Head and Neck Paragangliomas: A 30-year experience

Abstract

Background:

We aimed to review our experience and the changing trends in the management of head and neck paragangliomas (HNPG) over the last three decades.

Methods:

We retrospectively reviewed 103 patients with HNPG treated at our center (1986-2017). We included patients treated with surgery, radiotherapy and patients maintained under active surveillance.

Results:

Of the surgically treated patients (n=79), 20% (12/59) of the carotid body tumors (CBT) had a cranial nerve deficit as sequela compared to 95% (19/20) of the non-CBT. Radiotherapy controlled growth in all tumors treated with this modality (n=10). Of the initially observed patients, 70% (14/20) remained stable and did not require additional treatment. Stratifying by decades, there was a progressive increase in patients initially attempted to be observed and a decrease in upfront surgery. No deaths attributable to the HNPG were encountered.

Conclusions:

Surgery is an effective treatment for CBT. Nonsurgical treatment should be considered for non-CBT.

INTRODUCTION

Paragangliomas (PG) are a diverse group of neuroendocrine tumors arising from chromaffin cells within paraganglionic tissues of the autonomic nervous system. These tumors share a common embryologic origin from neural crest cells, and may extend from the skull base to pelvis.^1,2 Consequently, paragangliomas are often classified by their location, which incidentally has also been shown to delineate their endocrine activity. As such, head and neck paragangliomas (HNPG), which originate from parasympathetic tissue with relatively little catecholamines, lack endocrine activity. However, up to 5% of all HNPG do produce hormones.^1,3 Twenty different locations have been reported for HNPG alone, with the carotid body, vagal body, and jugulo-tympanic region making up the most common locations.³ Sporadic and familial variants exist in the population. Approximately 24% of all PG may be familial with this number increasing to 35% for HNPG.^2–4 Multiple susceptibility genes for PG have been identified with mutations of the succinate dehydrogenase enzyme (SDH) being the most common.^3–6 Familial variants tend to have a younger age of onset, a higher number of patients with multiple or bilateral PG, and positive family history.^2–4,7

Historically, surgical resection has remained the mainstay of treatment for HNPG. However, challenges to surgical management include the risk of injury to vital vascular structures with resultant transient ischemic attacks or major cerebrovascular accidents occasionally leading to permanent neurologic deficits. Moreover, cranial nerve (CN) deficits may be quite significant, depending on the location and size of the tumor as well as the surgical approach and experience of the surgeon. These morbidities may have longstanding cosmetic and or functional implications such as hoarseness, aspiration requiring permanent tracheostomy, gastrostomy, or deficits from other cranial nerve palsy.

Overall, paragangliomas are rare, with population studies noting an incidence of one to eight cases per million people.^1,4 Implications of their management and possible sequela generate the need for a thorough understanding of the disease and management options. With improved understanding of the natural history of these tumors, better identification of familial forms, and the increasing experience in nonsurgical treatment, the management paradigms for HNPG are changing. Consequently, an ever-increasing trend toward individualizing management based on patient factors as well as tumor biological behavior is now advocated for these patients. The purpose of our study was to review our experience in the management of patients with HNPG and the changing trends that we have implemented over the last three decades. Patient’s characteristics, type of treatment and long-term functional outcomes are reported comparing CBT with non-CBT due to the differences between these two groups.

MATERIALS AND METHODS

After approval from our Institutional Review Board, a retrospective review of medical records identified 178 consecutive patients with HNPG managed at our center from 1986 to 2017. We included all patients treated with surgery, radiotherapy and patients maintained under active surveillance. We excluded patients who were treated at an outside institution for the reference PG or those with incomplete data. Seventy-five patients were excluded leaving 103 patients available for analysis. Figure 1 shows the corresponding flow diagram and the description of the subtypes of PG included in the study. For patients with multiple PG we considered the first PG treated at our center as the reference for the data collection and analyses. A further explanation on the multiple PG was performed on a separate section.

Figure 1.

Flow diagram and description of the subtypes of paragangliomas included in the study.

