Von Hippel–Lindau and Hereditary Pheochromocytoma/Paraganglioma Syndromes: Clinical Features, Genetics, and Surveillance Recommendations in Childhood

Von Hippel–Lindau disease (vHL; OMIM #193300) is a multisystemic tumor predisposition syndrome characterized by benign and malignant tumors, including central nervous system (CNS) and retinal hemangioblastomas, clear cell renal cell carcinoma (RCC), pheochromocytoma (PHEO), pancreatic neuroendocrine tumors (pancreatic NET), endolymphatic sac tumor (ELST), and epididymal and broad ligament cystadenoma, as well as visceral (renal and pancreatic) cysts (1). Originally described in 1936 in the context of familial retinal angiomatosis by von Hippel (2) and CNS hemangioblastomas by Lindau (3), the incidence of vHL is estimated at approximately one in 36,000 (based on data preceding the advent of molecular genetic testing), and the lifetime penetrance approaches 100% by age 75 (4).

A clinical diagnosis of vHL can be established in one of two scenarios: (i) in an individual with a family history of vHL and the presence of a CNS or retinal hemangioblastoma, PHEO, or RCC; or (ii) in a simplex case (an individual with no family history) with ≥2 hemangioblastomas or ≥2 visceral tumors or one hemangioblastoma and one visceral tumor (5). Clinically, vHL is subdivided into five subtypes based on tumor spectrum and mutation type (Table 1). The clinical diagnosis is confirmed upon identification of a pathogenic variant in the VHL gene, the gold standard for vHL diagnosis (6).

Although many of the lesions associated with vHL present in the third and fourth decades of life, the age range of initial manifestations is wide (Table 2), and children are particularly vulnerable, as pediatric patients are at risk of developing hemangioblastomas and PHEO that can remain clinically occult until symptoms become severe. Moreover, when vHL-related tumors do occur, these malignancies can cause profound and lasting consequences (6–11).

Although most vHL-related tumors are histologically benign, morbidity due to mass effect can be significant (e.g., vision loss due to retinal hemangioblastoma; ref. 12). In addition, surgical intervention for tumors may be associated with operative/postoperative complications, such as hemorrhage from a large CNS hemangioblastoma (13). Historically, substantial mortality was attributable to RCC, pancreatic NET, and CNS hemangioblastoma (11). Fortunately, these risks have been mitigated in recent years by the institution of comprehensive surveillance paradigms, leading to early recognition of tumors and their multidisciplinary management (14–16).

VHL results from pathogenic variants in the VHL gene (6). vHL is inherited in an autosomal dominant pattern. Approximately 80% of individuals with vHL disease have an affected parent, and about 20% result from a de novo pathogenic variant. Genetic testing is indicated in first-degree relatives of individuals with pathogenic variants in VHL, as well as any child diagnosed with any of the following:

• Retinal angioma (hemangioblastoma)

• CNS hemangioblastoma

• Clear cell RCC

• PHEO or paraganglioma

• ELST

• Epididymal or adnexal papillary cystadenoma

• Multiple pancreatic cysts or pancreatic NET

• Multiple renal cysts

The type of variant identified in the VHL gene has been shown to account for differences in PHEO risk, with a strong genotype–phenotype correlation (Table 1; ref. 17). Truncating variants or exon deletions in the VHL gene are identified among individuals with vHL type I and confer a relatively low risk of PHEO (17, 18). In contrast, vHL type II is associated with missense variants that generally do not affect the protein structure and are associated with a relatively higher risk of PHEO (17).

The risk for RCC and hemangioblastoma in affected individuals may reflect the ability of the variant protein to regulate the hypoxia-inducible factor (HIF) pathway (19, 20). Higher HIF expression appears to result in lower risk of RCC and hemangioblastoma. In addition, specific missense variants in VHL are also associated with hereditary polycythemia. Most of these variants lead solely to the polycythemia phenotype; however, a few variants (L188V, G104V, V130I, and Y175C) have been associated with both polycythemia and PHEO (21, 22).

Early recognition and testing of at-risk individuals is key to the prevention of morbidity and mortality in vHL. Families with a history of vHL should be counseled regarding the importance of identifying at-risk children, performing genetic testing to identify mutation carriers, and initiating periodic surveillance. For example, children are susceptible to visual loss as result of retinal hemangiomas that, in the absence of surveillance, may go unnoticed until retinal damage becomes severe. CNS hemangioblastomas, if detected early, may be surgically excised with minimal damage to surrounding tissue.

