Preoperative ¹⁸F-FDG PET/CT in Pheochromocytomas and Paragangliomas Allows for Precision Surgery

Abstract

Background:

Fluorodeoxyglucose (¹⁸F-FDG) positron emission tomography/computed tomography (PET/CT) imaging is recommended in patients with metastatic pheochromocytoma (PC) and paraganglioma (PGL). There are no data on whether routine preoperative ¹⁸F-FDG PET/CT in all patients with PC/PGL impacts surgical management.

Objective:

The aim of this study was to determine whether routine preoperative ¹⁸F-FDG PET/CT imaging affects the surgical management of patients with PC/PGLs.

Methods:

We analyzed clinical, biochemical, genetic, and anatomic imaging data in 93 consecutive patients with PC/PGL who collectively underwent a total of 100 operations and who had preoperative ¹⁸F-FDG PET/CT imaging.

Results:

Of 100 operations, preoperative ¹⁸F-FDG PET/CT showed additional lesions compared to anatomic imaging in 15 cases. These patients were more likely to undergo an open surgical approach (P < 0.05). Presence of genetic mutation, redo operations, sex, age, or tumor size had no significant association with finding additional lesions on ¹⁸F-FDG PET/CT.

Conclusions:

Additional lesions detected on preoperative ¹⁸F-FDG-PET/CT imaging have an impact on the surgical approach in patients with PC/PGLs. Therefore, surgeons should routinely obtain ¹⁸F-FDG-PET/CT imaging in patients with PC/PGL to allow for a more precise surgical intervention.

Pheochromocytomas (PC) originate from chromaffin cells in the adrenal medulla, whereas paragangliomas (PGL) are tumors arising from the chromaffin cells of the sympathetic and parasympathetic ganglia. The estimated incidence of PC/PGL in hypertensive patients is 0.1% to 0.6%, and these tumors are found in 0.05% of patients during autopsy.¹ Metastatic PC/PGLs are defined as tumor sites wherein chromaffin cells are not present, such as the lymph nodes or bones, liver, and lung. The estimated incidence of metastatic PC/PGLs is 0.3 to 0.7 cases per million, which accounts for 2.4% to 26% of all PC/PGL cases.² PC/PGL may be sporadic or inherited. There has been considerable progress made in our knowledge of the susceptibility genes that predispose to PC/PGLs, and there are 15 well-characterized germline mutations related to these tumors, with some genotype–phenotype associations.³ The frequency of germline mutation in patients with PC/PGL is up to 40% in patients of all ages and can be as high as 80% in the pediatric population.⁴ Some inherited PC/PGLs are associated with tumor multiplicity, metastatic disease, and a risk of recurrent PC/PGL and preoperative knowledge of the presence of a germline mutation alters the surgical management of patients with PC/PGL.⁵

Advances in anatomic and functional imaging have rapidly constituted these imaging modalities as an indispensable part of the evaluation of patients with PC/PGL. Practice guidelines from both the United States and Europe recommend fluorodeoxyglucose (¹⁸F-FDG) positron emission tomography/computed tomography (PET/CT) scanning in patients with metastatic PC, a germline SDHB mutation, or PGL.^{6,7 18}F-FDG PET/CT was shown to have higher sensitivity compared to CT and magnetic resonance imaging (MRI) for detecting bone metastases in patients with PC/PGL.⁸ Although anatomic imaging with CT and/or MRI is highly accurate for localizing PC/PGL, routine preoperative imaging with 18F-FDG PET/CT in patients with PC/PGL may affect the surgical approach if a patient is found to have additional lesions. It may also change the indication for surgical management, for example, from curative to palliative intent if a patient is found to have widely metastatic disease.

The aim of this study was to determine whether routine preoperative ¹⁸F-FDG PET/CT scanning affcets surgical management and to examine what clinical and genetic factors may be associated with positivity on ¹⁸F-FDG PET/CT scanning in patients undergoing surgical intervention for PC/PGL.

