Clinical presentation of sporadic and hereditary pheochromocytoma/paraganglioma

Introduction

Pheochromocytomas and paragangliomas (PPGL) are rare neuroendocrine tumours of the autonomic nervous system, embryologically originating from the neural crest. While pheochromocytomas (PHEO) arise in the adrenal medulla, paragangliomas (PGL) can occur at various body sites and are traditionally divided into head-and-neck PGL (HNPGL) and PGL located in the thorax, abdomen, or pelvis. The incidence of PPGL is about 0.7 cases per 100,000 persons/year (Leung et al. 2021). The functionality of PHEO is defined by epinephrine, norepinephrine, and/or dopamine overproduction, whereas PGL can only excrete norepinephrine and/or dopamine (except Zuckerkandl organ PGL which secretes epinephrine). Also, some PPGLs are non-functional (most of HNPGLs). These differences in catecholamine secretion patterns can generate a different intensity of symptoms, but in general, most patients with PPGL present with non-specific complaints. Together with increasing use of medical imaging, this leads to an important percentage of incidentally discovered PPGL (~50%) (Gruber et al. 2019).

Importantly, up to 40% of patients with a PPGL have a hereditary predisposition (Lenders et al. 2020) – the highest heritability rate among human tumours. Major suggestive features of hereditary PPGL are familial history of the disease, bilateral cancers affecting paired organs, multiple primary PPGL in the same individual, recurrent or malignant disease, and early age of onset (i.e. <45 years old) (Amar et al. 2005, Lenders et al. 2014). This high percentage of hereditary PPGL makes genetic testing to identify patients and relatives at risk to (re)develop disease and thus enable tumour screening a key step in the management of these rare diseases. Although algorithms for targeted genetic testing might become redundant due to the increased efficiency and lower cost of next-generation sequencing (Ben Aim et al. 2019, Lenders et al. 2020), research on differences in disease presentation between sporadic and hereditary PPGL as well as on possible genotype–phenotype correlations in hereditary PPGL still is important for better surveillance and outcome prediction in patients and thus an important basis for future personalized therapy (Crona et al. 2019, Main et al. 2020).

In this study, we investigated the clinical presentation and disease characteristics of sporadic and hereditary PPGL and evaluated the applicability of genotype–phenotype correlations described in the algorithms for genetic screening in a cohort of Belgian patients.

Materials and methods

A retrospective analysis of electronic medical records of 67 consecutively registered patients diagnosed and/or treated with PPGL in the Ghent University Hospital (Belgium) between 2002 and 2020 was performed. Patients were identified through patient registries of the departments of Pathology, Surgery, and Endocrinology. Patients were divided according to anatomic location (PHEO vs PGL) and according to genetic status (hereditary vs sporadic PPGL; patients without genetic testing results (n = 7, all with PHEO) were not included in these latter analyses). General characteristics (age at diagnosis, gender, anthropometric parameters), method of discovery, clinical phenotype (symptoms, functionality, localization, presence of multiple primary tumours at diagnosis, malignancy (metastases or tumour recurrence), outcome) and hormonal phenotype (metanephrine, normetanephrine and dopamine levels), secretion pattern, and tumour diameter (assessed with CT/MRI scan) were compared between patients with sporadic vs hereditary PHEO/PGL.

Pathway to discovery was categorized as incidental/through familial genetic screening or symptomatic (i.e. due to symptoms related with catecholamine overproduction or local complaints) discovery. Symptoms were grouped into classic triad (headache, palpitations, sweating), non-specific symptoms (hypertension, paroxysms, weight loss, anxiety, vomiting, syncope), tumour location-related complaints (dysphagia, tinnitus, deafness, vertigo, bone pain), and cardiovascular (CV) events (myocardial infarction (MI), Takotsubo cardiomyopathy, arrhythmias, or stroke.) Patients were defined to have multiple tumours at diagnosis if they presented with bilateral disease or distant synchronous tumours. PPGL were considered functional if either plasma or urinary fractionated (nor)metanephrine levels were >1.5 × upper limit of the respective normal (ULN) reference ranges. A noradrenergic secretion pattern was defined as predominant increases of normetanephrine accompanied by metanephrine concentrations <1.5 ULN, whereas an adrenergic secretion pattern was defined as increments >1.5 ULN for both metabolites.

