Introduction

Paragangliomas and pheochromocytomas (PPGLs) are neuroendocrine tumors derived from the parasympathetic or sympathetic nervous system. They can be either functional or non-functional. Pheochromocytomas (Pheo) arise from the adrenal medulla and paragangliomas (PGLs) from the extra-adrenal paraganglia [1,2,3,4,5]. PGLs arising from head and neck (HNPGL) are usually derived from parasympathetic paraganglia and are non-functional. This contrasts with thoraco-abdominal-pelvic PGLs (TAPPGL), that are usually derived from the sympathetic ganglia, and together with Pheo, are typically functional active, producing an excess of catecholamines [1, 5,6,7].

PPGLs can be either sporadic or familial. Germline mutations in predisposition genes are now found in up to 40% of all PPGLs [1, 7]. Familial cases may be associated with specific syndromes such as multiple endocrine neoplasias (MEN) type 2 (RET mutation), Von Hippel Lindau (VHL) (VHL mutation), neurofibromatosis type 1 (NF1) (NF1 mutation) and familial PGL/pheochromocytoma [mutations in succinate dehydrogenase complex (SDHx)] [2, 6, 8]. Mutations in other genes such as those encoding fumarate hydratase (FH), transmembrane protein 127 (TMEM127) and MYC-associated factor X (MAX) are also associated with familial PPGLs [8,9,10,11]. Novel genes are continuously being added to this list.

Germline mutations in SDHx are responsible for up to 30% of the cases [2, 7, 12, 13]. Four distinct syndromes – PGL1–4 – are related to mutations in the different SDHx subunits, but the precise genotype–phenotype correlation is still uncertain [2, 3]. The most common SDHx genetic defects in the Portuguese population are not completely described.

Despite the majority of PPGLs cases being benign, 10–20% of the cases can be malignant. There are some germline variants (like those affecting the SDHB subunit) and particular anatomic locations (such as TAPPGLs) that are linked to malignancy [1, 4, 13,14,15].

Our general objective was to characterize the clinical features and genotype-phenotype associations in a group of SDHx-mutated PPGL patients. Our specific objectives were to study the most frequent genetic alterations, to characterize the clinical features and to study treatment outcomes in SDHx-mutated PPGL patients.

Material and methods

A retrospective analysis of all germline SDHx-mutated PPGL cases followed in our center was performed. The cases were retrieved from SDHx genetic tests performed at our institution between 1986 and October 2016. Genetic tests were performed using Sanger sequencing or next-generation sequencing (NGS) for the detection of missense and indel mutations, and multiplex ligation-probe amplification (MLPA) for the detection of large deletions and insertions.

Information was gathered based on clinical records, laboratory results, imaging exams, pathology reports, and genetic tests results.

Confidentiality was kept throughout the study and all patients or representatives signed informed consent. This study was approved by our center’s ethics committee and conducted in accordance with the Declaration of Helsinki. The authors have followed the protocols of their center on the publication of data.

Clinical status and treatment outcomes at last follow-up visit were established using the Response Evaluation Criteria in Solid Tumors (RECIST). Excessive catecholamine secretion was assumed when more than two-fold of total and/or fractioned urinary metanephrines, and/or total and/or fractioned urinary catecholamines and/or urinary vanillylmandelic acid was observed. Multifocality and not metastases was assumed when tumors were located at usual PGL sites. Malignancy or metastatic potential was defined by the presence of metastases or a gross invasion of adjacent structures.

Data management and statistical analysis were performed using Excel Software Microsoft® and SPSS 23 IBM®. Categorical variables were compared using Chi-square and Exact F statistic tests. Numeric variables were compared using the Mann–Whitney statistic test.

Results

Thirty-one cases were first investigated but one case was excluded due to an early loss to follow-up. Therefore, 30 patients were included in the study for further analysis.

Clinical and genetic characterization

Clinical and genetic characteristics of the PPGLs patients included in the study are presented in Table 1.

Table 1 Clinical and genetic characteristics of PPGLs patients
Full size table

At presentation, 96.7% (29/30) of the patients had symptoms, mainly hypertension or a cervical mass.

Excessive catecholamine secretion was detected in 33.3% (10/30) of the patients, unknown or undetermined in 10% (3/30) and doubtful in 6.7% (2/30). Catecholamine secretion was within the normal range in 50% (15/30).

There was no difference between age at diagnosis according to the affected SDHx subunit (p = 0.15).