METHODS

One hundred and fifteen patients underwent 127 operations for PC/PGL at the National Institutes of Health Clinical Center between November 2009 and July 2017. The patients had biochemical and or histologic evidence of PC/PGL. All patients were enrolled in clinical protocols (NCT00004847, NCT01005654) at the National Institutes of Health Clinical Center after written informed consent. Ninety-three patients who underwent 100 operations had all three preoperative imaging modalities with CT, MRI, and ¹⁸F-FDG PET/CT and were included in this study (Fig. 1). The study was approved by the Institutional Review Boards of the National Cancer Institute and the Eunice Kennedy Shriver National Institute of Child Health and Human Development. The operations were classified into first-time operations or reoperations. A reoperation was defined as a patient who had a previous operation at the same site as the newly diagnosed tumor. Preoperative genetic testing was performed for known PC/PGL susceptibility genes (RET, VHL, NF1, SDHA, SDHB, SDHC, SDHD, MAX, FH, TMEM127, and SDHAF2) in a Clinical Laboratory Improvement Amendments–certified laboratory in all patients.

FIGURE 1.

Patients included in the study cohort based on imaging modalities performed and categorized by initial versus redo operations and the surgical approach used.

Imaging with ¹⁸F-FDG PET/CT, CT, and MRI

¹⁸F-FDG PET/CT, CT, and MRI were performed before 100 operations for PC/PGLs, within 6 months before the intervention (Fig. 1). ¹⁸F-FDG PET/CT imaging was performed approximately 60 minutes after intravenous administration of 10 mCi of ¹⁸F-FDG in patients who weighed <90kg and 15 mCi of ¹⁸F-FDG in patients who weighed >90kg.Blood glucose levels were confirmed to be <150mg/dL in all patients before radiotracer administration. The patients were scanned from the base of the skull to the mid-thighs. A noncontrast CT scan of the head, neck, chest, abdomen, and pelvis was used for attenuation and anatomical localization. A Siemens SOMATOM Force CT scanner (Washington,DC) was used to obtain multidetector helical images of the chest, abdomen, and pelvis with 2mm cuts following intravenous contrast administration, 1.4mL/s of Isovue-300, as well as oral Omnipaque contrast administration. A Siemens MAGNETOM Aera 1.5-Tesla MRI scanner (Washington, DC, USA) was used to obtain multiplanar and multisequence images of the abdomen and pelvis following intravenous injection of Gadavist. All scans were compared with preoperative CT and MRI results, and the official board-certified radiologist’s interpretations were used to determine positive and negative lesions. All patient history, biochemical findings, genetic testing results, and imaging findings were reviewed preoperatively in a multidisciplinary case conference with a board-certified nuclear medicine physician, anatomic radiologist, endocrinologist, and endocrine surgeons to reach a consensus on the imaging findings and planned operative interventions.

We analyzed prospectively collected demographic (age, sex, ethnicity/race) and clinical [body mass index (BMI), laboratory, operative, pathologic, and genetic data] information. The laboratory data included preoperative and postoperative 24-hour urine and plasma dopamine, epinephrine, norepinephrine, free metanephrine, and free normetanephrine. Patients with serum or urine laboratory values elevated to 2 times the upper limit of normal or higher were defined as having functional tumors. If the postoperative levels in patients who had a functional tumor decreased to <2 times the upper limit of normal, patients were defined as having biochemical remission but in 13 patients, there was no abnormal biochemistry preoperatively.

Statistical Analyses

t Test, Fisher exact test, and the chi-square test were used to analyze the data as appropriate. We tested for associations between demographic characteristics, clinical variables, genetic testing results, and biochemical data by ¹⁸F-FDG-PET/CT imaging findings, as well as a direct comparison to anatomic imaging findings with respect to the surgical approach selected. Two-tailed P values were used and reported. A P value <0.05 was considered statistically significant. IBM SPSS Statistics Data Editor (New York, NY) and Microsoft Excel (Redmond, WA) were used for statistical analyses.

