Current Management of Pheochromocytoma/Paraganglioma: A Guide for the Practicing Clinician in the Era of Precision Medicine

Pheochromocytomas and paragangliomas (PCC/PGLs) are rare, mostly catecholamine-producing neuroendocrine tumors of the adrenal gland (PCCs) or the extra-adrenal paraganglia (PGL). They can be separated into three different molecular clusters depending on their underlying gene mutations in any of the at least 20 known susceptibility genes: The pseudohypoxia-associated cluster 1, the kinase signaling-associated cluster 2, and the Wnt signaling-associated cluster 3. In addition to tumor size, location (adrenal vs. extra-adrenal), multiplicity, age of first diagnosis, and presence of metastatic disease (including tumor burden), other decisive factors for best clinical management of PCC/PGL include the underlying germline mutation. The above factors can impact the choice of different biomarkers and imaging modalities for PCC/PGL diagnosis, as well as screening for other neoplasms, staging, follow-up, and therapy options. This review provides a guide for practicing clinicians summarizing current management of PCC/PGL according to tumor size, location, age of first diagnosis, presence of metastases, and especially underlying mutations in the era of precision medicine.


Introduction
Pheochromocytomas/paragangliomas (PCCs/PGLs) are rare neuroendocrine tumors, mostly catecholamine-producing and originating from chromaffin tissue derived from the neural crest. The incidence of diagnosed PCC/PGL is about 0.8/100,000 patients/year [1] with around 30-40%  Percentage of germline mutations [2] Germline 100% Germline 25% Germline 20% Germline 0% This further justifies subjecting all PCC/PGL patients to genetic testing for germline mutations [39]. Somatic tumor mutations may indeed also influence prognosis, since different somatic mutations correlate with different metastatic risks in a similar manner as germline mutations do (Table 1), although, of course, this is still the subject of ongoing research. Moreover, a personalized molecular-targeted therapy approach could result from somatic mutations such as RET, NF1, HRAS, MAX, TMEM127, or those of defective pseudohypoxia signaling, much as it does from the molecular testing in other cancer entities such as differentiated thyroid carcinoma, bronchial carcinoma, or breast cancer [40,41].

Conclusion/Practical Tips
We recommend genetic testing for germline mutations for all PCC/PGL patients; testing of tumor material, where available, for somatic tumor mutation analysis can also provide useful additional information.