Patients were genetically tested for germline pathogenic variants by ‘sequence by synthesis’ or direct sequencing technology, performed at the Center for Medical Genetics, Ghent University Hospital, Belgium, using KAPA HyperCap technology (Roche Diagnostics) or other depending on the diagnosis date. The gene panel included the most frequently mutated genes in PPGL: RET, VHL, NF1, MAX, SDHA, SDHB, SDHC, SDHD, SDHAF2, and TMEM127. Depending on phenotypic features and/or year of diagnosis, patients were tested for a variable number of these genes. For patients with hereditary and apparently hereditary phenotype (metastatic disease/multiple tumours at diagnosis), we evaluated the applicability of genotype–phenotype correlations, suggested in the algorithm for genetic screening (Lenders et al. 2014), to which we added SDHB immunohistochemistry results. This study was approved by the Ethical Review Board of Ghent University Hospital (BC-10125) and conducted in accordance with the Declaration of Helsinki.

All statistical procedures were performed using Statistical Package for Social Sciences 23.0 software package (SPSS Inc.). Continuous variables were described in terms of mean ± s.d. if their distribution was normal according to the Kolmogorov–Smirnov test and in terms of median and range, otherwise. For statistical analysis, we used non-parametric correlations, Kruskal–Wallis test, Fisher test, and Student’s t-test where appropriate. A two-sided P-value < 0.05 was considered to indicate statistical significance.

Results

Discussion

This study shows that there were no important differences in pathways to diagnosis, clinical, or biochemical phenotype of sporadic vs hereditary PPGL in a cohort of Belgian patients. Genotype–phenotype correlations identified in our cohort are largely similar with those from the literature, although in one patient with multiple PGL locations, no pathogenic germline mutation could be identified. Also, the algorithm for genetic testing (Lenders et al. 2014) did not fit in 3 out of the 24 patients with a documented hereditary disease which confirms that a multiple gene-panel to screen these patients should be considered clinical state-of-the-art.

In this retrospective cohort, we observed a predominance of PHEO cases but more than one-third of our patients presented with a PGL, which is within the range reported in other series (Amar et al. 2005, Leung et al. 2021). Further, 23% of patients with PHEO showed a hereditary predisposition, whereas for PGL, this was 64%. Indeed, it is known that PHEO are more frequently sporadic than PGL (Karasek et al. 2010, Hensen et al. 2011). Among all genetically tested PPGL patients, 40% of patients carried a genetic predisposition which corroborates literature findings in which up to 40% of the PPGL are reportedly caused by a germline mutation, with mutations in SDHD/SDHB/VHL genes being most common (Boedeker et al. 2007, Karasek et al. 2010, Eisenhofer et al. 2011, Lenders et al. 2014, Muth et al. 2019). In our cohort, SDHD mutations were most frequent, followed by SDHB. It should be noted, however, that 7 out of our 67 patients were not genetically tested and that the gene panel used at the time of screening had a variable composition.

Remarkably, we observed a predominant right/left localization in sporadic/hereditary PHEO (and similar but non-significant trends in PGL), which is an aspect that was not described earlier. We did not find any significant correlation between tumour localization (right/left) and tumour diameter, metanephrine levels, occurrence of CV events, or pathway discovery (data not shown) and suppose that these differences in lateralization are coincidental rather than of clinicopathological importance.

The distribution of incidentally/genetically screened vs symptomatically discovered PPGL was equal among both sporadic and hereditary PPGL. These findings emphasize the non-specific character of the symptoms in these diseases, along with an increasing rate of imaging-based diagnosis. In other studies, the incidental or screening-based finding of the tumour was the first mode of diagnosis (Falhammar et al. 2018, Gruber et al. 2019, Cvasciuc et al. 2020). We found that the classic triad was most frequently reported in patients with sporadic PHEO, while in hereditary PHEO, symptoms were more often non-specific but without statistically significant differences. The frequency of classic triad was higher in our cohort than in other studies, where it was reported in 10–28% of cases (Kopetschke et al. 2009, Falhammar et al. 2018, Geroula et al. 2019, Cvasciuc et al. 2020). We also observed more prevalent CV events in patients with PHEO as compared to PGL. Although this might result from diagnostic bias and the higher number of functionally active tumours in the former, future research should address if patients with PHEO really are at increased risk of CV events. In our cohort, only in one patient with Takotsubo cardiomyopathy occurring shortly after abdominal manipulation, a direct link with the PHEO could be presumed.