Data evaluated from patientś medical records included preoperative assessments, imaging studies, surgical reports, postoperative examination, and functional and treatment outcomes. Preoperative assessments focused on history and physical examination findings and patient demographics including sex, age at presentation, and presenting symptoms among others. Surgical approach, estimated blood loss (EBL) and complications were evaluated from operative reports. For patients initially managed with observation, the change in growth rate was calculated selecting the axis that grew more during observation (cm/year). The primary end points assessed were type of treatment, changes in management approach through decades, and long-term functional outcomes. The secondary end points evaluated included the development of local, regional and distant recurrence, and survival outcomes, specifically alive with or without disease.

To compare variables between groups we used Student’s t test to compare means and Pearson’s chi-squared test or Fisheŕs exact test to compare categorical variables. A P value of less than 0.05 was considered statistically significant. All statistical analyses were conducted using SPSS (v25.0, IBM Corporation; Somers, NY).

RESULTS

There were 68 patients identified with CBT and 35 identified with non-CBT. From the latter, 24 were Glomus Vagale (GV), 8 were Glomus Jugulare (GJ), 2 were Glomus Jugulotympanicum (GJT) and 1 patient had a PG located in the neck not originating from any specific vascular or nervous structure (Figure 1). Patients’ demographics and details of clinical presentation are shown in Table 1. Mean age at diagnosis was 50 years (range, 18-86). A female predominance was noted in the entire cohort, being more pronounced in the non-CBT group. Most patients presented with a neck mass (69%). When analyzed by groups, non-CBT had a significantly lower percentage of patients complaining of a neck mass, and a higher percentage of patients experiencing hoarseness or dysphagia. At presentation, only one patient in the CBT group had a cranial nerve palsy (1.5%), which was a cranial nerve X deficit. Conversely, in the non-CBT group, 42.9% of patients had a cranial nerve X deficit at presentation. Additionally, in the group of patients with CBT, tumors were classified according to Shamblin’s classification, with 8 patients (11.8%) falling under Shamblin I, 41 under Shamblin II (60.3%) and the 19 remaining were classified as Shamblin III (27.9%).⁸

Table 1.

Patients’ demographics and clinical presentation.

Characteristics		Overall N (%)	CBT N (%)	NON-CBT N (%)	P value
Patients	N (%)	103	68 (66.0%)	35 (34.0%)
Age	(mean, range) years	50.4 (18.9-86.6)	49.1 (18.9-81.1)	52.9 (29.6-86.6)	0.237
Sex	Male Female	37 (35.9%) 66 (64.1%)	27 (39.7%) 41 (60.3%)	10 (28.6%) 25 (71.4%)	0.265
Laterality	Left Right	55 (53.4%) 48 (46.6%)	39 (57.4%) 29 (42.6%)	16 (45.7%) 19 (54.3%)	0.262
Size	(mean, range) cm	3.8 (0.9-8.0)	3.7 (0.9-8.0)	3.9 (1.4-6.6)	0.535
Symptoms	Neck mass Hoarseness Incidental Dysphagia	71 (68.9%) 17 (16.5%) 13 (12.6%) 11 (10.7%)	55 (80.9%) 1 (1.5%) 9 (13.2%) 0 (0.0%)	16 (45.7%) 16 (45.7%) 4 (11.4%) 11 (31.4%)	< 0.001 < 0.001 1.000 < 0.001
Duration of symptoms	(mean, range) months	2.2 (0.0-40.0)	2.3 (0.0-40.0)	2.0 (0.1-26.0)	0.831
Physical examination	V cn palsy VII cn palsy VIII cn alterations * IX cn palsy X cn palsy XI cn palsy XII cn palsy	1 (1.0%) 2 (1.9%) 5 (4.9%) 8 (7.8%) 16 (15.5%) 3 (2.9%) 7 (6.8%)	0 (0.0%) 0 (0.0%) 0 (0.0%) 0 (0.0%) 1 (1.5%) 0 (0.0%) 0 (0.0%)	1 (2.9%) 2 (5.7%) 5 (14.3%) 8 (22.9%) 15 (42.9%) 3 (8.6%) 7 (20.0%)	0.340 0.113 0.004 < 0.001 < 0.001 0.037 < 0.001
Imaging	CT MRI MRA Angiogram	90 (87.4%) 79 (76.7%) 27 (26.2%) 18 (17.5%)	58 (85.3%) 48 (70.6%) 21 (30.9%) 12 (17.6%)	32 (91.4%) 31 (88.6%) 6 (17.1%) 6 (17.1%)	0.535 0.050 0.133 0.949
Follow-up	(median, IQR) years	3 (1-6)	2 (1-6)	4 (2-6)	0.160

^*Hearing loss or tinnitus.