Biochemistry
Measurements of plasma free or urinary metanephrines are the recommended screening tests for the biochemical diagnosis of PCC/PGL [42], with recent evidence now establishing the superiority of the plasma over urinary tests [6]. However, high accuracy of the plasma test can only be assured using appropriately established measurement methods and reference intervals, combined with correctly applied preanalytics, especially sampling of blood in the supine position [6,[43][44][45]. Under these conditions, plasma concentrations of normetanephrine, metanephrine, or methoxytyramine more than two-fold above upper cut-offs of reference intervals indicate a high probability of PCC/PGL even at low pre-test prevalence of disease [8,46]. Combined increases of two or more metabolites also suggest a high probability of PCC/PGL. In other cases, intra-patient longitudinal comparisons can be useful to confirm progression of disease (i.e. increased disease burden), which in most cases is slow and involves a doubling-time of over two years [8]. Biochemical testing for PCC/PGL in patients screened due to signs and symptoms of apparent catecholamine excess should always be performed before imaging [42]. Moreover, it is important to instruct the patients to abstain from caffeine, black tea, nicotine, alcohol, bananas, cheese, almonds, nuts, chocolate, eggs, or vanilla three days prior to assessment of plasma free or urine metanephrines. Tricyclic antidepressants, serotonin re-uptake inhibitors, ephedrine, cocaine, and metamphetamine may lead to false positive results.
As shown in Table 1, pseudohypoxic cluster 1 type PCC/PGL are linked to a noradrenergic (predominantly increased normetanephrine) phenotype because the tumors lack the enzyme, phenylethanolamine N-methyl transferase (PNMT), which converts noradrenalin to adrenalin [47,48]. Moreover, a substantial proportion of PCCs/PGLs in SDHx mutation carriers also appear to show some additional deficiency in the enzyme dopamine-β-hydroxylase, which catalyzes the final step in catecholamine synthesis-the conversion of dopamine to noradrenalin; patients with such tumors show increased production of the dopamine metabolite 3-methoxytyramine [8,49,50]. Thus, all cluster 1 PCCs/PGLs may be diagnosed by elevated normetanephrine levels, with or without increases in 3-methoxytyramine. However, measurements of urinary 3-methoxytyramine are not useful and only plasma 3-methoxytyramine is of significant diagnostic utility [6]. Moreover, SDHx mutation carriers often show reduced synthesis and secretion of normetanephrine and in rare cases lack the initial rate-limiting enzyme of catecholamine synthesis (tyrosine hydroxylase); such cases may be termed non-functional or biochemically silent [9]. SDHx mutation carriers can therefore be diagnosed preferentially by normetanephrine and 3-methoxytyramine while, in cases of silent phenotypes, circulating chromogranin A may be a useful marker [51][52][53]. It is important to note that there are multiple other diseases, conditions, and medications leading to false positive chromogranin A results. Diseases potentially leading to elevated chromogranin A levels are other endocrine diseases/tumors (pulmonary/gastrointestinal neuroendocrine tumors, medullary thyroid carcinomas, pituitary tumors, hyperthyroidism, hyperparathyroidism, small cell lung cancer, prostate cancer, breast cancer, ovary carcinoma), systemic inflammatory diseases (systemic rheumatoid arthritis, chronic bronchitis, chronic airway obstruction in smokers), renal insufficiency, gastrointestinal disorders (chronic atrophic gastritis, pancreatitis, inflammatory bowel disease, irritable bowel syndrome, liver cirrhosis, chronic hepatitis, hepatocellular carcinoma, pancreatic and colon cancer), cardiovascular diseases (arterial hypertension, cardiac insufficiency, acute coronary syndrome), drugs (proton pump inhibitors, H2 blockers, steroids), and pregnancy. Therefore, it is especially important to instruct the patient to withhold proton pump inhibitors for at least one week, optimally for two to three weeks, prior to assessment of chromogranin A. H2 blockers should be withheld at least two days prior to chromogranin A assessment if possible.
Since biochemical testing may not be applicable in patients with non-functioning PCC/PGL, clinicians have to rely on repeat imaging for surveillance.

Conclusion/Practical Tips
Especially in the case of a history/suspicion of PGLs, metastatic disease or in SDHx mutation carriers, additional assessments of plasma 3-methoxytyramine (but not urinary 3-methoxytyramine) are useful [17] since highly elevated 3-methoxytyramine levels may suggest the presence of metastatic tumors [9,17]. Such patients may be triaged for pre-operative staging (if possible by radionuclide imaging, see below) [39].
In the case of a suspected or diagnosed non-functional PGL/PCC, chromogranin A should be measured (no proton pump inhibitor at least one week prior to assessment).
A recent meta-analysis of pooled PCC/PGL detection by radionuclide imaging showed the highest sensitivity (93%) for 68 [74].

Conclusion/Practical Tips
For PCC screening in the case of elevated metanephrines, CT (or MRI in young patients) should usually be sufficient.
However, in all patients with PCC/PGL (except for patients with low risk of metastatic disease, i.e., epinephrine secreting PCC <3-5 cm), whole-body CT or MRI or radionuclide imaging is recommended prior to surgery in order to rule out metastatic disease/multiplicity which may affect the decision regarding surgery [39].

Biopsy Is Not Recommended
There is no recommendation for performing biopsy in PCCs/PGLs. The European guideline for adrenal incidentalomas [77] claims that biopsy of an adrenal tumor should only be performed if: There is another extra-adrenal malignancy in the patient's history; The tumor is non-functioning (especially non-functioning PCCs/PGLs); Not judged as benign on imaging; Biopsy would change patient management.
Therefore, biopsy should only be considered in special cases of non-functioning potential PCCs/PGLs when all biochemistry is negative and it would change patient management, for example, due to another extra-adrenal malignancy in the patient's history. Otherwise, the adrenal mass should generally be surgically resected due to the risk of a catastrophic complication related to a biopsy of a functioning PCC/PGL [78].