Anatomical distribution

Anatomical distribution was assessed in 86.7% (n = 26) of the cases by computed tomography (CT) scan, in 63.3% (n = 19) by magnetic resonance imaging (MRI) and in 50% (n = 15) by both imaging modalities. All patients that performed MRI had HNPGL, except for two who had abdominal PGL (one was a 15 years-old male and the other a female in reproductive age). Octreoscan was performed in 50% (n = 15) of the patients, metaiodobenzylguanidine (MIBG) scan in 30% (n = 9), 68Gallium-DOTANOC positron emission tomography/computed tomography (68Ga-DOTANOC PET/CT) in 26.7% (n = 8) and 18F-fluoro-deoxy-glucose PET/CT (FDG-PET/CT) in 6.7% (n = 2).

Table 2 shows the sensitivity of each nuclear medicine exam in evaluating primary lesions.

Table 2 Sensitivity of nuclear medicine exams in evaluating primary lesions
Full size table

As for the HNPGL, 57% (12/21) of the patients had a carotid PGL, 33% (7/21) a jugulotympanic PGL and 29% (6/21) a vagal PGL (some patients had HNPGL in more than one location). The majority of TAPPGLs were in the abdomen in 57.1% (4/7), followed by mediastinum in 28.6% (2/7) and pelvis in 14.3% (1/7).

Concerning multifocality, 16.7% (3/18) of SDHB, 55.6% (5/9) of SDHD and 0% (0/3) of SDHC affected patients showed a multifocal PPGL.

Table 3 shows an overview of the SDHx distribution of PPGL cases by anatomical site, catecholamine secretion, malignancy, and multifocality rates.

Table 3 SDHx distribution of PPGL cases by anatomical location, catecholamine secretion, malignancy, and multifocality
Full size table

Malignancy

From all patients studied, 30% (9/30) were considered malignant. HNPLGs were malignant in 19% (4/21), TAPPGLs in 83% (5/6) and Pheo in 60% (3/5) of the cases. Metastatic SDHB patients involved bone in 75% (6/8), lung in 62.5% (5/8), liver in 50% (4/8), lymph nodes in 37.5% (3/8) and kidney in 12.5% (1/8) of the cases. As for the metastatic SDHD patient, he had bone, lung and liver metastases. Two patients with bone metastases developed pathological fractures and spinal cord compression. For metastases control, surgery was performed in two patients and radiotherapy in four.

Genetic findings

A positive family history of SDHx-related PPGL was present in 20% (6/30) of the patients.

The most frequent genetic findings in this population were SDHB exon-1 deletions, affecting 26.7% (8/30) patients, followed by SDHD exon-4 deletions, affecting 16.7% (5/30) patients. No direct familial relationship was found between PPGL patients.

All the identified mutations are listed in Table 4.

Table 4 Germline mutations of SDHB, SDHC, and SDHD found among the 30 PPGL patients studied
Full size table

Associated tumors

Five patients had benign thyroid nodules (two with SDHB germline mutations and three with SDHD), two patients had papillary thyroid cancer (one case was SDHB germline mutated and also had intestinal polyposis—and the other had a SDHD germline mutation), one had breast cancer and multiple myeloma (SDHB mutated), one had adrenal adenoma (SDHC germline mutation) and one had a uterine myoma (SDHB mutated).

Treatment outcomes

Surgery was performed in 80% (24/30) of the patients. Seven primary PGL lesions from six patients were not submitted to surgery (five carotid HNPGL, one thoracic and one pelvic PGL). One patient with a nonfunctioning bilateral carotid HNPGL was not submitted to surgery or any other intervention and was selected for watchful waiting. He showed stable disease during follow-up. There was a complete resection (R0) rate in 60% (3/5) of the Pheo, 50% (3/6) of the TAPPGL and in 33.3% (7/21) of the HNPGL. After R0 surgeries all patients achieved remission, except two, who were metastatic ab initio. All patients with R1 and R2 surgeries have persistence of disease except one who had a R1 surgery and is in remission.

Pre-surgical vascular embolization was performed in 43.3% (13/30) of the patients (11 HNPGL, one TAPPGL and in hepatic metastases of one patient), with a success rate (defined by post-embolization angiographic vascularization pattern) of 46.2% (6/13).

As for reoperations, 23.3% (7/30) of patients needed more than one surgery and 10% (3/30) of patients needed three surgeries. The causes for multiple surgeries were multiple PGLs in three patients, local relapses in three (all SDHB mutated) and distant metastases in the other three.

Radiotherapy was used in 40% (12/30) of the patients, intensity-modulated radiotherapy (IMRT) in 36.7% (11/30) and Gamma Knife in 3.3% (1/30). Radiotherapy was used in eight HNPGL, five bone metastases, one pelvic PGL and one lung metastasis. All lesions treated with radiotherapy showed stability or partial response, and none progressed during a median follow-up of 26 months (minimum 7.6; maximum 125).