Abbreviations: cn, Cranial nerve; CT, Computerized Tomography; MRI, Magnetic Resonance Imaging; MRA, Magnetic Resonance Angiography.

Analyzing family history of PG, there were 15 patients (14.6%) in our cohort who reported family history of PG. Mean age was 43.5 years (range, 23–60), 60% of patients were females, and 7 patients had multiple PG (46.7%). All patients for whom genetic testing was performed harbored a mutation, with the SDH-B gene being the most commonly implicated. One of the 6 patients where secretory activity was analyzed, showed secretion of normetanephrine and norepinephrine.

Treatment modality and changes in treatment philosophy through the last 30 years are shown in Table 2. Overall, a higher percentage of patients with CBT were treated with upfront surgery compared to patients with non-CBT, and a higher percentage of patients with non-CBT were treated with upfront radiotherapy compared to patients with CBT. Stratifying the results by decades, we observed a progressive increase in the number of patients that were initially attempted to be observed (and only treated if rapid increase in size or symptoms), and a decrease in upfront surgery as the treatment of choice.

Table 2.

Overall management and changes in management through time.

Management	Overall (n = 103)	CBT (n = 68)	NON-CBT (n = 35)	P value
OVERALL -30 years-
Observation*	20 (19.4%)	11 (16.2%)	9 (25.7%)	0.246
Upfront Surgery +/− Radiotherapy	75 (72.8%)	56 (82.4%)	19 (54.3%)	0.002
Upfront Radiotherapy alone	8 (7.8%)	1 (1.5%)	7 (20.0%)	0.002
CHANGES THROUGH DECADES
1986-1995	N = 21	N = 18	N = 3
Observation*	1 (4.8%)	1 (5.6%)	0 (0.0%)
Upfront Surgery +/− Radiotherapy	20 (95.2%)	17 (94.4%)	3 (100.0%)
Upfront Radiotherapy alone	0 (0.0%)	0 (0.0%)	0 (0.0%)
1996-2005	N = 38	N = 25	N = 13
Observation*	3 (7.9%)	2 (8.0%)	1 (7.7%)
Upfront Surgery +/− Radiotherapy	31 (81.6%)	23 (92.0%)	8 (61.5%)
Upfront Radiotherapy alone	4 (10.5%)	0 (0.0%)	4 (30.8%)
2006-2017	N = 44	N = 25	N = 19
Observation*	16 (36.4%)	8 (32.0%)	8 (42.1%)
Upfront Surgery +/− Radiotherapy	24 (54.5%)	16 (64.0%)	8 (42.1%)
Upfront Radiotherapy alone	4 (9.1%)	1 (4.0%)	3 (15.8%)

^*Observation alone or followed by active treatment (surgery or radiotherapy) if rapid increase in size or symptoms

Out of the 20 patients that were initially attempted to be managed with observation, the group of patients that observation was followed by active treatment (surgery or radiotherapy) (n = 6, 30%) compared to patients that were only observed (n = 14, 70%), showed a shorter duration of observation (mean 2 years, range 1-5 vs mean 6 years, range 1-15; P = 0.003) and a trend towards a faster growth rate (mean 0.9cm/year, range 0.1-2.5 vs mean 0.1cm/year, range 0.0-0.4, P = 0.068). For patients treated with radiotherapy, mild dermatitis and mucositis were observed in 40% of the cases. Radiotherapy was able to control growth in all treated tumors (70% stayed stable and 30% decreased in size). Hoarseness and dysphagia due to vagus nerve dysfunction were the most common sequelae, all of which were present at diagnosis, before radiotherapy.