Conclusion/Practical Tips
We strongly discourage image-guided biopsy of any adrenal or suspicious retroperitoneal mass without pre-biopsy biochemical case detection testing.

Immunohistochemistry: Biomarkers
Besides typical morphology, well-known biomarkers for PCC/PGLs are positive immunohistochemistry for chromogranin A, synaptophysin, and S100; however, this does not allow differentiation from any other neuroendocrine tumors. This in turn may be a problem if the patient does not show a typical clinical presentation or biochemistry for a PCC/PGL. In these cases, negativity for keratin and site-specific transcription factors for neuroendocrine tumors as well as positivity for tyrosine hydroxylase (except for non-functioning HN PGLs) and GATA-3 immunohistochemistry seem to be valuable biomarkers to confirm the diagnosis of a PCC/PGL in most cases, as recently reviewed [20]. Future potential relevant biomarkers suggesting metastatic potential may be TERT structural variants [79], BUP1 [80], chromogranin B [81], ERBB2 [82], and/or the well-known SDHB immunohistochemistry [83].

Follow-Up
It should be emphasized that all patients with a history of a PCC/PGL are at risk of recurrence-even after complete (R0) resection-and any PCCs/PGL may have metastatic potential [19,39]. This is especially important for those patients with large (≥5-6 cm) tumors [64]. Moreover, extra-adrenal location, a noradrenergic or dopaminergic biochemical phenotype, high chromogranin A levels, young age <20 years, multiplicity, and, most importantly, pseudohypoxic cluster 1-related germline mutations (especially SDHA/B), are associated with a higher metastatic risk and an adverse prognosis once metastasis is found [16,17,[84][85][86][87][88].
Furthermore, pathologists cannot safely determine from histological findings if a PCC/PGL is "malignant" or has "metastatic potential" since the existing grading systems [26,27] are of limited predictive power and well-validated biomarkers for metastatic diseases are missing [19,28,29,39]. Thus, for all patients with a history of a PCC, follow-up for at least 10 years is suggested [39,42]; for high-risk PCC patients (germline mutation, young age <20 years, large tumor ≥5-6 cm, for SDHB carriers tumor size ≥3-3.5 cm) and all PGL patients lifelong surveillance is recommended-at least consisting of a yearly clinical investigation and assessment of urinary or plasma metanephrines and 3-methoxytyramine (see above) [39,42,88]. In cases of biochemically-silent PCCs/PGLs, additional imaging every one-two years is suggested [39]. In order to minimize radiation exposure, MRI is the preferred imaging modality for follow-up but it can miss tumors in unusual locations.
All patients at risk for a new PCC/PGL due to a germline mutation should also be offered lifelong surveillance [39,88,89]. However, currently, the details of such lifelong surveillance are unclear-especially if and how often imaging should be performed. Accordingly, due to the strong inter-patient heterogeneity depending on the underlying germline mutations and disease characteristics, Crona et al. suggest a personalized surveillance program using genetic mutations together with disease characteristics (Table 3) [32, 42,89]. Table 3. Follow-up depending on the underlying germline mutation and disease characteristics (after presumably curative surgery).