Lifelong surveillance is, therefore, mandated given ongoing risks for tumor development and malignant growth with increasing age, and evidence that longitudinal surveillance in vHL limits morbidity and mortality (14–16). Multiple groups have developed surveillance guidelines (Table 3). Historically, these guidelines were based on expert consensus, with early screening targeted to tumor spectrum, risk levels, and typical ages of presentation (7, 23–25). Recent efforts have employed mathematical modeling methodologies, informed by the available clinical data, to predict the ideal ages to initiate surveillance and the most appropriate frequency for ongoing screening (26).

Our current consensus recommendations for screening of vHL-related tumors incorporate the elements of existing paradigms into a new surveillance regimen, with an emphasis on pediatric patients (Table 4). Broadly, the age-specific tumor risk, the youngest reported ages of occurrence, presumed growth rate, and the potential clinical impact of tumor progression were all considered in developing these new surveillance recommendations based on risk in children and adolescents.

Regardless of age, every individual with vHL should undergo an annual history and physical examination, including blood pressure assessment and a comprehensive neurologic evaluation, assessing for deficits including evidence of visual disturbance or hearing impairment. Ideally, these visits would be conducted by a medical provider experienced in caring for individuals with vHL, who has access to a multidisciplinary team with expertise in managing vHL-associated tumors. At these visits, education should also be provided on the signs and symptoms that could raise concern.

Beginning from birth, ophthalmology exams should be performed annually, with particular attention to the retina to monitor for retinal hemangioblastomas. This is consistent with most current screening paradigms. These exams should be conducted by an ophthalmologist experienced in pediatric retinal evaluation. In addition, starting at 2 years of age, PHEO surveillance should commence with blood pressure checks at every medical visit (using standard tables based on age and height; ref. 27), and with annual plasma or urine metanephrine levels. This screening interval is in agreement with standard practices, but the youngest age to initiate PHEO screening advocated in existing screening protocols is 5 years of age. There are multiple reports of PHEO occurring in children at younger ages, driving the impetus to advance the onset of screening (10, 26). We note that younger patients with PHEO bore missense variants (type II/III) rather than truncating or deletion variants (type I). This may inform future iterations of recommendations with respect to the age to initiate PHEO surveillance. In the course of surveillance, attention should be given to avoiding potentially interfering foods and medications that may confound interpretation of biochemical testing (Supplementary Table S1).

By 5 years of age, biennial audiology evaluations should be commenced to screen for ELST. This is in agreement with current screening paradigms. By 8 years of age, biennial MRI of the brain and spine should begin to monitor for CNS hemangioblastomas. Our recommendation to start this longitudinal surveillance at 8 years is significantly younger than dictated by other guidelines. Although the risk of CNS hemangioblastomas prior to adolescence is relatively low, multiple instances have been reported earlier in childhood (6, 23, 24). Moreover, these tumors may cause substantial morbidity with progression, particularly those lesions associated with peritumoral cysts that are more commonly found in younger individuals with vHL and that may progress more rapidly (28). As the risk of CNS hemangioblastomas rises in adolescence, and given the approximately 7% risk of hemangioblastomas developing in the interval between every-other-year MRIs, consideration may be given to increasing the screening frequency to annually starting in mid-adolescence, in contrast to the biennial surveillance recommended in established guidelines (4). The risk of rapidly growing hemangioblastomas, however, needs to be better quantified before this approach can be formally recommended. Cranial imaging should include thin cuts through the internal auditory canals, as a complement to audiology evaluations for ELST screening due to the possibility of ELST detection prior to development of audio-vestibular symptoms (29). Because spinal hemangioblastomas can occur at any segment of the spinal cord, it should be emphasized that spinal imaging should extend the entire length of the cord and not be restricted to only the cervical cord (30).

Our proposed surveillance for visceral manifestations of vHL diverges significantly from established recommendations. Primary screening for PHEO with imaging is not advocated given the high sensitivity of biochemical screening measures. Therefore, the characteristics of RCC (31) and pancreatic NET (6, 9, 23, 24) should drive the timing of screening initiation and the interval of ongoing surveillance. We recommend that surveillance for visceral manifestations of vHL should be implemented at 10 years of age, with annual MRI of the abdomen to maximize the sensitivity and maintain the consistency of the screening methodology for RCC and pancreatic NET detection. To ensure optimal detection for RCC, this abdominal MRI should be performed per institutional protocol used for renal evaluation. Ultrasound may be considered to complement the MRI, whereas CT is reserved for rare circumstances where biochemical abnormalities are detected and MRI is contraindicated. The latter places high value on reduction of exposure to ionizing radiation, in accordance with the position of the VHL Alliance (23).

