We reviewed the medical records of 78 patients with the histopathological or clinical diagnosis of PPGLs who visited Severance Hospital, Seoul, Korea between 2006 and 2019, and have been managed by a single, well-experienced endocrinology specialist (corresponding author Y.R.). Histopathological diagnosis of PPGL was based on histologic confirmation by pathologists. Four patients did not undergo surgery due to having an unresectable mass or individual preference. However, they were clinically diagnosed with PCCs because of the apparent adrenal mass on imaging studies such as computed tomography (CT), positron emission tomography, or iodine-131-labeled metaiodobenzylguanidine scintigraphy and elevated catecholamine level in the plasma or urine.

Of the 78 patients, 57 were selected for inclusion in this study (Fig. 1). Patients who refused to undergo any form of genetic testing due to personal or economic reasons were excluded (n=18). Also, patients who underwent Sanger sequencing, but no mutation was found, were excluded (n=2). These exclusion criteria were applied since we could not evaluate the genetic background or rule out the possibility of patients having other meaningful mutations in terms of the pathogenesis of PPGLs, since the Sanger sequencing included only a very limited number of genes (i.e., RET, NF1, VHL). In the same context, a patient who had a silent mutation in the RET gene by Sanger sequencing was also excluded (n=1) because it was considered as having no mutation. As a result, the included patients are those who underwent targeted NGS in our hospital (n=41), and those who had germline mutations found either by Sanger sequencing in our hospital (n=11) or in other hospitals (n=5).

The medical records of the 57 patients were thoroughly reviewed. Information, including age, age of onset, sex, family history, location of the tumor, tumor size, metastasis, recurrence, presence of other tumors, discovery route, symptoms and signs, histologic examination results, and biochemical profile, were extracted. When a patient had at least one family member with PPGL or second-degree relatives with a syndrome related to PPGL, the patient was considered to have a family history. The tumor size was represented by its largest diameter measured histologically or on images if histological findings were not present. Only cases with distant metastasis counted as malignant PPGL, conservatively with high specificity but relatively low sensitivity [22]. It is known that distant metastasis is the only reliable factor that can distinguish malignant PPGL among other clinical, biochemical, or histological features [23,24]. The time at which malignancy was detected was also recorded. Recurrence was defined as the reappearance of PPGL, i.e., presence of a newly discovered mass on imaging studies with elevation in catecholamines levels after the complete removal of the tumor by surgical resection [25,26]. Recurrence included both local and distant recurrence. The tumor discovery route was classified as an incidental finding on CT or ultrasonography regardless of the presence of symptoms and suspected tumors based on the presenting symptoms known to correlate with PPGL when the patient visited the hospital. An incidental finding of a mass in the adrenal gland is called adrenal incidentaloma, and additional workup was done for further evaluation and proper diagnosis of the mass. Uncommon discovery routes included the detection of gene mutations before PCC was detected in a family study. For biochemical profile, plasma or 24-hour urine epinephrine, metanephrine, norepinephrine, normetaepinephrine, dopamine, and urinary vanillylmandelic acid levels were measured. PPGL was considered functional if the catecholamine levels were elevated beyond the normal upper limit, adrenergic when epinephrine/metanephrine level was predominantly elevated; noradrenergic when norepinephrine/normetanephrine level was predominantly elevated; and adrenergic/noradrenergic when both epinephrine/metanephrine and norepinephrine/normetanephrine levels were elevated. When all catecholamine levels were normal, the tumor was classified as silent [27].

The research protocol was approved by the Institutional Review Board of Severance Hospital, South Korea (IRB No. 4-2019-4088). Informed consent was waived by the board.

Targeted NGS was strongly recommended to all patients who were diagnosed with PPGLs as a routine test since January 2017. Before targeted NGS was implemented in our hospital (pre-NGS period), Sanger sequencing was conducted selectively in patients who demonstrated features of familial PPGLs, such as MEN2, NF1, and VHL disease. Even after adopting targeted NGS (post-NGS period), Sanger sequencing was the diagnosis of choice for patients who had syndromic features that strongly implied specific gene mutations and family members of patients with germline mutations.

We conducted targeted NGS in 41 patients with PPGLs from March 2017 to October 2019. We used our hospital’s novel NGS panel to detect gene mutations of 400 genes for hereditary endocrine disorders (Supplemental Table S1). There were 19 PPGL-related genes in the panel (alpha thalassemia mental retardation syndrome X-linked [ATRX], BRAF, Egl-9 family hypoxia-inducible factor 1 [EGLN1], endothelial PAS domain-containing protein 1 [EPAS1], fibroblast growth factor receptor 1 [FGFR1], fumarate hydratase [FH], Harvey rat sarcoma viral oncogene homolog [HRAS], kinesin family member 1B [KIF1B], Myc-associated protein X [MAX], NF1, RET, SDH A/AF2/B/C/D [SDHA/SDHAF2/SDHB/SDHC/SDHD], transmembrane protein 127 [TMEM127], tumor protein 53 [TP53], and VHL) [17]. Our hospital’s targeted NGS followed the 2017 American College of Medical Genetics and Genomics (ACMG) guidelines [28]. Of the five categories in ACMG guidelines, our hospital only reported three categories: pathogenic, likely pathogenic, and variant of unknown significance (VOUS).