Conclusion/Practical Tips
In patients with elevated metanephrines prior to PCC/PGL surgery, plasma/urine metanephrines and 3-methoxytyramine (plasma only) should be assessed on pain-free recovery three-six weeks after surgery [39]; in patients with elevated chromogranin A prior to PCC/PGL surgery, chromogranin A should be assessed three-six weeks after surgery [39]; in the rare patients with non-functional PCC/PGL (no tumoral synthesis of catecholamines) or with postoperatively elevated metaneprines or 3-methoxytyramine, imaging should be performed three-four months after presumed complete surgery (if possible with the recommended radionuclide imaging, see above) [39].
We suggest especially radionuclide imaging three-four months after presumed complete resection of a metastatic/multiple PCC, any PGL, or high-risk (SDHA/B) mutation carrier in the case of post-operative abnormal biochemistry or non-functional tumor (Table 3).
At least 10 years follow-up is recommended for all patients with a history of PCC, but if the tumor was initially ≥5-6 cm, lifelong follow-up is recommended.
Lifelong follow-up is recommended for all high-risk patients (every patient with a germline mutation, PGL, young age <20 years, large tumor size ≥5-6 cm, for SDHB carriers tumor size ≥3-3.5 cm, multiplicity/recurrence, noradrenergic/dopaminergic phenotype at the initial presentation, moderately to poorly differentiated PCC according to the GAPP classification system).
A personalized surveillance program is suggested depending on the underlying germline mutation and disease characteristics (Table 3).
For patients with high-risk mutations (especially SDHA/B), we suggest clinical/biochemical control every six months and MRI every one-two years (consider CT for suspected lung involvement or use an alternate approach using CT and MRI).
For completely resected metastatic PGL/PCC, we suggest clinical/biochemical control every six months, MRI six months and 12 months after surgery, then annually. CT can be also used, but with more caution and less frequently since it possesses a radiation risk. An alternate approach using CT and MRI is also an option. Additional radionuclide imaging may be considered every two-three years, especially in the case of high-risk mutations (SDHA/B). Other risk factors may apply as well, but their validation may be needed.
For staging purposes of metastatic disease, we suggest whole body cross-sectional CT or MRI every four-six months and radionuclide imaging every one-two years depending on whether PRRT is considered as a treatment option; a patient's age and the growth rate/grading of the tumor are also important factors to be considered.

Perspectives
For all above recommendations on follow-up, it should be considered, although it appears logical, that earlier diagnosis of PCC/PGL by follow-up surveillance should provide for better outcomes. This has not been established by any evidence-based study. One recent study provides suggestive evidence for improved outcomes with surveillance in mutation carriers [95], but the retrospective nature of that study and associated uneven matching of populations were confounders that do not allow concrete conclusions about the benefits of follow-up. The international multicenter prospective cohort study PROSPHEO including patients with a history of PCC/PGL, newly-diagnosed PCC/PGL, or mutations in PCC/PGL susceptibility genes from different centers in Germany and Switzerland over 18 years may eventually be able to answer questions concerning benefits of follow-up surveillance as well as the optimal follow-up procedures.

Therapy