Analyzing patients treated surgically, either upfront or after observation, (n = 79), only 12.7% underwent pretreatment embolization, being more frequent in the non-CBT group. Only 4 CBT underwent pretreatment embolization, all of them were Shamblin III. Taking into consideration all patients treated surgically, mean EBL was 271 cc (range, 5-1800). When analyzing EBL based on pretreatment embolization, mean EBL for patients who underwent pretreatment embolization was not different from the mean EBL of patients who did not (466cc vs 241cc, P = 0.06). Among all patients treated with surgery, the specimen size was larger, and the length of hospitalization was longer in the group of patients with non-CBT tumors (Table 3). In the immediate postoperative setting, the most frequent complications were hematomas, and the most commonly affected structures were the facial nerve, the vagus nerve and the hypoglossal nerve, all of them more frequently affected in the non-CBT group.

Table 3.

Patients treated with surgery, either upfront or after observation (n = 79).

Characteristics	Overall (n = 79)	CBT (n = 59)	NON-CBT (n = 20)	P value
Pretreatment embolization No Yes	69 (87.3%) 10 (12.7%)	55 (93.2%) 4 (6.8%)	14 (70.0%) 6 (30.0%)	0.014
Estimated blood loss Mean (range) cc	271 (5-1800)	235 (5-1800)	383 (50-1100)	0.114
Tumor size (pathology) Mean (range) cm	3.5 (1.1-8.5)	3.2 (1.1-7.0)	4.6 (1.7-8.5)	0.012
Length of hospitalization Mean (range) days	4 (1-15)	4 (1-15)	7 (1-14)	0.001
STRUCTURES SACRIFICED DURING SURGERY
NEURAL
VII cn	1 (1.3%)	0 (0.0%)	1 (5.0%)	0.253
IX cn	2 (2.5%)	0 (0.0%)	2 (10.0%)	0.062
X cn	16 (20.3%)	0 (0.0%)	16 (80.0%)	< 0.001
XI cn	2 (2.5%)	0 (0.0%)	2 (10.0%)	0.062
XII cn	5 (6.3%)	1 (1.7%)	4 (20.0%)	0.013
Sympathetic chain	3 (3.8%)	2 (3.4%)	1 (5.0%)	1.000
VASCULAR
Internal Jugular Vein	4 (5.1%)	1 (1.7%)	3 (15.0%)	0.048
Internal Carotid Artery	1 (1.3%)	1 (1.7%)	0 (0.0%)	1.000
External Carotid Artery	13 (16.5%)	13 (22.0%)	0 (0.0%)	0.031
Other Vessels	15 (19.0%)	14 (23.7%)	1 (5.0%)	0.098

Abbreviations: cn, Cranial nerve.

Long-term functional outcomes are shown in Table 4. Overall, regardless of treatment modality, the most common definite sequela was vagus nerve palsy (28.2%). In the last physician’s assessment, most of these patients reported hoarseness and dysphagia to some degree, most of them due to confirmed cranial nerve palsy, while some patients did not show alterations in the clinical evaluation. Most cranial nerve deficits were reported in a higher percentage of patients with non-CBT compared to patients with CBT. If we only analyze the surgically treated patients (n = 79), 20% of the CBT cohort (12/59) had a cranial nerve deficit as a sequela of the surgery compared to 95% of the non-CBT cohort (19/20). From the 4 patients that needed a percutaneous endoscopic gastrostomy (PEG), 3 of them maintained PEG dependency at last follow-up. All patients who maintained PEG dependency were patients with non-CBT treated surgically. The only patient that needed a tracheostomy, was decannulated 3 months after surgery.

Table 4.

Functional outcomes.