As expected, patients with hereditary PPGL presented more often with new or recurrent disease (8/24) than those with sporadic PPGL (2/36). This justifies a life-long follow-up for patients with hereditary PPGL. However, one patient with a sporadic PHEO presented with a recurrence after 13 years, supporting the European Society of Endocrinology guideline proposing lifelong follow-up for all high-risk patients (i.e. young patients, or with large tumours, or with PGL) (Plouin et al. 2016).

Regarding genotype–phenotype correlations in those patients with hereditary PPGL, we found that anatomical distribution of the tumours and secretion pattern according to the mutated gene largely correspond to the literature. Specifically, patients with SDHD mutations presented mostly with HNPGL (9/13), a presentation also described in a Dutch cohort (Hensen et al. 2011). Most SDHD-related PPGL are benign, which is confirmed by our findings. One out of 10 patients with SDHD mutation had metastatic disease after a median follow-up duration of 12.1 years. In patients with pathogenic SDHB mutations, PHEO occurrence and abdominal/pelvic PGL (7/8) were more frequently observed, also as previously reported (Muth et al. 2019). In contrast to other reports where the malignancy rate was relatively high (van Hulsteijn et al. 2012, Tufton et al. 2017, Main et al. 2020), none of these eight patients developed malignant disease, although the median follow-up was only 7 years.

Tumour SDHB expression as determined by immunohistochemistry largely corresponded with the literature on patients with hereditary disease (van Nederveen et al. 2009). However, in eight PHEO patients with negative SDHB immune-staining, no pathogenic mutation in SDHx-genes could be demonstrated. In one of these patients, a somatic SDHB mutation was found, three patients had metastatic disease, and four did not have a hereditary phenotype based on clinical and anamnestic observations. Although reported sensitivity and specificity of SDHB immunohistochemistry for detecting SDHx germline mutations range between 90% and 100% and 60% and 97%, respectively (van Nederveen et al. 2009, Evenepoel et al. 2015), our findings suggest a more variable degree of reliability. This variability could be related to somatic SDHx mutations (not routinely tested at our centre), unknown germline pathogenic variants, or to non-sensitive or non-specific immunohistochemical staining.

Apparently hereditary PPGL with negative genetic screening represent a challenge for the clinician. In our cohort, two out of three PHEO with metastatic disease had negative SDHB immunostaining and noradrenergic phenotype, but we could not identify a pathogenic SDHx gene mutation. VHL was not tested due to the lack of availability of this gene sequencing at the time of diagnosis; however, generally VHL mutations are not associated with negative SDHB staining (van Nederveen et al. 2009, Evenepoel et al. 2015). These two patients had no family history of PPGL but were under 45 years old. In these apparently hereditary cases, there are some possibilities. First, the disease is related to pathogenic SDHx mutations escaping detection by DNA sequencing (deleterious mutations in untranslated, intronic or promoter regions, low-level mosaicism in lymphocytes) or by epigenetic silencing of SDHx genes. Second, the spectrum of genetic mutations involved in PPGL development is incompletely known and continuously evolving so we cannot exclude a germline mutation which was not tested or discovered yet. For instance, the FH gene in which pathogenic mutations can also lead to the development of PPGL is not yet tested in all centres or patients. Third, it could be that these apparently hereditary PPGL are caused by a somatic mutation with an aggressive potential (e.g. ATRX, TERT mutations).

In conclusion, we found that patients with PGL more frequently have a hereditary predisposition than those with PHEO, contributing to the fact that diagnosis is generally made earlier in PGL. Although diagnosis in both PHEO and PGL was mostly due to related symptoms, patients with PHEO more often presented with CV comorbidities than those with PGL. The only minor differences between sporadic and hereditary PPGL and sometimes discordant genotype–phenotype correlations challenge the management of these diseases. Knowledge of clinical behaviour and evolution in tumours with somatic mutations is still in progress and more detailed tumour genotype–-phenotype research in sporadic PPGL should be done. Patients with apparently sporadic PPGL but with aggressive behaviour should be tested for somatic mutations. In addition, genetic diagnostic laboratories should extend the gene panel for PPGL with other newly discovered genes such as FH, EPAS, MDH2, SLC25A11, and others, at least until whole exome/genome sequencing is routinely available. Also, the inclusion of these genes should be accompanied by the development of clinical guidelines for follow-up and management of these patients in whom germline variants will be identified.

Declaration of interest

All authors have nothing to disclose.

Funding

This study did not receive any specific grant from any funding agency in the public, commercial or not-for-profit sector.