Surgery is always the therapy of choice of non-metastatic PCC/PGL, whenever possible [96,97]. However, surgery of non-functioning HN PGLs has to be carefully balanced against surgery-related morbidities, especially for the cranial nerves for vagal and jugular PGLs [98,99]. In cases of a high risk related to surgery, radiotherapy/radiosurgery (gamma-knife/cyberknife) might be a less invasive option with non-curative but controlling outcomes [100][101][102].
In patients with hereditary PCCs, cortical sparing surgery should always be considered since there is frequently a high risk of bilateral PCCs in hereditary disease, and cortical-sparing surgery was not associated with decreased survival despite PCC recurrence in 13% of cases in a very recently published study [103].
With metastatic disease, primary tumor resection should be recommended if feasible in order to alleviate cardiovascular and other symptoms from catecholamine excess or from tumor invasion, and to minimize the target for radiopharmaceutical therapies [31, 104,105]. Moreover, several studies have shown that surgical resection of the primary tumor is associated with improved survival even with metastatic disease [104,106,107]. In addition, complete metastatic surgery may be considered in oligo-metastatic PCC/PGL on a case-by-case decision, although there is only little evidence for such an approach from single case reports [108,109].
Watchful waiting with frequent follow-up may be the optimal initial approach in patients with non-functioning HN PGL, especially without evidence of significant tumor growth and/or compression of surrounding structures.
Conventional external beam radiation therapy (cEBRT) or radiosurgery (gamma-knife/cyberknife) are well-established methods in the case of bone metastases and also may play a significant palliative role in oligo-metastatic scenarios [102,110].
Minimally-invasive procedures such as radiofrequency ablation, cryoablation, and ethanol injection may be considered in the treatment of metastatic PCC/PGL, especially in oligo-metastatic disease [111,112].
Bisphosphonate or denosumab therapy should be considered in the case of bone metastases by analogy with other types of neuroendocrine tumors.
Adequate blood pressure control with alpha adrenoceptor blockade at least 10-14 days prior to surgery is essential in functioning PCCs/PGLs to prevent severe cardiovascular events during surgery [12,113,114]. In palliative scenarios, alpha adrenoceptor blockade should also be considered-balanced against side effects-to alleviate hormonal symptoms and prevent complications from catecholamine excess [42]. It has been conventional to use phenoxybenzamine at starting doses approximating 10 mg 2-3x per day, although other similar drugs such as doxazosin and prazosin have been used. There is no clear evidence for the superiority of one alpha-blocker for the pre-operative blockade of PCC/PGL patients, as previously reviewed [115]. Nevertheless, perioperative hypertension seems to be slightly better controlled with phenoxybenzamine (especially in those patients with high catecholamine or metanephrine levels), although with more pronounced postoperative hypotension. Indeed, there were fewer side effects in the doxazosin group [115]. Moreover, in functioning metastatic PCCs/PGLs, pre-treatment alpha blockade is recommended prior to initiation of therapy to prevent symptomatic catecholamine release in response to locoregional or systemic treatment. Furthermore, it is important to mention that beta adrenoceptor blockers must not be given prior to initiation of an adequate alpha adrenoceptor blockade [114].