Characteristics	Overall N = 103	CBT N = 68	NON-CBT N = 35	P value
Hoarseness	24 (23.3%)	7 (10.3%)	17 (48.6%)	< 0.001
Dysphagia	23 (22.3%)	6 (8.8%)	17 (48.6%)	< 0.001
Baroreceptor alterations	2 (1.9%)	1 (1.5%)	1 (2.9%)	1.000
Transient Ischemic Attack	1 (1.0%)	1 (1.5%)	0 (0.0%)	1.000
Xerostomia	1 (1.0%)	1 (1.5%)	0 (0.0%)	1.000
V cn palsy	1 (1.0%)	0 (0.0%)	1 (2.9%)	0.340
VII cn palsy	7 (6.8%)	2 (2.9%)	5 (14.3%)	0.043
VIII cn alterations*	7 (6.8%)	0 (0.0%)	7 (20.0%)	< 0.001
IX cn palsy	6 (5.8%)	0 (0.0%)	6 (17.1%)	0.001
X cn palsy	29 (28.2%)	4 (5.9%)	25 (71.4%)	< 0.001
XI cn palsy	7 (6.8%)	2 (2.9%)	5 (14.3%)	0.043
XII cn palsy	15 (14.6%)	4 (5.9%)	11 (31.4%)	0.001
Sympathetic chain palsy	5 (4.9%)	2 (2.9%)	3 (8.6%)	0.334
First bite syndrome	6 (5.8%)	2 (2.9%)	4 (11.4%)	0.177
IX cn pain syndrome	2 (1.9%)	1 (1.5%)	1 (2.9%)	1.000
GAN amputation neuroma	1 (1.0%)	1 (1.5%)	0 (0.0%)	1.000
PEG	3 (2.9%)	0 (0.0%)	3 (8.6%)	0.037
Tracheostomy	0 (0.0%)	0 (0.0%)	0 (0.0%)	-

^*hearing loss or tinnitus.

Abbreviations: cn, Cranial nerve; GAN, greater auricular nerve; PEG, percutaneous endoscopic gastrostomy.

A subset analysis in patients with multiple PGs (n = 12, 11.7%) is shown in Table 5. Mean age was 44.2 years (range, 30-60). The most common treatment approach was surgery. Seven patients reported family history of PG (58.3%), 3 patients underwent genetic testing, with 2 of them resulting in a mutation in the SDH gene. Secretory activity was studied in 4 patients, with one of them showing high levels of normetanephrine and norepinephrine.

Table 5.

Patients with multiple paragangliomas (n = 12).

Age (y)	Sex	PG type	Benign / Malignant	PG treatment	Synchronous or Metachronous	Time difference if metachronous	Sequela	Family history	Genetic testing	Secretory
30	M	L – CBT R – CBT R – PGV R – PGM	Benign Benign Benign Benign	Surgery (1st) Surgery Observation Surgery	Synchronous	N/A	Right V, VII, IX, X, XI, XII cn palsy	Father bilateral CBT Uncle and cousins PG	Not studied	Not studied
37	F	L – PGV R – CBT	Benign Benign	Surgery (1st) Out Surgery	Metachronous	2 years	Left IX, X, XI cn palsy	Father bilateral CBT Sister 4 PG and Pheo Brother with 4 PG	SDHD c.57delG	No
31	F	R – CBT L – PGV	Benign Benign	Surgery (1st) Surgery	Synchronous	N/A	Left X, XII cn palsy	None	Not studied	Not studied
58	M	L – CBT R – CBT	Benign Benign	Surgery (1st) Observation	Synchronous	N/A	None	Sister CBT	Not studied	Not studied
60	F	L – CBT L – PGS	Benign Benign	Surgery (1st) Surgery (1st)	Synchronous	N/A	None	Brother CBT Brother PGJ	Not studied	Not studied
41	F	L – PGV L – PGM	Benign Benign	Surgery (1st) Surgery	Synchronous	N/A	Left X cn palsy	None	Not studied	Not studied
31	F	L – PGJ R – CBT	Benign Benign	Surgery (1st) Out Surgery	Metachronous	6 years	Left VII cn palsy	Cousin PG	Not studied	Not studied
38	M	R – CBT L – CBT	Benign Benign	Surgery (1st) Surgery	Synchronous	N/A	None	None	Not studied	Not studied
38	M	RP – PG L – CBT	Benign Malignant*	Surgery (1st) Surgery + RT	Synchronous	N/A	Left X cn palsy	Aunt and 2 cousins PGs	Not studied	NME and NEE
53	M	L – CBT L - PGV	Benign Benign	Observation Observation	Synchronous	N/A	None	Father, brother, nephew carriers of mutation but unknown presence of tumors	SDHB 423+1 G>A	Not studied
60	F	L – PGM L – PGV	Benign Malignant*	Surgery (1st) Out Surgery + CTRT	Metachronous	12 years	Left X and XII cn palsy, first bite sd	None	Not studied	No
47	F	R – CBT L – CBT L – PGV	Benign Benign Benign	Surgery (1st) Observation Observation	Synchronous	N/A	None	None	None	No

^*Malignant because of positive lymph nodes

Abbreviations: y, Years; M, male; F, female; PG, paraganglioma; L, left; CBT, carotid body tumor; R, right; PGV, PG vagale; PGM, mediastinal PG; PGS, sympathetic chain PG; PGJ, PG jugulare; NME, normetanephrine; NEE, norepinephrine; RP - PG, retroperitoneal PG; Out, outside institution; RT, radiotherapy; CT, chemotherapy; cn, cranial nerve, Pheo, pheocromocytoma; SDH, succinate dehydrogenase.

