SDH indicates succinate dehydrogenase; SDHA, succinate dehydrogenase subunit A; SDHB, succinate dehydrogenase subunit B; SDHC, succinate dehydrogenase subunit C; SDHD, succinate dehydrogenase subunit D.
Customize your JAMA Network experience by selecting one or more topics from the list below.
McCrary HC, Babajanian E, Calquin M, et al. Characterization of Malignant Head and Neck Paragangliomas at a Single Institution Across Multiple Decades. JAMA Otolaryngol Head Neck Surg. 2019;145(7):641–646. doi:10.1001/jamaoto.2019.1110
What are the genetic trends associated with malignant paragangliomas, and what are the roles of surgery, irradiation, and chemotherapy in the treatment of these tumors?
In this cohort study of 70 patients with paragangliomas at any body site, 17 had malignant paragangliomas, and 5 were found to have succinate dehydrogenase subunit B mutations. Surgery is the mainstay of therapy, with 4 patients found to have malignant disease after selective neck dissection.
Genetic evaluation should be integrated into the care of patients with paragangliomas, and a selective neck dissection should be performed at the time of initial resection owing to the inability to detect malignant disease prior to surgery.
Malignant head and neck paragangliomas (HNPGLs) are rare entities, and there are limited data regarding optimal treatment recommendations to improve clinical outcomes.
To classify succinate dehydrogenase (SDH) germline mutations associated with malignant HNPGLs, evaluate time from diagnosis to identification of malignant tumor, describe locations of metastases and the functional status of malignant HNPGLs, and determine the role of selective neck dissection at the time of initial surgical resection.
Design, Setting, and Participants
A retrospective cohort study was completed of patients diagnosed with paragangliomas on various sites on the body at an academic tertiary cancer hospital between the years 1963 and 2018. A subanalysis of HNPGLs was also completed. Data regarding diagnosis, gene and mutation, tumor characteristics and location, and treatments used were reviewed between February 2017 and March 2018.
Main Outcomes and Measures
Mutations of SDH genes associated with benign and malignant HNPGLs, treatments used, time to the discovery of malignancy, and location of metastasis.
Of the 70 patients included in the study, 40 (57%) were male, and the mean (SD) age was 47 (21.1) years. Of patients with tumors isolated to the head and neck, 38 (54%) had benign HNPGLs, which were associated with mutations in the genes SDH subunit B (SDHB) (n = 18; 47%), SDH subunit C (n = 2; 5%), and SDH subunit D (n = 18; 47%). Among those with malignant HNPGLs, all but 1 patient had mutations in SDHB (n = 5; 83%); 1 patient had no mutation associated with their disease. The average age at diagnosis for malignant HNPGLs was 35 years, while benign tumors were diagnosed at an average age at 36 years. All patients with malignant disease underwent surgery. Four patients were found to have metastasis at the time of selective neck dissection. Among patients with malignant HNPGLs, 5 (83%) were treated with adjuvant radiation, and 1 (17%) was treated with adjuvant chemotherapy.
Conclusions and Relevance
Malignant HNPGLs are rare entities that are difficult to diagnose and are typically identified by the presence of regional or distant metastasis. The results of this study found the prevalence of malignant HNPGLs to be 9%. These data suggest that it is beneficial to perform a selective neck dissection at the time of tumor excision. All patients with malignant HNPGLs but 1 had SDHB mutations.
Paragangliomas (PGLs) are hypervascular tumors of neuroendocrine origin that arise in extra-adrenal autonomic paraganglia. They may be functional (eg, catecholamine secreting) or nonfunctional. Functional PGLs are typically associated with sympathetic paraganglia found in the thorax, abdomen, and pelvis, whereas their nonfunctional counterparts are frequently associated with parasympathetic paraganglia. Paragangliomas may be sporadic or part of an inherited syndrome, with up to 50% considered familial in nature.1
In the head and neck, PGLs are most commonly encountered at the bifurcation of the carotid artery and involve the carotid body, but they can also arise in the middle ear (eg, jugulotympanic tumors), along the track of the vagus nerve (eg, glomus vagale), or very rarely in the larynx.2,3 Head and neck paragangliomas (HNPGLs) typically present as slow-growing, painless lateral neck masses but can progress and manifest with deficits in cranial nerves VII, IX, X, XI, and XII; dysphagia; hoarseness; or Horner syndrome.4,5 Although a majority of HNPGLs are nonfunctional, recent evidence suggests that up to one-third of these tumors are biochemically active; thus, proper evaluation of urine or serum catecholamines should be included in the examination of these patients.6 Many HNPGLs are benign tumors, but up to 19% may be malignant in nature.7 Malignancy is determined by the presence of metastases to nonendocrine tissue, because histologic characterization appears to be insufficient to determine malignancy.8
The majority of hereditary PGLs are found to be associated with a mutation in genes encoding succinate dehydrogenase (SDH) subunits (or cofactors), an enzyme that contributes to both the tricarboxylic acid cycle and the electron transport chain as part of mitochondrial complex II. Paraganglioma syndrome type 1 is specifically associated with mutations in the SDH subunit D (SDHD) gene, while the SDH subunit B (SDHB) gene is identified as the susceptibility gene for PGL syndrome type 4.9,10 Importantly, the type of SDH mutation can vary based on the location of the tumor and if there is an increased risk of cancer. Previous studies have identified that SDHB gene mutation carriers are specifically more prone to developing malignant disease.11 However, PGLs can also be associated with mutations in other genes associated with hereditary cancer risk, such as VHL, MEN1, RET, SDHAF2, and more recently, MAX and TMEM127.12
The purpose of this study is to classify SDH germline mutations that are associated with malignant HNPGLs, evaluate the time from diagnosis to discovery of malignant tumor, describe locations of metastases and the functional status of these tumors, and determine the role of a selective neck dissection at the time of surgery.