Targeted Endoradionuclide Therapy Using [ 131 I]MIBG or Radiolabeled Somatostatin Analogs (PRRT)
The best investigated and well established therapy for metastatic PGL/PCC is [ 131 I]MIBG therapy, which is preferentially recommended for slow-growing [ 123/131 I]MIBG positive metastatic PCCs/PGLs [32,34], although the studies are very heterogeneous and most are retrospective [118]. In a meta-analysis of 17 studies with 243 PCC/PGL patients on [ 131 I]MIBG therapy, there was a complete response in 3% of patients, a partial response in 27% of patients and stable disease in 52% of patients: 40% of patients showed a partial hormonal response [118]. In two of these studies, the mean progression-free survival (PFS) of [ 131 I]MIBG treated patients was 23.1 and 28.5 months, respectively [118]. However, due to the unknown status regarding tumor progression prior to therapy in most of

Targeted Endoradionuclide Therapy Using [ 131 I]MIBG or Radiolabeled Somatostatin Analogs (PRRT)
The best investigated and well established therapy for metastatic PGL/PCC is [ 131 I]MIBG therapy, which is preferentially recommended for slow-growing [ 123/131 I]MIBG positive metastatic PCCs/PGLs [32,34], although the studies are very heterogeneous and most are retrospective [118]. In a meta-analysis of 17 studies with 243 PCC/PGL patients on [ 131 I]MIBG therapy, there was a complete response in 3% of patients, a partial response in 27% of patients and stable disease in 52% of patients: 40% of patients showed a partial hormonal response [118]. In two of these studies, the mean progression-free survival (PFS) of [ 131 I]MIBG treated patients was 23.1 and 28.5 months, respectively [118]. However, due to the unknown status regarding tumor progression prior to therapy in most of these studies, no conclusion can be drawn from the rated stabilization observed in these studies [118]. One study included patients only after progressive disease, but PFS and overall survival (OS) were not reported in this study [119]. Objective responses were mainly observed in patients with soft tissue metastases with a prolonged PFS, but there was no demonstrated impact on survival [118].
The most common side effects are anorexia, nausea, vomiting, and hematologic toxicity with grade 3-4 neutropenia in 87% of patients and grade 3-4 thrombocytopenia in 83% [118]. However, in long-term survivors there is a risk of myelodysplastic disorders [120]. One problem with conventional [ 131 I]MIBG therapy is the relatively low specific activity of the radiopharmaceutical (most of the MIBG molecules are not 131 I-labeled) potentially leading to less uptake into the tumor and life-threatening side-effects. Very recently, high-specific activity (HSA) [ 131 I]MIBG that consists almost entirely of 131 I-labeled molecules has been developed [117]. A recent multicentric phase II study led to the showed partial responses (23%, 15/64) or stable disease (44/64, 69%) as the best objective response. The median OS was 36.7 months (95% confidence interval, 29.9-49.1 months) with 18 months for patients who received one therapeutic dose and 44 months for those who received two therapeutic doses [117]. The most common treatment-related side effects were nausea, myelosuppression, and fatigue [117], with a higher rate of hematologic toxicity compared to conventional [ 131 I]MIBG-therapy.
Since most PCCs/PGLs strongly express somatostatin receptor subtype 2 (SSTR2) [121,122], PRRT using radiolabeled somatostatin analogs has been studied in several small studies [123][124][125][126][127][128][129][130][131][132] and recently reviewed [116]. Therefore, latterly, PRRT-a treatment option that has already been approved for gastro-enteropancreatic and pulmonary neuroendocrine tumors in many countries [133]-has also been suggested as an effective treatment option for metastatic PCCs/PGLs [116]. PRRT planning is always preceded by somatostatin receptor imaging as part of a theranostic approach. Although most studies consist of small numbers and limited follow-up, in one direct comparison study the percentage of patients with tumors showing disease stabilization was significantly greater after PRRT using [ 177 [127]. Further clinically controlled studies (e.g., NCT04029428, NCT03923257, NCT03206060) should be awaited with regard to recommendations and guidelines, including the selection of the radionuclide ( 90 Y, 177 Lu; another potential future option is the use of α-emitting radionuclides [134]), doses and dose regimens, but also the consideration of specific risk constellations (nephrotoxicity, patient age, co-medication with SSA).
Most common treatment-related side effects of PRRT which is already well-studied in neuroendocrine tumors [133] are renal or hematological (bone marrow) toxicities that can be minimized by adequate precautions and proper and safe dosing [135][136][137][138].
In order to maximize radionuclide uptake in the tumor with minimal risk to organs at risk (especially the kidney), an internal patient-specific dosimetry prior to PRRT is a potential future approach (personalized PRRT) [139][140][141][142]. However, currently, a fixed empiric dose is applied in analogy to the NETTER-1 trial in most centers [133] and dosimetry is not a standard procedure.
Interestingly, there are few or no data on the use of unlabeled long-acting somatostatin analogs, such as octreotide LAR or lanreotide autogel, in therapy. These are administered once a month, and in patients with other forms of neuroendocrine tumors (NETs), especially pancreatic and midgut NETs, they cannot only inhibit hormonal secretion but, according to two large-scale trials, attenuate tumor progression [143,144]. As such agents are usually well-tolerated with generally mild adverse effects, they may be trialed in patients with the aim of lowering catecholamine secretion and possibly stabilizing tumor growth, especially when 68  One head-to-head comparison study indicates that somatostatin receptor targeted PRRT may be superior to conventional [ 131 I]MIBG-therapy regarding treatment response, PFS and OS in metastatic PCC/PGL, especially in the subgroup of PGLs, and may be recommended for slow-growing metastatic PCC/PGL.