RESULTS

Ninety-three patients underwent 100 operations from November 2009 to July 2017. Five patients underwent >1 operation for PC/PGL. The demographic and clinical characteristics of the study cohort are summarized in Table 1.

TABLE 1.

Demographic, Clinical, and Biochemical Data of Study Cohort

Variables		93 patients
Age, y, mean (range)		42 (11–76)
BMI, kg/m2, mean (range)		27 (17–46)
Sex (female/male)		50/43
Type of operation		100 Operations
Initial videoscopic transabdominal/redo videoscopic transabdominal		49/1
Initial open/redo open		15/31
Initial videoscopic retroperitoneal/redo videoscopic retroperitoneal		2/2
PC/PGL/both		62/23/4
Surgery duration, * min, mean (range)		233 (65–615)
Genetic mutation
Present		46
Not present		54
Functional status of tumor (≥2 times the upper limit of normal) at the time of the operation
Functional		87
Nonfunctional†		13
Biochemical remission‡ (2 × upper limit of normal)
Remission/no remission		67/13
No information available		15
Plasma Biochemistries, pg/mL
Preoperative Values (Reference Range)	Mean	Range
Dopamine (0–25)	16	0–3112
Epinephrine (0–57)	25	0–2418
Norepinephrine (84–794)	859	72–14,755
Metanephrine (12–61)	43.5	0–5996
Normetanephrine (18–112)	589	16–23,350
Postoperative Values	Mean	Range
Dopamine (0–25)	9	0–507
Epinephrine (0–57)	5	0–185
Norepinephrine (84–794)	281	100–3236
Metanephrine (12–61)	18	0–357
Normetanephrine (18–112)	70	16–2282
Urinary Biochemistries, μg/24 h
Preoperative Values	Mean	Range
Dopamine (65–400)	280.5	20–9704
Epinephrine (0–21)	6.4	0–789
Norepinephrine (15–80)	132	19–4287
Metanephrine (44–261 normotensive) (<400 hypertensive)	170	0–22,920
Normetanephrine (103–390 normotensive) (<900 hypertensive)	1628	155–31,175
Postoperative Values	Mean	Range
Dopamine (65–400)	199.5	0–683
Epinephrine (0–21)	2.4	0–51
Norepinephrine (15–80)	41.5	4–1837
Metanephrine (44–261 normotensive) (<400 hypertensive)	71.5	0–1558
Normetanephrine (103–390 normotensive) (<900 hypertensive)	328.5	110–15,374

^*Operative time was from skin incision to skin closure.

^†Thirteen patients did not cross the threshold of biochemical positivity.

^‡No biochemical information was available in 15 operations.

Comparison of ¹⁸F-FDG PET/CT and Anatomic Imaging Findings

Before 15 operations, ¹⁸F-FDG PET/CT was discordant with CT and MRI. ¹⁸F-FDG PET/CT showed additional lesions that were suspicious for metastatic disease to the bone, liver, or retroperitoneal lymph nodes or additional primaries that were not seen on CT or MRI (Table 2). These additional lesions were skeletal metastases before 4 operations (Fig. 2), retroperitoneal lesions before 7 operations (Fig. 3), and liver lesions before 2 operations. In one of the operations, both additional liver and retroperitoneal lesions were found. In one of the operations, the additional lesion was consistent with a splenule on histologic evaluation. Of the 8 operations with additional retroperitoneal avidity on ¹⁸F-FDG PET/CT, 5 had additional lesions found during the operation (Fig. 3); furthermore, in 3, no lymphadenopathy was observed intraoperatively, so a lymphadenectomy was not performed (Fig. 4). These 3 patients were biochemically cured and had no persistent/recurrent disease on follow-up imaging. An open surgical approach was used significantly more when additional lesions were detected on ¹⁸F-FDG PET/CT with 11 of 15 operations performed using an open approach (P < 0.05). Age (P = 0.54), sex (P = 0.78), and BMI (P = 0.93) were not significantly associated with finding additional lesions on ¹⁸F-FDG PET/CT.