When a pathogenic or likely pathogenic mutation was found by targeted NGS or germline mutation of PPGL-related susceptibility gene was detected by Sanger sequencing, a patient was considered to have mutation of known pathogenicity. If a patient had a mutation of known pathogenicity, the genetic analysis result counted as positive, which indicates PPGL with a germline mutation of a susceptibility gene. If a patient had VOUS mutations only, the genetic analysis result was considered negative. Meanwhile, some patients had VOUS mutations of genes that are known to be related to the pathogenesis of PPGL. Although physician assumed that these VOUS of PPGL-related genes might have clinical significance, they were not considered positive as the pathogenicity of mutations has not been proven yet due to insufficient data.

For patients with germline mutations of PPGL-related genes, information on mutated genes and specific loci were collected. Mutated genes were classified into molecular clusters as previously described: pseudohypoxia, Wnt signaling, and kinase signaling group [15].

Family members of genetically confirmed patients were recommended to undergo genetic analysis. Fourteen individuals, who were family members of nine PPGL patients with germline mutations, underwent Sanger sequencing for the same mutation as detected in the respective family member with PPGL.

Clinical characteristics were presented as mean±standard deviation for parametric continuous variables, as median (interquartile range) for nonparametric continuous variables, and as counts (percentage) for categorical variables. For continuous variables, the Shapiro-Wilk test was used to test for normality. Clinical characteristics were compared according to the genetic analysis results and molecular clusters of patients of all periods. For parametric continuous variables, comparisons were made using a t test between two groups and one-way analysis of variance (ANOVA) between three groups. Comparisons were conducted using the Kruskal-Wallis test for nonparametric continuous variables and chi-square test for categorical variables. A two-sided P<0.05 was considered to be statistically significant. All statistical analyses were performed using Stata version 14.2 (Stata Corp., College Station, TX, USA).

Of the 78 patients with PPGLs, 61 patients (78.2%) were diagnosed with PCC, and the other 17 patients (21.8%) were diagnosed with PGL. After excluding patients, we extensively reviewed medical records of 57 patients (73.1%) who underwent targeted NGS or were found to have germline mutations of a susceptibility gene by direct sequencing. The descriptive clinical characteristics of these patients are presented in Table 1. The mean age of onset was 42.8±14.8 years, and 31 patients (54.4%) were females. Forty patients (70.2%) were diagnosed with PCC, and the other 17 (29.8%) were diagnosed with PGL.

The location of PCC was either the right (40.0%), left (40.0%), or bilateral (20.0%) adrenal glands. While three PGLs (17.7%) were parasympathetic as they were located in the head and neck region, including the mediastinum, 14 PGLs (82.4%) were sympathetic, located in thoracoabdominopelvic regions, including retroperitoneum, pancreas, para-aortic space, and bladder [29]. All of parasympathetic PGLs were biochemically silent. The median tumor size was 3.7 cm. The metastasis and recurrence rates were 26.3% and 19.3% for PPGL, respectively. Among 15 patients with metastasis, seven cases were synchronous, whereas eight cases were metachronous. The site of distant metastasis included bone (n=5), lymph nodes (n=5), liver (n=3), and lungs (n=3). Eighteen patients (31.6%) had previously diagnosed tumors in other organs: medullary thyroid carcinoma (MTC, n=8), papillary thyroid carcinoma (n=3), renal cell carcinoma (n=3), and hemangioblastoma (n=3) among others. Forty-three patients presented with symptoms related to PPGLs, including hypertension (n=26), palpitation (n=17), headache (n=8), weight loss (n=6), abdominal pain (n=5), dizziness (n=4), and sweating (n=3). PPGL was detected incidentally by CT or ultrasonography in 21 patients (41.2%), whereas 28 patients (54.9%) had symptoms of PPGL at the first visit. In two patients (3.9%), gene mutations were first detected during a family study, and the following workup revealed the presence of PCCs. Functional tumors accounted for 70.2% of the cases, with the noradrenergic type being the most common (52.6%).

Fourteen patients received Sanger sequencing in pre-NGS period, while the other 43 patients underwent genetic testing by either direct sequencing (n=2) or targeted NGS (n=41) in post-NGS period.

Among patients of all periods, 28 patients had positive genetic analysis results, as they were found with germline mutations of PPGL-related genes with pathogenicity, detected either by targeted NGS (n=2 for pathogenic, n=10 for likely pathogenic) or Sanger sequencing (n=16). The other 29 patients were regarded to have negative genetic analysis results. Among those with negative results, seven patients had VOUS mutations of susceptibility genes, while 22 patients had only VOUS mutations of genes unrelated to PPGL.

Meanwhile, to obtain a more accurate prevalence of germline mutations among PPGL patients, we took into consideration only the patients who received gene analyses during post-NGS period. Since 2017, the year at which targeted NGS was recommended to all patients, a total of 43 PPGL patients underwent genetic testing (n=41 for targeted NGS, n=2 for Sanger sequencing). Among those 43 patients, 14 patients had positive results, so the proportion of germline mutations is estimated to be 32.6% (14 out of 43).