Patients diagnosed with a PGL at any body location and who underwent genetic testing or counseling were included in the cohort of patients. Among patients with PGLs, we further evaluated patients with HNPGLs, including carotid body tumors, jugulotympanic tumors, and glomus vagale tumors. Patients included in the study were seen at the Huntsman Cancer Institute between the years 1963 and 2018. Eligibility criteria included mutation of an SDH gene (SDHA, SDHB, SDHC, SDHD, or SDHAF2) or presence of a malignant HNPGL. There were no age restrictions for participants.
Patients were then seen by a head and neck surgeon after the initial diagnosis was made. Patients with malignant PGLs were considered those with demonstrated disease in lymph nodes, bone, or distant organs. Patients underwent testing to assess the functional status of the tumors, including plasma or urine catecholamines, metanephrines, chromogranin A, and VMA. All patients with HNPGLs, both benign and malignant, were analyzed to abstract information regarding age at diagnosis, type of HNPGL, functional status of tumor, mutation analysis, surgical interventions, adjuvant therapy, pathology, and information regarding familial trends of disease. All patients with SDH gene mutations were enrolled in the Cancer Genetics Study at Huntsman Cancer Institute. Institutional review board approval was given by the University of Utah, and written informed consent was obtained from all study participants.
A total of 70 patients diagnosed with a PGL at any body location and who had an SDH gene mutation or malignant disease were included in the study. Among these participants, 17 (24%) were found to have malignant tumors, with 6 (9%) malignant PGLs isolated to the head and neck (Table 1). Of the tumors isolated to the head and neck, 38 (54%) were found to be benign HNPGLs (Table 2). Among participants with malignant tumors of the head and neck, all but 1 had mutations of the SDHB gene, with the following mutations identified: R230L, P197R, R230H, R242C, and 423+1G>A. Another patient diagnosed at 16 years old had no identifiable mutation found after undergoing a multigene test that evaluated all known genes associated with PGL. The patient ultimately had negative results for all genes both with sequencing and duplication/deletion studies. Patients with non–head and neck malignant tumors had mutations associated with the genes SDHA (n = 1; 9%), SDHB (n = 8; 73%), and SDHD (n = 2; 18%) (Figure 1). Benign HNPGLs were associated with mutations in the genes SDHB (n = 18; 47%), SDHC (n = 2; 5%), and SDHD (n = 18; 47%) (Figure 2).
The average age at diagnosis for malignant HNPGL tumors was 35 years (range, 16-65 years), while diagnosis of benign HNPGLs occurred at an average age at 36 years (range, 32-41 years). Among participants with malignant tumors, 3 (50%) were male and 3 (50%) were female, while among patients with benign tumors, 20 (53%) were male and 18 (47%) were female. Imaging used for surveillance of all HNPGLs included the following: magnetic resonance imaging (MRI) of the neck (n = 11; 26%), computed tomography of the neck (n = 10; 22%), MRI of the whole body (n = 8; 18%), MRI of the brain (n = 6; 14%), neck ultrasound (n = 2; 5%), positron emission tomography–computed tomography (n = 1; 2%), and other distant disease surveillance, including MRI of the abdomen, pelvis, and chest (n = 6; 13%).