TABLE 2.

¹⁸F-FDG PET/CT Findings and Demographic, Clinical, and Genetic Features of Patients With PC/PGL

	18F-FDG PET/CT Discordant (15)	18F-FDG PET/CT Concordant (85)	P
Mean age, y (SD*)	44 (15)	41 (16)	0.549
Sex
Female	7/15	45/85	0.781
Mean BMI, kg/m2 (SD)	27 (5.8)	27.1 (7)	0.934
Mean duration of the operation, min (SD) †	271 (98)	227 (110)	0.158
Mean size, cm (SD)	3.9 (2.1)	4.2 (3.9)	0.685
Genetic status
Initial operation/redo	7/8	59/26	0.137
Germline mutation	8	38	0.584
SDHB	4	14	0.464
MEN2A, NF1, VHL	3	12	1.00
Biochemical remission‡	8	59	0.032
Open surgical approach/laparoscopic	11/4	35/50	0.026
Single/multiple PC/PGLs	8/7	59/26	0.244

^*SD, standard deviation.

^†Operative time was from skin incision to skin closure.

^‡No biochemical information was available in 20 patients.

FIGURE 2.

A 29-year-old woman presented with a paraganglioma in the left renal hilum, with the tumor displacing the left renal vein. (A) Additional lesion noted on whole-body anterior 3-dimensional maximum intensity projection (MIP) ¹⁸F-FDG image, a midline lesion (arrow) with an SUV of 6.4. Among the lesions resected, only the lesion corresponding to SUV of 6.4 was proven to be a paraganglioma on histology. The specimen contained 6 lymph nodes with no evidence of metastatic disease. (B) axial fused ¹⁸F-FDG PET/CT image of the abdomen showing a lesion with an SUV of 6.4, anterior to the distal aorta. (C) Corresponding level on axial arterial phase CT with no discernable lesion. (D) Corresponding level of the abdomen on axial T2-weighted MRI with no discernable lesion. Histology of the distal aortic lesion showed a paraganglioma.

FIGURE 3.

A 55-year-old woman with a germline SDHB-mutation and bilateral adrenal pheochromocytoma. An additional lesion was noted on axial fused ¹⁸F-FDG PET/CT (arrow) with an SUV of 3.3 (A) is not discernable on whole-body anterior 3-dimensional maximum intensity projection (MIP) ¹⁸F-FDG image (B). Corresponding level on axial arterial-phase CT (C). Corresponding level on axial T2-weighted MRI (D). The additional lesion was not seen intraoperatively and after resection of bilateral pheochromocytomas the patient was biochemically cured.

FIGURE 4.

A 52-year-old man with a 6.4-cm right retroperitoneal paraganglioma. (A) Additional lesions were found on whole-body anterior 3D maximum intensity projection (MIP) ¹⁸F-FDG image showing a lesion in the left transverse process of T9 thoracic vertebrae (blue arrow) with an SUV of 22.3, and a right jugular foramen lesion eroding into the petrous bone (black arrow). In addition, other lesions including right rib (black arrow), right proximal humerus (black arrow), right posterior sacrum (black arrow), and left proximal femoral shaft (black arrow). (B) Axial fused ¹⁸F-FDG PET/CT image of the chest/abdomen shows the left transverse process lesion in the T9 thoracic vertebrae (arrow) with an SUV of 22.3. (C) Corresponding level of image in (B) on axial non-contrast CT bone window.

Of the 15 patients who had additional lesions detected on ¹⁸F-FDG PET/CT, 6 patients had adrenal involvement and 9 patients had abdominal PGLs. A germline mutation was found before 7 operations (3 SDHB, 1 SDHD, 2 VHL, and 1 FH). The presence of a germline mutation in a PC/PGL susceptibility gene (P = 0.58) and the presence of a specific type of gene mutation (P = 0.46) were not associated with finding additional lesions on ¹⁸F-FDG PET/CT. In addition, tumor size was not significantly associated with detecting additional lesions on ¹⁸F-FDG PET/CT that were not seen by either CT or MRI (P = 0.68).