Although longstanding evidence supports genotype–phenotype correlations in vHL, these are generally not considered in tailoring screening recommendations, as the correlations are still being characterized. This may change as data accrued from systematic surveillance become available. It should be recognized that the tumor surveillance in vHL is time-consuming and may incur substantial financial and psychosocial burdens (32), as discussed further in the genetic counseling article in this CCR Pediatric Oncology Series (33). However, these burdens may be diminished by experienced multidisciplinary teams through care coordination and enhanced education.

We propose a tumor surveillance paradigm for individuals with vHL based on specific risks in childhood and adolescence. Although these recommendations derive from existing paradigms, in considering screening onset and intervals, we placed high priority on the earliest ages of tumor onset, potential tumor growth rates, and the clinical impact of delayed detection of these tumors in children and adolescents. We have not made specific recommendations for adults with VHL, and practitioners should rely on the existing screening regimens for adults with VHL (Table 3).

Despite our reliance on available data, these guidelines remain largely based on expert opinion. The next logical step would incorporate a prospective assessment of clinical outcomes of individuals with vHL, screened according to these proposed guidelines. In addition, future advances in early detection methodologies may be realized through identification of reliable biomarkers for the aberrant vascular proliferation occurring in vHL. Finally, no measures are currently available to prevent individuals from developing vHL manifestations. Future investigations, however, could include the identification and application of strategies to inhibit aberrant vascular growth (34). The main impediments to all of these avenues of future study are the relatively low prevalence of vHL and the prolonged duration over which associated tumors may arise. These characteristics of the condition make accrual of sufficient numbers of affected patients, and/or samples, for these studies challenging and will depend on collaborative multi-institutional efforts.

Hereditary paraganglioma and PHEO syndromes (HPP) are characterized by rare and usually benign tumors of neural crest origin that are symmetrically distributed along the paravertebral axis from the base of the skull and neck to the pelvis. In addition to paraganglioma/PHEOs, patients with HPP syndromes can develop renal cancers, gastrointestinal stromal tumors (GIST), pituitary adenomas, and other rare tumor types. The genes in which pathogenic variants are known to cause HPP syndromes collectively include the SDHx genes, a group of multiple nuclear genes encoding subunits of the succinate dehydrogenase (SDH) enzyme complex. This enzyme complex catalyzes the conversion of succinate to fumarate in the Krebs cycle and serves as complex II of the electron transport chain. SDHx genes include SDH subunits A to D and SDH assembly factor 2 (SDHAF2), which is a stabilizing protein required for the flavination of SDHA (35). Other non-SDHx genes associated with hereditary paraganglioma/PHEOs include the MAX gene encoding a member of the helix-loop-helix leucine zipper family of transcription factors that regulates cell proliferation, TMEM127 that encodes a transmembrane protein that negatively regulates mTOR (36, 37). Recently HIF2α, EGLN1, and KIF1β have been implicated in the development of PHEO/paraganglioma (38, 39). It should also be borne in mind that PHEO/paraganglioma are components of other hereditary tumor predisposition syndromes associated with the RET, VHL, NF1, and FH genes, described elsewhere.

As with vHL, although the tumors associated with HPP syndrome are most frequently histologically benign, they can result in significant clinical morbidity related to mass effect, cranial nerve palsies, or hypertension/tachycardia resulting from catecholamine excess. If left untreated, a subset of these tumors will metastasize. A high malignant potential has been specifically recognized for SDHB-related tumors and is associated with tumor size at the time of diagnosis (40, 41). The increased metastatic potential and aggressive nature of SDHx-related paragangliomas compared with de novo paragangliomas without underlying germline predisposition must be taken into account when developing early tumor surveillance in patients with HPP.

Paragangliomas in the skull base, neck, and upper mediastinum are primarily associated with the parasympathetic nervous system and generally do not hypersecrete catecholamine or other hormones, although a subset can secrete dopamine. In contrast, paragangliomas in the lower mediastinum, abdomen, and pelvis are typically associated with the sympathetic nervous system and usually hypersecrete catecholamines. PHEOs are catecholamine-secreting paragangliomas confined to the adrenal medulla.

The most current studies suggest that up to 35% of PHEO/paraganglioma are hereditary (42). The diagnosis of HPP syndrome is based on molecular genetic testing, which should be offered to all patients with paraganglioma or PHEO. However, recognition of bilateral paraganglioma/PHEO and/or a strong autosomal dominant familial presentation of paraganglioma/PHEO may also lead to a clinical diagnosis of HPP. Assessment of the family history should include specific inquiry regarding relatives with sudden death, and it should detail the spectrum of tumors associated with HPP syndrome, some of which are known by different names (paragangliomas, e.g., may be called glomus tumors or extra-adrenal PHEOs). The medical history should include symptoms of catecholamine excess, such as elevation in blood pressure, headaches, diaphoresis, and palpitations, and also symptoms attributable to mass effect, such as dysphagia, hearing loss, and dysarthria. The exam should include assessment for arrhythmias and palpable masses (35).