Two patients were treated with 177Lutethium peptide receptor radionuclide therapy (177-Lu-PRRT) and other two patients with 131-MIBG. All these patients had distant metastases and progressed after treatment.

After all treatment modalities, complete remission was observed in 33.3%, stable disease/partial response in 53.3% and disease progression in 13.3% of the patients. During follow-up, one disease-related death occurred, and significant disease and treatment-related morbidities were observed in 66.6% of the cases.

Table 5 shows the correlation between SDHx mutated subunit and treatment outcomes.

Table 5 Outcomes of various treatment modalities
Full size table

The most frequent complication was permanent cranial nerve deficit, found in 13 patients (61.9%) of the HNPGL, from which 9 were surgery-related. Pathologic bone fractures occurred in three patients, and nephrectomy and surgery-related gastrointestinal ischemia/bleeding occurred in two patients each. The other complications found in one patient each were pelvic neurological deficit, chronic bicytopenia related to 131-MIBG treatment, renal vein thrombosis, surgery-related pneumothorax and surgery-related abdominal abscess.

Discussion

In this study, we have characterized the clinical features and genotype-phenotype associations in germline SDHx-mutated PPGL cases followed in our center and we describe the largest series of Portuguese patients with SDHx related PPGL.

In our cohort, women were more frequently affected than men, which contrasts to previous publications [6, 16]. This difference may be justified by the small number of patients included in these studies.

There was no significant difference between age at diagnosis according to the affected SDHx subunit. There is some controversy in the literature regarding this issue since some series [3, 6], but not others [7], report a later onset of disease in SDHB mutated patients.

Germline mutations in SDHD and SDHC were more likely to develop HNPG (all cases in both groups) than SDHB affected patients who were more prone to develop TAPPGL and Pheo. This is in accordance with previous publications [2,3,4,5,6,7, 13, 17].

We also confirmed that patients with SDHD mutation are more likely to develop multifocal PPGLs, whereas SDHB and SDHC mutated patients are more likely to have a unifocal disease [2,3,4, 7, 13].

The percentage of SDHB mutated PPGL with catecholamine secretion was lower than we expected, perhaps because half of them were HNPGL, which are usually non-functional tumors [2]. The routine measurement of 3-Methoxytyramine could have identified additional secretory PGLs, but this test is not routinely available in our center [7, 14].

SDHC mutated patients had non-functional HNPGL, which is in agreement with other publications [7].

We also confirm that SDHB mutations were the most frequently associated with malignancy, followed by SDHD mutations [1,2,3,4,5,6,7, 13, 14, 17, 18].

In agreement with previous publications, the majority of our patients did not have a family history of PPGL. This can be justified by the incomplete penetrance of these mutations, maternal imprinting of SDHD subunit mutations, the possibility of de novo mutations and limited access to specialized health care in the ancestral as suggested by others [3, 7].

The most common genetic alteration was a 15678 bp deletion in exon-1 of SDHB described in northern Portuguese, Galician and Brazilian patients. This highlights the possible founder effect of this mutation in the Portuguese population [4, 16, 17].

Two patients showed the SDHB mutation in exon-2 c.127 G > C (p.Ala43Pro), also described in some French patients [6, 17, 19].

One patient of Dutch origin had a SDHD mutation in exon 3 c.274 G > T, p.Asp92Tyr which is a founder mutation in the Netherlands, responsible for 68% of the cases of hereditary PPGLs in that country [2, 6, 17, 20].

It has been described that patients with SDHx mutations are prone to develop other tumors, such as gastrointestinal stromal tumors, renal carcinomas or pituitary adenomas which were not observed in our cohort [7, 17]. Two of our patients had a papillary thyroid carcinoma and another a breast cancer and a multiple myeloma.

Despite the elevated malignancy rate documented in our study, there was only one death, which can be justified by the slow progression rate of these tumors [4, 14].

Systemic treatment of PPGLs with MIBG and PRRT were disappointing since the four patients submitted to these treatment modalities showed progressive disease after therapy.

SDHB patients and HNPGLs had the worse prognosis, the first related to malignancy, and the latter to cranial nerve deficits, unresectable disease, and multimodality interventions.

Due to the limited number of cases, it is difficult to retrieve major conclusions about the efficacy of different treatment modalities, when analyzed regarding the subtype of SDHx subunit affected. Nevertheless, it is clear that progressive disease after surgery was more common in SDHB patients.

Conclusion

This single center study is the most complete Portuguese cohort in the literature and helps to understand the behavior of tumors based on their genotype and anatomical location.