervical segment) was partially (n = 10) or completely (n = 8) compressed on contrast-enhanced CT or MRI scans, and this was similarly visualized on the 3-D models (Figure 5B and C).

Figure 5.

(A) Relationship between the paraganglioma (T) with the internal jugular vein (IJV, green arrow) on contrast enhanced computed tomography (CT) image; (B) IJV on the left side was partially compressed; (C) IJV was completely compressed on a 3-D printed model, blue arrow: external jugular vein (EJV), red arrow: internal carotid artery (ICA).

Surgical removal of these tumors led to some expected postoperative complications including vocal cord paralysis (n = 5, 3 for bilateral CBT, 2 for vagal paragangliomas), dysphagia (n = 2, 1 for bilateral CBT, 1 for vagal paraganglioma), and facial paresis (n = 1, concurrent GJT and CBT). However, symptoms related to vocal cord paralysis resolved over time, and only 2 patients developed permanent vocal cord paralysis (patient #8 & 9), who ultimately received vocal cord injection medialization therapy. No vascular rupture, transient ischemic attack, stroke, intracranial infection, cerebral spinal fluid leak, nor mortality was encountered in this cohort. To date, no recurrences were noted on follow-up examinations and imaging.

Discussion

Surgical extirpation of paragangliomas arising in the head and neck region carries potential risks of vascular rupture and nerve injury. ¹⁵ Therefore, typically a team approach composed of a vascular surgeon and a head neck surgeon is required to facilitate complete and safe resection. ¹⁶ Auxiliary diagnostic methods such as the 3-D model described in this study can help appreciate the complex relationships between the tumor and the surrounding neurovascular bundle prior to and during the operation. Ultimately, 3-D modeling can lead to improved operative planning and increased overall safety of these procedures.

All 14 patients involved in the present study demonstrated a familial predisposition, and the SDHD mutation was detected in all 10 evaluated patients. ⁶ Concurrent presence of a pheochromocytoma is occasionally encountered in patients with head and neck paragangliomas. ¹⁷ Pheochromocytoma might lead to an intraoperative crisis of severe hypertension; therefore, the surgical and anesthesia teams must be aware of these tumors and well versed in strategies to manage perioperative alterations in blood pressure. One patient in this cohort underwent resection of bilateral pheochromocytomas prior to manifestation of the head and neck paragangliomas. The removal of neck paragangliomas went smoothly, and no significant alteration of blood pressure was detected.

3-D modeling is not a novel technique as it has been successfully applied in oral and maxillofacial reconstruction as well as to transoral approach and resection of post-styloid tumors.^13,14 3-D models of both the printed and the computer-generated versions can more reliably distinguishing the critical neurovascular structures and the potential degree of involvement by these tumors, which is helpful for surgeons to make decision during the operation. We find that some aspects of the challenging anatomic relationships are more fully appreciated when the surgeon can hold the model in their hands and orient the model from all desired perspectives; while the computer-generated model allows for a secondary approach to also seek to understand the anatomy from various perspectives. For the 14 patients included in this study, these models give important preoperative information including tumor size, partially or circumferential encasement the ICA, tumor relationship with the IJV, distance from the tumor to the cranial base and the carotid bifurcation, and the relationships of multiple ipsilateral tumors. The established preoperative surgical plans and strategies based on the 3-D model were appropriately validated during the operation.

For patients with multiple paragangliomas, application of a 3-D model provided a more comprehensive and much improved assessment of the relationship between the tumor and the great vessels (ICA, common carotid, and the IJV), which is helpful for coming into a plan to determine and prioritize the appropriate surgical resection side. For patients with bilateral CBT or multiple vagal paragangliomas, the choice of resection was routinely selected on the side with lower Shamblin classification or less potential occurrence of complications. ¹⁸ If no postoperative issues such as vocal cord palsy or dysphagia occurred during resection of the initial side, continued resection of the contralateral lesion was considered, ¹⁹ as well as observation or radiotherapy may be indicated depending on patient preference. We used the same strategy for treatment of multicentric bilateral paragangliomas; however, only 3 patients agreed to resection of the contralateral lesion.

Although a 3-D model can clearly demonstrate the degree of ICA encasement by the tumor, a circumferentially encased ICA is not an absolute indication to reconstruct the ICA. Temporary great vessel occlusion and same stage reconstruction is always necessary for resection of Shamblin II or III CBTs. ²⁰ In our experience, if the tumor had a high T2-weighted imaging MRI signal on MRI, the attempt to separate the tumor from the ICA could be possible, and the tumors in 4 (80%) of 5 patients were successfully separated without ICA reconstruction. However, the surgical team needs to be prepared for vascular sacrifice or reconstruction in cases where the tumor cannot be completely dissected away. Especially challenging are cases where the tumors are located in the upper parapharyngeal space with partially or circumferentially encased ICA and managing the ICA in this area deserves careful planning and strategy. Reconstruction close to the cranial base cannot be routinely performed similarly to reconstruction down at the carotid bifurcation. Therefore, preoperative DSA and permanent embolization and occlusion of the vessel or preoperative placement of a vascular stent may become an alternative option.^21,22 At our institution, only those patients who have systolic pressure higher than 60 mm Hg and a patent anterior and posterior communications at the Circle of Willis are considered to be safe candidates to perform such a procedure.

The decision to dissect the concurrent CBT and GJT in the ipsilateral neck at one stage is determined by multiple factors. ²³ Following extirpation of the CBT, the damage and intraoperative bleeding for resection of the GJT should be decreased as possible. Anterior rerouting of the facial nerve is often required to facilitate complete resection of GJTs, ²⁴ and some modified techniques sparing the facial nerve for resection of GJTs were also reported. ²⁵ As described in the present study, the 3-D model can help delineate the size and involved areas of the GJT. Preoperative modeling allowed for improved preoperative surgical planning for such cases. In 3 (75%) of 4 patients with concurrent presence of CBT and GJT, establishment of a retrofacial corridor was performed, and complete resection of the tumor was achieved without anterior rerouting of the facial nerve. However, for one patient with wide expansion involving the adjacent structures, anterior rerouting of the facial nerve was necessary predicted by the preoperative 3-D model.

The authors recognize there are significant limitations of this study. This cohort contains limited sample size, and it is difficult to perform a case–control comparative study with larger sample size given the rare occurrence of multiple head and neck paragangliomas; and as such, the implications of a 3-D model for management of multiple paragangliomas deserve further assessment. Moreover, the process of a 3D reconstruction and printing contains multiple procedures, and different colored materials are used for ultimate printing. 3D pieces of the involved structures (tumor, bone, vessels, etc) are printed individually and then assembled to create the final model, the process of which may introduce errors. Furthermore, the maximum printing resolution (0.1 mm in the current study) is also a restricting factor for discrimination of tiny structures. Therefore, a bias may be present for usage of a 3-D model for establishment of a surgical plan. In addition, the decision to resect these multiple paragangliomas is determined by multiple factors, ²⁶ critical multidisciplinary team discussion and the sufficient experience of the surgical team are prerequisites as part of the decision-making process for each case. ²⁶

Conclusion

3-D models provide valuable preoperative information on tumor relationship to critical neurovascular structures at risk during potential surgical resection. The use of such models can help establish a safe preoperative strategy, which is beneficial for guiding the intraoperative procedure for successful and safe resection of multiple head and neck paragangliomas.