Eight patients had a malignant PG (7.8%). The most frequent criterion for malignancy was the presence of metastatic lymph nodes (n = 4, 50%; 2 were CBT and the other 2 were PGV), and all of them were treated with surgery and adjuvant RT. Both cases with locally infiltrative pattern (1 CBT and 1 non-CBT, were also treated with surgery and adjuvant RT). For patients that were considered to have a malignant PG because of distant metastasis, both occurred in patients with CBT. One of the distant recurrences was treated with CTRT, and the other was managed through RT, embolization and surgery.

During the study period, only 2 patients developed a local recurrence, there were no regional recurrences, and other 2 patients were diagnosed with distant metastasis (as described above). Both local recurrences were seen in patients with a GJ treated surgically, and the recurrences were localized in the skull base. With a median follow-up of 3 years (IQR 1-6), no deaths attributable to the HNPG were encountered. At last physician’s assessment, out of the 79 patients treated surgically, 72 patients (91.1%) had no evidence of disease, and 7 (8.9%) were alive with disease (three had residual disease after incomplete initial removal, two had local recurrence, and two had distant metastases).

DISCUSSION

HNPG are classified based on their anatomical location with CBT comprising the most common location. In our series, CBTs were encountered in 66% of all HNPG, a figure comparable to recent reports in the literature citing 45-75%.^6,9–12 Incidences of other less common HNPG such as PGV range between 7-40%, being 23.3% in our study.^6,9 Variability in the literature of reported incidences of the less frequent HNPG may be reflective of selection biases inherent to the individual series. Similar results than other authors in terms of age and sex distribution were found in our series.^10,13–18 A female predominance of nearly 2:1 was noted, being more pronounced for non-CBT.^5,16 The average age at presentation in all subgroups was the fifth decade of life, an observation consistent throughout most case series.^6,9 The most common form of presentation was the presence of an asymptomatic neck mass, followed by hoarseness, incidental finding, and dysphagia.¹⁰ For CBT, the second most common form of presentation was an incidental finding, indicating that CBTs are frequently asymptomatic upon initial evaluation.⁹ Less than 40% of patients present with any cranial nerve deficit.^5,6,16,19,20 In our series, only one patient from the CBT group presented with a cranial nerve deficit (1.5%), while in the non-CBT group, up to 41 cranial nerve deficits were reported in the 35 patients.

Historically, HNPG were considered locally aggressive tumors requiring surgical excision. However, several studies have shed light into the fact that these tumors are slow growing and indolent, remaining asymptomatic and quiescent for a prolonged period of time.^6,20,21 Combined with the fact that surgery will often result in resection of vital neurovascular structures, especially in non-CBT, resulting in permanent morbidity, it becomes obvious why a paradigm shift in treatment towards non-surgical options has emerged. In our series, no deaths attributable to the HNPG were encountered. Moreover, we observed a trend towards a higher percentage of patients being observed through the decades, and a decreasing percentage of patients undergoing upfront surgery. Amongst patients selected for observation, only those with rapid increase in size or symptoms were shifted to active treatment. In a large series of PG from a single institution studied across 60 years, the authors observed the same trend of change in treatment strategy.²²

A major finding that allows us to differentiate CBT from non-CBT is the fact that the rate of adverse functional outcomes is significantly lower for the CBT group, which has also been demonstrated by other authors.^5,9–11,23 In our series, overall, patients with CBT showed a lower percentage of sequelae compared to patients with non-CBT. Furthermore, if we only analyze the surgically treated patients, 20% of the CBT cohort had a cranial nerve deficit as sequela of the surgery compared to 95% of the non-CBT cohort. Comparatively, complications of HNPG after radiation therapy appeared to be much better tolerated and we observed a 100% control rate (arrest of growth) with radiotherapy in our series. Several groups have also indicated excellent local tumor control, with rates between 85% and 100%.^24–31 Moreover, of the 20 patients that were initially observed, 70% remained stable over the years or with minimal growth, and only 30% needed further active treatment due to rapid growth or clinically significant symptoms. Other authors have found similar percentage of patients that remain stable throughout the period observation, ranging from 42-79%.^6,32,33 These findings guide us towards considering non-surgical options for non-CBT.