Only 1 patient had a malignant tumor on diagnosis, and the average time to the discovery of malignancy was approximately 1 year. Locations of malignant tumors in the head and neck included paracervical (n = 1), glomus vagale (n = 2), and carotid body (n = 3). All patients with malignant HNPGLs underwent excision with selective neck dissection. Positive lymph nodes were found in 5 patients, with another patient found to have lymph node involvement on recurrence. Two patients were known to have regional spread prior to surgery, while 4 patients were found to have lymph node involvement at the time of selective neck dissection. Distant bony metastases were found in 3 patients with malignant HNPGLs. Among patients with malignant HNPGLs, 5 (83%) were treated with adjuvant radiation and 1 (17%) was treated with adjuvant chemotherapy. Only one patient was found to have a malignant HNPGL tumor that was functional and positive for chromogranin A. There was 1 death reported among patients with malignant disease, and 2 patients had disease recurrence after initial treatment. There was minimal loss to follow-up among this cohort of patients; however, patients with known malignant disease did have closer follow-up than those with benign disease.
Although there has previously been a paucity of literature on malignant HNPGLs, this retrospective cohort study of patients with PGLs followed some patients for more than 50 years. Malignant HNPGLs are rare entities, with literature citing a frequency of malignancy between 0.016% and 19%.7,8,13-16 When PGLs arise in the head and neck, they are commonly found at the carotid bifurcation; this study similarly shows a higher rate of malignant HNPGLs arising from the carotid body. The data in this study strengthen the existing body of literature on malignant HNPGLs because of the extensive follow-up time and tracking of families with SDH mutations. Owing to the rare nature of malignant HNPGLs, guidelines for the appropriate surgical management of these tumors has continued to evolve.
Historically, surgery has been the mainstay of therapy for HNPGLs. All of the patients presented in this cohort underwent surgical excision of their masses and selective neck dissection. Our recommendations for a selective neck dissection at the time of resection originates from 2 important concepts. First, determining malignancy may not be possible until pathology specimens are reviewed. Currently, there are no cytological, histological, immunohistochemical, or molecular criteria that can classify malignant PGLs prior to surgery.17 The second impetus for completing a selective neck dissection is that appropriate treatments may be delayed without pathologic evidence of malignancy, which may lead to further progression of disease, particularly if there is poor follow-up. Although prior studies have recommended a comprehensive neck dissection only after evidence of metastatic disease during tumor resection,8 recent literature does support selective neck dissection, including levels IIA, IIB, and III.18
Over the years, there has been growing evidence that there is a role for adjuvant radiation for the treatment of malignant tumors. Typically, treatment recommendations are made based on the location, extent of disease, number of PGLs present, and medical comorbidities.19 Hinerman and colleagues19 evaluated the use of external radiation therapy among a cohort of 104 patients with HNPGLs, of which 6 patients were found to have malignant disease. Among these patients, there was a total of 6 recurrences noted, but none of those recurrences was above the clavicle. The authors did suggest, however, that patients with malignant PGLs require higher doses of radiation therapy. Another study by Moskovic and colleagues20 evaluated a series of 14 patients treated with radiation therapy, with 12 of those patients ultimately developing disease progression. Thus, there is still some lack of clarity of the benefit of radiation therapy alone or with surgery for the treatment of malignant HNPGLs. All but 1 of the patients in the present study cohort received radiation therapy. One mortality was observed among patients undergoing surgical resection and radiation, with this patient’s care being complicated by an atypical meningioma and recurrence of a malignant PGL at a lumbar vertebra.
There is a growing body of evidence regarding the use of chemotherapy agents for the treatment of HNPGLs. Patel and colleagues21 reported a 15-year experience of using chemotherapy to treat patients with PGLs, ultimately concluding that chemotherapy should be used in patients with metastatic or unresectable disease. However, the present study only included 2 patients with metastatic HNPGLs, with 1 patient showing disease progression and the other showing no change after treatment with chemotherapy. Lee and colleagues8 also reported their outcomes with chemotherapy when used in combination with surgery and radiation, or used as monotherapy. The findings of their study suggested that chemotherapy should be reserved for patients with rapidly growing tumors, which is unlike most malignant PGLs that grow at a slow rate. Ultimately, overall survival may not be largely affected by the use of chemotherapy, but there is some evidence that it assists with reducing symptoms or improving short-term disease outcomes.22 In the present cohort, only 1 patient received chemotherapy in the treatment of their metastatic disease, and this was the only patient who had a functional tumor. This patient also received chemotherapy treatment in the setting of an aggressive recurrence after primary surgical resection. To date, this patient has been found to have stable disease after several rounds of treatment with cyclophosphamide, vincristine, and dacarbazine.