GISTs occurring in the setting of pathogenic SDHx germline variants have distinct clinical features. The SDHx-related GISTs are almost always located in the stomach, arising from the interstitial cells of Cajal in a submucosal location. Multifocal gastric tumors are common. Because of their gastric location, by far the most common presentation for SDHx-related GISTs is gastric bleeding.

Pathogenic missense and truncating variants in SDH B, C, and D underlie the majority of patients with HPP syndromes (after excluding other syndromes such as vHL, multiple endocrine neoplasia and, less frequently, neurofibromatosis and hereditary leiomyomatosis and RCC). Variants in SDHA and SDHAF2, albeit rare, have also been reported. Although SDHx-related GISTs can be due to many of the different SDH genes (including SDHC epimutations), the SDHA germline mutations seem to be the most common germline mutations associated with GISTs (43). In the setting of tumor susceptibility, the SDHx genes act as tumor suppressors. The tumorigenesis mediated by mutated SHDx genes is thought to result from elevations in cellular succinate concentrations that stabilize the HIFα transcription factor by inhibiting prolyl hydroxylases, thus preventing HIFα degradation by ubiquitination (44). High levels of HIFα are strongly implicated in promoting tumor growth and metastasis through the role of HIFα in initiating angiogenesis and regulating cellular metabolism to overcome hypoxia (45). In addition, several reports suggest that SDHx-related tumors display a hypermethylator phenotype associated with downregulation of key genes involved in neuroendocrine differentiation (46, 47). TMEM127 and MAX also function as tumor suppressors, and loss of heterozygosity has been demonstrated in PHEO/paraganglioma tumor tissue.

Susceptibility to paraganglioma/PHEO, as well as other associated tumors, is either inherited in a strict autosomal dominant fashion (SDHA, SDHB, SDHC, and TMEM127) or in an autosomal dominant fashion modified by parent of origin. SDHD, SDHAF2, and MAX variants present with this parent-of-origin–dependent tumorigenesis, wherein tumor formation occurs almost exclusively in the context of paternal transmission of the variant due to maternal imprinting (i.e., affected children of carrier fathers will present with disease, but affected children of carrier mothers will not).

Penetrance estimates have been defined for variants in SDHD and SDHB. For SDHD variants, the penetrance is about 90% for probands and relatives identified through variant testing (48). Penetrance estimates for SDHB variant carriers have been more variable due to ascertainment and analytic differences between studies. Early studies estimated an approximately 77% risk for paraganglioma by age 70 (49); however, more recent estimates based on testing of extended families have estimated the lifetime risk to be between 30% and 50% (50) and may be even as low as 14% by age 60 (51). Variants in SDHA, SDHAF2, and SDHC are rare, and specific penetrance estimates have not been calculated. A single-center review of eight probands identified with SDHC variants noted that none of them had a family history of paraganglioma, suggesting that the penetrance is likely incomplete (52). SDHAF2 variants have been described in a large Dutch family. In this kindred, 11 of 16 individuals (69%) who had paternally inherited the variant were found to have paraganglioma (50, 53). Cumulative penetrance of TMEM127-associated PHEOs has been estimated to be 32% from a large family with 34 members (54).

Patients with SDHB variants have a positive family history in 33% of cases, present with single tumors at a mean age of 25 to 30 years (range, 6–77), and are strongly associated with extra-adrenal sympathetic paragangliomas, mainly in the abdomen and pelvis. About 20% may also have PHEOs, and, as discussed, paragangliomas in these patients have a substantial propensity to metastasize (25, 55–59). Paraganglioma development in SDHB variant carriers is generally associated with higher morbidity and mortality than pathogenic variants in the other SDHx genes due to the greater risk for malignancy and metastasis. In comparison with individuals with sporadic, malignant paraganglioma, those with malignant disease and a germline SDHB variant have shorter survival (55). Although less common than malignant extra-adrenal sympathetic paragangliomas, malignant PHEOs do occur and may be more common among individuals with a germline variant in SDHB compared with other PHEO-predisposing loci.