Chemotherapy
The second most studied therapy recommended for rapidly progressing metastatic PCC/PGL is conventional chemotherapy with cyclophosphamide, vincristine, and dacarbazine (CVD, Averbuch scheme: Cyclophosphamide 750 mg/m 2 , vincristine 1.4 mg/m 2 , and dacarbazine 600 mg/m 2 on day 1 and dacarbazine 600 mg/m 2 on day 2, at 21 day intervals) [145,146]. In a meta-analysis of the largest studies, CVD led to a complete response in 4%, a partial response in 37%, and stable disease in 14% of patients [145]. PFS was only reported in two of these studies with 20 and 40 months, respectively. However, due to missing information regarding tumor progression prior to therapy, no valid conclusions can be drawn from these PFS data [145]. There is only one study solely including patients with progression prior to therapy which showed a radiological and clinical response in 33% of patients [105]. This is also the only study showing a significant survival benefit for patients who responded to CVD chemotherapy: There was a significant effect of response to chemotherapy on median OS (according to a multivariate Cox proportional hazard model analyses) [105]. The median OS of radiological responders was 6.4 years versus 3.7 years for non-responders [105]. The CVD regimen has been shown to be especially effective in SDHB-related PCC/PGLs [145,147,148].
Prolonged CVD chemotherapy (median of 20.5 cycles) in 12 patients with SDHB mutations led to a total response in 83% of patients [partial response 8/12 (66.7%) patients, complete response 2/12 (16.7%) patients, assessed by Response Evaluation Criteria in Solid Tumors (RECIST)] and a PFS/OS of 930 and 1190 days, respectively [148]. On a case-by-case basis, prolonged CVD therapy can be considered, especially for metastatic SDHB-related PCC/PGLs.
Monotherapy with the DNA-alkylating chemotherapeutic temozolomide, an oral metabolite of dacarbazine, showed a partial response (33%) or stable disease (47%) in a total of 80% of patients with SDHB mutations, and thus may be used as a single agent treatment, or alternatively could be considered as a maintenance regime for tumor stabilization subsequent to six-nine cycles of CVD chemotherapy (150 mg/m 2 on days 1-5, at 28 day intervals) [149,150]. Down-regulation of the DNA repairing enzyme O-6-methylguanine-DNA methyltransferase (MGMT) via hypermethylation in SDHB mutated tumors appears to lead to increased susceptibility of SDHB-related PCCs/PGLs to temozolomide [149,[151][152][153].
In the case of intolerance to temozolomide monotherapy, a combination of a metronomic scheme with long-term low-dose temozolomide (75 mg/m 2 per day for three weeks followed by one week off treatment) and high-dose lanreotide autogel (120 mg s.c. every 14 days) may be an alternative in order to stabilize PCC/PGL growth (low MGMT levels seem to be beneficial), as reported for two patients [150].
Whether adjuvant chemotherapy with four-six cycles of CVD after surgery is beneficial, has not as yet been studied. There are no data providing evidence for adjuvant chemotherapy.
On a case-by-case decision, prolonged treatment with CVD chemotherapy with 20 cycles of CVD is suggested, especially in the case of patients with SDHB mutations.
Alternatively, temozolomide monotherapy (150 mg/m 2 on days 1-5, at 28 day intervals) or a metronomic scheme with temozolomide (75 mg/m 2 per day for three weeks followed by one week off treatment) may be considered, either as initial therapy or following stabilization with CVD.