Comparison of clinical characteristics, including age, age of onset, sex, family history, location of the tumor, tumor size, metastasis, time of metastasis discovery, recurrence, presence of other tumors, discovery route, and biochemical profile, between positive and negative genetic analysis results is presented in Table 1. PPGL patients with germline mutations were more likely to have a family history (P=0.001). Although not statistically significant, patients with germline mutations tended to have a lower age of onset and prevalence of bilateral PCCs and a higher rate of metastasis and recurrence.

Among 28 patients with positive results, 19 different germline mutations from seven susceptibility genes were discovered. The most frequently mutated genes were SDHB (n=9), followed by RET (n=8), VHL (n=6), NF1 (n=2), MAX (n=1), SDHA (n=1), and SDHD (n=1) (Fig. 2). Sixteen patients had classical familial syndromes. Of these, eight had MEN2, six had VHL syndrome, and two had NF type 1. Meanwhile, seven VOUS mutations were found in four types of genes related to PPGL; SDHB (n=2), EPAS2 (n=2), KIF1B (n=1), SDHC (n=1), and TMEM127 (n=1) (Supplemental Table S2).

Seventeen out of the 28 positive mutations (60.7%) were missense mutations (Fig. 3). Eight mutations (28.6%) were frameshift mutations, and one mutation (3.6%) was a nonsense mutation.

Mutated genes with pathogenicity were classified into three molecular clusters: pseudohypoxic tricarboxylic acid (TCA) cycle-related group including SDHB, SDHA, and SDHD (n=11); pseudohypoxic VHL/EPAS1-related group including VHL (n=6); and a kinase-signaling group including RET, NF1, and MAX (n=11). Clinical characteristics were compared between the molecular clusters (Table 2). The diagnosis of PCC or PGL significantly differed between the clusters (P=0.003). While the TCA cycle-related group patients were comparable in terms of the type of PPGL (PCC 36.4% vs. PGL 63.6%), the kinase signaling group consisted of PCCs only (PCC 100.0% vs. PGL 0.0%). Patients in the kinase signaling group were more likely to have a family history (63.6%), followed by those in the pseudohypoxic VHL/EPAS1-related group (33.3%), while the pseudohypoxic TCA-cycle related group had no family history (P= 0.006). Metastasis rate was the highest in the pseudohypoxic TCA cycle-related group (63.6%), followed by the pseudohypoxic VHL/EPAS1-related group (50.0%), and the kinase signaling group had no patient with distant metastasis (P=0.006). The kinase signaling group had the highest percentage of patients who had other tumors (72.7%), followed by the pseudohypoxic VHL/EPAS1-related (50.0%) and pseudohypoxic TCA cycle-related (9.1%) groups (P=0.010). A significant difference in biochemical profile between the molecular clusters was detected (P=0.036). Notably, functional PPGLs in the pseudohypoxic group were all noradrenergic type. Although there was no statistically significant difference, all the pseudohypoxic TCA cycle-related group cases had unilateral PCC. Head and neck PGL was detected in only the pseudohypoxic TCA cycle-related group.

Detailed clinical and genetic characteristics, including the specific mutation loci, of patients with positive genetic analyses results are presented in Table 3. Information on patients with VOUS of PPGL-related genes can be found at Supplemental Table S2. While most of the genes have already been reported in previous studies, four novel mutations were discovered in the present study: EPAS1 c.1250G>A (p.Gly417Glu), NF1 c.6215delA (p.His2072LeufsTer10), NF1 c.6777del (p.Gly2260fs), and SDHC exon 2–6 duplication.

Fourteen family members of nine patients underwent Sanger sequencing for family study. Eight of these family members (57.1%) had the same germline mutation as that of the patient: SDHB (n=4), RET (n=3), and NF1 (n=1). Those with germline mutations underwent screening workup, including history taking, physical examination, measurement of plasma or urine catecholamine levels, and imaging such as CT for the detection of PPGLs. For genes known to cause syndromic diseases, further evaluation was performed, i.e., thyroid sonography for patients with RET mutation considering the possibility of MTC. As a result, two persons were newly diagnosed with PCCs. Both daughters of a patient with PCC and MEN2A (Case no. 17) had the same RET gene mutation. While one only had C-cell hyperplasia, another (Case no. 13) was diagnosed with PCC, MTC, and MEN2A. Similarly, the daughter (Case no. 18) of a patient with PCC and MEN2A (Case no. 20) was found to have RET gene mutation and was diagnosed with PCC, MTC, and MEN2A. The daughter of another patient with PCC and NF1 (Case no. 12) was also diagnosed with NF1 due to the same mutation in the NF1 gene; although she did not have PPGL, typical clinical features of neurofibromatosis, café-au-lait spots, and neurofibroma were noted. Four family members who had SDHB mutations currently showed no phenotype, as no evidence of PPGL was found after thorough screening workup, including biochemical and radiological tests, and they did not have any symptoms and signs. This is probably due to the low penetrance of the SDHB gene [30] and the possibility of delayed manifestation.