This study also adds to the body of literature regarding the genetics of PGL tumors. In terms of benign HNPGLs, Neumann and colleagues23 analyzed SDH mutations among a cohort of 598 patients, with the SDHD mutation found to account for 51.4% of all benign HNPGL mutations. Among the present cohort of 38 patients with a benign HNPGL, the distribution of SDH mutations were as follows: SDHB, 47%; SDHC, 5.3%; and SDHD, 47%. These data are consistent with prior publications that have found that malignant HNPGLs are frequently seen in patients with SDHB mutations.24 All of the patients with malignant disease in the present cohort of patients were found to have the SDHB mutation except for 1 patient (Figure 1). This patient with no mutation fits into the 10% of patients younger than 20 years who may have no identifiable mutation.
Ultimately, all patients diagnosed with a PGL benefit from counseling with genetic counselors for hereditary PGLs and malignancy risk, regardless of the disease site or age at diagnosis. Recent literature suggests that genetic testing is paramount if the risk of a heritable mutation is greater than 10% or if genetic testing affects treatment recommendations.25 Among head and neck tumors, should the patient test positive for the SDHB mutation, this ultimately could make surgery the most realistic treatment option owing to the inherent risk of the tumor being malignant.
Although this study strengthens the current literature on malignant HNPGLs, there are several limitations. This study followed some patients in the cohort for more than 50 years, but there is still a small sample size of only 6 patients with malignant HNPGLs, which limits further statistical analysis regarding trends among these malignant tumors. The largest cohort of malignant HNPGLs to date is 59 patients from the National Cancer Database.8 Although the sample size of the present study is small, the experience is reflective of how rare malignant HNPGLs are and provides insight into the typical experience of a large academic institution. Another limitation of this study is that the follow-up time for patients varies, particularly among patients with benign disease. Part of this is related to the large catchment area our institution serves, including 6 surrounding states. Because of long distances to see specialists, many patients may have opted to follow up with local otolaryngologists, which may have limited the overall evaluation of the long-term outcomes among this cohort of patients. Nonetheless, we are able to provide a clinical picture of the outcomes among patients with PGLs across a large geographical area. For this rare disease, future studies should aim to collect prospectively defined key clinical, tumor, and treatment data from investigators at multiple institutions and assess treatment efficacy with a randomized clinical trial nested within a multi-institutional cohort design.
We present a cohort of 70 patients with PGLs arising from various sites of the body, including 38 that were isolated to the head and neck. The prevalence of malignant HNPGLs was 9%, with a total of 6 patients found to have malignant disease. Surgery is the mainstay of treatment among these rare tumors, and we ultimately recommend a selective neck dissection at the time of surgery owing to poor criteria for diagnosing malignant disease prior to surgery. Genetic testing and counseling should be integrated early in the care of these patients owing to their influence on future medical and surgical recommendations.
Accepted for Publication: April 6, 2019.
Corresponding Author: Luke Buchmann, MD, Division of Otolaryngology–Head and Neck Surgery, Department of Surgery, University of Utah School of Medicine, 50 N Medical Dr, SOM 3C120, Salt Lake City, UT 84132 (firstname.lastname@example.org).
Published Online: June 13, 2019. doi:10.1001/jamaoto.2019.1110
Author Contributions: Drs Buchmann and McCrary had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: McCrary, Calquin, Carpenter, Casazza, Naumer, Kohlmann, Cannon, Monroe, Buchmann.
Acquisition, analysis, or interpretation of data: McCrary, Babajanian, Calquin, Casazza, Naumer, Greenberg, Kohlmann, Cannon, Monroe, Hunt, Buchmann.
Drafting of the manuscript: McCrary, Babajanian, Calquin, Buchmann.
Critical revision of the manuscript for important intellectual content: McCrary, Carpenter, Casazza, Naumer, Greenberg, Kohlmann, Cannon, Monroe, Hunt, Buchmann.
Statistical analysis: McCrary, Calquin, Casazza, Monroe.
Obtained funding: McCrary.
Administrative, technical, or material support: McCrary, Casazza, Naumer, Greenberg, Kohlmann, Cannon.
Study supervision: Carpenter, Casazza, Cannon, Monroe, Hunt, Buchmann.
Conflict of Interest Disclosures: Dr Monroe reports receiving grants from the National Institutes for Health outside of the submitted work. The Genetic Counseling Shared Resource is supported in part by the National Cancer Institute (grant P30 CA2014) and awarded to the Huntsman Cancer Institute and the Huntsman Cancer Foundation. No other disclosures were reported.
Meeting Presentation: This study was presented at the American Head & Neck Society 2019 Annual Meeting held during the Combined Otolaryngology Spring Meetings; May 1–2, 2019; Austin, Texas.
Additional Contributions: We thank the Huntsman Cancer Institute for providing a great deal of services to our patients, including genetic evaluation. This project would not have been possible without the assistance of our genetic counselors. They were not compensated for their contributions.