Targeted Therapy and Immunotherapy
Different receptor tyrosine kinase inhibitors (TKIs) (sunitinib, cabozantinib, axitinib, lenvatinib, and pazopanib) are currently under evaluation as treatment options for metastatic PCCs/PGLs. They all have anti-angiogenic effects and may be interesting therapy options for cluster 1 and cluster 2 mutated PCCs/PGLs.
The best studied TKI is sunitinib, which is already approved by the FDA and by the European Medicines Agency (EMA) for pancreatic neuroendocrine tumors, renal cell cancer, and gastrointestinal stromal tumors. Sunitinib leads to inhibition of VEGF1/2 receptors, platelet-derived growth factor-β receptor (PDGFR) and RET. The largest retrospective study included 17 patients of which 14 were evaluable for tumor responses to sunitinib (dose: 37.5 mg or 50 mg) [155]. Of the 14 patients, a total of 8/14 (57%) of the patients showed a partial response (3/14, 21%) or stable disease (5/14, 36%) [155]. However, the median PFS was only 4.1 months, although there was a much longer PFS in the responders compared to the non-responders. The PFS of the three partial responders was 11, 12, and 4.5 months, respectively [155]. The PFS of three patients with stable disease was 27, 8, and 6 months, respectively, and two other patients with stable disease experienced no progression until the end of the observation period (36 months). One of these patients remained on targeted combination therapy with sunitinib and the mTORC1 inhibitor rapamycin for 1.5 years until the end of the observation period [155]. In the non-responders, the PFS was 0.4-4 months. A total of 6/8 (75%) of the patients with stable disease or partial responses in this study were SDHB mutation carriers, indicating clinical benefit especially for SDHB-related PCCs/PGLs [155].
Importantly, the largest prospective sunitinib phase II multicenter study has recently been published (regime: 50 mg sunitinib daily for four weeks, followed by two weeks off treatment corresponding to one cycle): The total disease control rate (stable disease or partial response) was 83% (95% CI: 61-95%): 3/23 (13%) patients showed a partial response. All responders were carriers of germline mutations (SDHA, SDHB, RET). The median PFS was 13.4 (95% CI: 5.3-24.6) months [156]. The RET-mutated patient with MEN2A was still on sunitinib therapy by the time the study was published (after 64 cycles) which may implicate that RET and SDHx mutation carriers as benefiting the most [156].
At present, sunitinib is being investigated in the first randomized placebo-controlled phase II clinical trial in advanced PCC/PGL (FIRST-MAPP, NCT01371202), recruitment (n = 74) has been completed and results are pending.
Common side effects of sunitinib are fatigue, nausea, vomiting, diarrhea, taste changes, heartburn, severe hypertension, and myelosuppression.
Another TKI, the c-Met inhibitor cabozantinib, which was more effective than sunitinib in renal cell carcinoma and in human PCC/PGL primary culture [157][158][159], is currently being investigated in a phase II clinical study in 11 metastatic PCC/PGL patients (initial dose: 60 mg, titrated down on the basis of tolerability) (NCT02302833). Preliminary data have shown tumor size reduction and disease stabilization in most patients, with a median PFS of 11.2 months [160]. The side effects of cabozantinib are similar to those of sunitinib.
The TKI axitinib (AG-013736) is currently being investigated in a phase II non-randomized clinical trial including 14 patients with metatstatic PCC/PGL (NCT01967576). Moreover, another phase II clinical trial is evaluating the efficacy of the TKI lenvatinib, inhibiting VEGFR1/2/3, in advanced PCC/PGL (NCT03008369). The pazopanib trial had to be terminated due to gastrointestinal and severe cardiovascular events [161].
Treatment with the mTORC1 inhibitor everolimus led to disease stabilization in five out of seven patients with advanced PCC/PGL in a small phase II study (NCT01152827) [162], but another study showed no effect of the agent on its own [163]. However, one patient treated with 25 mg sunitinib in combination with 4 mg of the mTORC1 inhibitor rapamycin (as mentioned above) experienced maintained long-term disease control [155]. Accordingly, it is possible that targeted therapies in combination may be more effective at lower doses compared to single treatment approaches.
Consistently, we have already shown in several in vitro and in vivo studies that targeted drug combinations show synergistic anti-tumor effects in PCCs/PGLs [159,[164][165][166][167]. Moreover, very recently, we established a method to screen multiple targeted drug combinations-some are already in use for other types of cancers-ex vivo in human PCC/PGL primary cultures of individual patient tumors [159]. These data may then be correlated to the signaling pathway alterations and the individual genetic background of the tumor [159]. This will hopefully pave the way to customized combination therapy to target individual patient tumors.
Pseudo-hypoxia may prevent the immune system from recognizing cluster 1-related PCCs/PGLs through inactivation of cytotoxic T-cell lymphocytes, activation of immune-suppressive monocytes and increased expression of the immune checkpoint protein programmed death-ligand 1 (PD-L1) and its receptor [169][170][171]. Therefore, immunotherapy is currently being studied in advanced PCC/PGL in two different phase II clinical studies [nivolumab plus ipilimumab (NCT02834013) and pembrolizumab (NCT02721732), respectively].

Conclusion/Practical Tips
In the case of progression after chemotherapy or radionuclide therapy, or if chemotherapy or radionuclide therapy are not possible/tolerated by the patient, the TKIs sunitinib (37.5 mg or 50 mg daily) or cabozantinib (60 mg daily, or titrated down to a tolerable dose) may be considered, especially for RET and SDHx-mutation carriers.
Several targeted therapies (PI3K inhibitors in combination with mTORC1 inhibitors, HIF-2α inhibitors, PARP inhibitors, SSTR2 analogs, HDAC inhibitors, DNA demethylating agents) and immunotherapy are currently under investigation and may have strong potential for future personalized therapy approaches.

Outlook
Customized combination therapy (targeted therapy combinations [159], combinations of targeted therapy with immunotherapy, or targeted therapy combined with PRRT) to target individual patient tumors depending on their underlying germline/somatic mutation and disease characteristics, are likely to be the future directions of therapeutic options for these fascinating but complex tumors.