For patients undergoing surgery, the need for preoperative embolization remains controversial, with some authors advocating that it reduces the amount of blood loss and the sacrifice of vascular structures, while others defend that it is not recommended for CBT because of the inability to adequately embolize the tumor-feeding vessels, since they mainly arise from the adventitia surrounding the common, external and internal carotid arteries.^10,12,34 In our study, pretreatment embolization was performed in only 12.7% of patients, being more common in the non-CBT group (30%). There were no differences in the comparison of EBL between patients who underwent pretreatment embolization and those who did not. Also in patients selected for surgery, there was a significant difference of the specimen size between CBT and non-CBT, suggesting that the patients with non-CBT who were selected for surgery were those who had a large tumor with clinical symptoms or those who were observed and significantly grew. Moreover, non-CBT patients had a significantly longer length of hospitalization probably due to the higher rates of cranial nerve deficits and postoperative complications in these patients.

Multicentricity rates in the literature range from 7% to 58%,^1,6,19 with reports of much higher incidence for familial PG.^35–41 In our series, approximately 12% presented with multiple PG in the entire cohort, and it increased to 46.7% in the subset of patients with a family history of PG. A high percentage of patients with a family history of PG may harbor a paraganglionic sensitizing mutation.⁵ Mutations usually occur in the succinate dehydrogenase (SDH) gene, with mutations in SDH-B, SDH-C, and SDH-D being the most common. Review of the literature suggests that predictors for germline mutations are male sex, younger age (< 40 years), multicentricity, malignant HNPG, or family history of PG. Therefore, in these patients, genetic analysis of SDH genes and subsequent genetic counseling are recommended.^12,42

For secretory and malignant tumors surgery is still the mainstay of treatment. Secretory PG are rare in the head and neck, being reported in less than 5% of cases, matching our results.^1,3,10,43,44 Generally, PG are considered malignant only if metastases to non-neuroendocrine tissue are demonstrated. Rates of malignancy range between 2-19% in the literature, with differences depending on site (2–4% for PGJT, 6% for CBT, and 16–19% for PGV). ^12,44–47 In our series, the rate of malignancy was approximately 8%, being similar between CBT and non-CBT groups (7.4% vs 8.6%).

In summary, it is now recognized that up-front surgical management is often not the optimal approach for HNPG.⁴⁸ Multiple factors should be taken into consideration when deciding the most appropriate treatment, such as age, comorbidities, prior treatment, likelihood of developing additional lesions, and preexisting cranial nerve deficiencies. Primary surgery should be considered in young healthy patients with small to medium size CBT, secreting tumors, when there is a concern for malignancy, when we observed rapid growth or increase in symptomatology, and when radiotherapy cannot be offered. Conversely, for the remaining cases, observation with serial imaging can be offered in order to avoid unnecessary morbidity while still providing acceptable tumor control.

The present study has inherent limitations due to its retrospective nature, and it is therefore susceptible to biases. Unfortunately, given the incidence of HNPG, the feasibility of a well-controlled prospective review is not possible. Moreover, the need for such a study may not be necessary as more recent population studies demonstrate the indolent course of these tumors. As treatment modalities change in the coming years, further population-based studies from other cohorts will help corroborate the Dutch findings.

CONCLUSIONS

Within HNPG, CBT and non-CBT are separate entities, with different presentation, optimal treatment approach and sequela. A trend to distinguish them and guide our decisions accordingly has been seen through the years. Our experience in the management of HNPG has shown that surgery is an effective treatment for CBT but in non-CBT surgery is associated with significant speech and swallowing functional impairment. Therefore, nonsurgical treatment either with active surveillance or radiotherapy should be considered as an alternative in patients with non-CBT. A better understanding of the natural course of HNPG, has resulted in a more conservative management.