Patients with SDHD, SDHC, and SDHAF2 variants are more frequently associated with parasympathetic skull-base and neck paragangliomas (49, 52, 53). Individuals with SDHD variants present at a mean age of 28 to 31 years (range, 12–70) with multiple tumors, whereas those with SDHC variants most commonly develop a solitary tumor (about 77%), with a mean age of diagnosis of 38 years of age (range, 15–40; refs. 25, 57, 58). Eighty percent of reported SDHC-related paragangliomas originate in the head and neck, and malignant disease has been rarely reported (2%; ref. 52). Only head and neck tumors, predominantly carotid body tumors (70%), have been reported in SDHAF2 variant carriers, with an average age of onset of 33 years (range, 22–47). However, all descriptions of the SDHAF2 phenotype are based on a single family (53, 60). Germline SDHA variants have been observed in association with both PHEOs and paragangliomas (sympathetic and parasympathetic; ref. 60).

Tumorigenesis is rare in the first decade of life among individuals predisposed to HPP. The youngest patient that we have identified with paraganglioma was 6 years at diagnosis (59). After reviewing the literature and clinical practice across many different institutions and countries, we recommend initiating tumor surveillance at 6 to 8 years of age. Although we acknowledge specific genetic lesions drive varying clinical presentations, we do not think that the genotype–phenotype relationships are currently well enough defined to justify distinct surveillance paradigms based on genotype. Therefore, we advocate a single regimen for all carriers of variants associated with HPP syndromes, with a need to accumulate enough data over time to make more gene-specific recommendations.

Surveillance recommendations for HPP syndromes resemble those for vHL with respect to PHEO/paraganglioma screening (Table 5). However, we recognize the existence of silent (nonsecretory) paragangliomas, particularly those occurring in the head and neck, which occur with greater frequency than in vHL and, thus, emphasize the importance of radiologic surveillance in HPP syndrome in addition to biochemical screening. This screening will also facilitate identification of RCCs, which have rarely been associated with germline SDHx variants. Although limited data exist, whole-body imaging was recently studied among individuals with SDH-associated paraganglioma syndromes and identified six tumors among asymptomatic carriers. Sensitivity and specificity of whole-body imaging were 87.5% and 94.7%, respectively, whereas biochemical surveillance yielded sensitivity and specificity of 37.5% and 94.5%, respectively (61). Although there is increasing interest in the use of contrast-enhanced ultrasound (CEUS; ref. 62), there are insufficient data to recommend CEUS over MRI for routine surveillance. As some paragangliomas are predominantly dopaminergic (particularly those associated with SDHB, C, D), plasma methoxytyramine analysis is advocated in addition to metanephrines (63–65). Finally, chromogranin A, a NET marker, has been demonstrated to enhance detection of PHEO/paragangliomas by 22%, with minimal impact on specificity, and can be included in annual surveillance (66).

There are currently no data to support routine imaging for GISTs in at-risk patients. Formal studies of the sensitivity of whole-body MRI for submucosal gastric masses have not been performed, but one would anticipate that the sensitivity will be relatively low for small GISTs. Imaging approaches more likely to detect submucosal gastric GISTs include CT with oral and intravenous contrast, with or without FDG-PET. Both of these imaging modalities carry the risk of exposure to radiation. Given the relatively low incidence of GIST in HPP, the benefits are unlikely to outweigh the risks of exposure to radiation and follow-up of false positives. On the basis of this information, we do not advocate imaging surveillance specifically for GIST detection in patients with HPP syndrome. Because essentially all GISTs in HPP patients are gastric (in some series, 95% of patients present with gastric bleeding and/or anemia, and in some cases, GISTs can present with rapid bleeding and significant anemia; refs. 67–69), based solely on a consensus of opinion, we recommend an annual complete blood count to screen for incident anemia during surveillance. The effectiveness of this approach and the frequency of incidental findings requiring further evaluation with hematology referral or endoscopy are not known. There are insufficient data to define the typical age of onset of GIST in this population, so we propose that the age of onset and the interval for surveillance is the same as for PHEO/paraganglioma (at least until such time as future evidence dictates otherwise).

The identification of the genetic basis of the HPP syndromes has led to significant advances in clinical care for these patients, providing prognostic insights as well as opportunities for early detection and treatment of component tumors. Further research is needed to clarify how individual genotypes can more reliably predict phenotypic behavior, and whether and how surveillance should differ based on genotype. In addition, there remains an urgent and unmet clinical need to develop improved therapies for patients with metastatic PHEO/paragangliomas.

W. K. Kohlmann reports receiving commercial research grants from Myriad. No potential conflicts of interest were disclosed by the other authors.

G.E. Tomlinson was supported by the NCI under award number 5P30CA054174-21.