Seven of 15 cases wherein additional lesions were detected were first-time operations for the given site of disease, and the remaining 8 cases were reoperations. Patients undergoing reoperative surgery did not demonstrate a statistically significant increase in the rate of finding additional lesions on ¹⁸F-FDG PET/CT compared to patients having a first-time operative intervention for PC/PGL. In addition to these 15 cases, the findings on ¹⁸F-FDG PET/CT influenced the surgical decision before 2 other operations. One patient with a germline MEN1 mutation had bilateral adrenal masses, and the PC was lateralized to the left side by ¹⁸F-FDG PET/CT scanning. The lesion was resected and was consistent with a PC on histologic examinations, and the patient’s biochemistries normalized after the operation. Another patient with an NF1 germline mutation had an incidental finding of an adrenal mass during an evaluation for abdominal pain with a normal biochemical profile. The lesion was avid on ¹⁸F-FDG PET/CT, and the patient underwent an adrenalectomy and was found to have a PC on histology.

Factors Associated With an Open Surgical Approach

The following variables were significantly associated with an open surgical approach: additional lesions on ¹⁸F-FDG PET/CT (P = 0.026), presence of SDHB mutation (P = 0.004), and presentation with persistent/recurrent disease (P < 0.01). After performing a linear regression analysis, redo surgery (P < 0.01) and the presence of SDHB (P = 0.016) remained statistically significant factors for an open surgical approach. The tumor size of the largest lesion was not significantly associated with an open surgical approach; however, operations with >1 abdominal PC/PGL resected were more likely to have an open resection (P = 0.005). Patients with a higher BMI had more laparoscopic resections of PC/PGL (P = 0.027).

DISCUSSION

There have been significant advances in our understanding of the pathophysiology and genetic predisposition for PC/PGL and in imaging modalities to detect PC/PGL. In the present study, we report the frequency of detecting additional lesions in an unselected cohort of patients with PC/PGL who had routine ¹⁸F-FDG PET/CT imaging as compared to CT and MRI. Moreover, we found that this information affected the surgical management of patients. Additional factors associated with an open surgical approach were the presence of the germline mutation SDHB, BMI, and operations with >1 abdominal PC/PGL resected. In 4 patients, the presence of additional lesions on preoperative ¹⁸F-FDG PET/CT also altered the indication and goal of surgical intervention from curative intent to palliation of symptoms and signs associated with excess catecholamine in a subset of patients. One patient achieved biochemical remission. In 2 of the patients, plasma-fractionated normetanephrine and norepinephrine decreased by >75% after the operation, and their symptoms have resolved. In 1 patient, the entire abdominal lesion was not able to be resected because of local invasion; therefore, there was no change in her biochemical profile.

Malignancy in PC/PGL is defined by the presence of metastatic lesions in lymph nodes or distant sites where chromaffin cells are not present.² The absence of accurate biochemical markers to assist in the detection of metastatic tumors makes imaging the sole method to preoperatively establish the diagnosis of metastatic PC/PGL, which would potentially modify operative planning and the indication for an operation. The recent European guideline recommended selective use of ¹⁸F-FDG PET/CT preoperatively to screen for metastatic lesions in certain high-risk groups, such as patients with PGL and patients with SDHB germline mutations.⁷ The United States Endocrine Society Clinical Practice guidelines recommend using ¹⁸F-FDG PET/CT in patients with known metastatic disease.⁶ We found that patients not in these guideline categories were found to have additional disease on ¹⁸F-FDG PET that was not seen on CT scan and MRI, and that this information impacted the surgical approach. The reported sensitivities of ¹⁸F-FDG PET for PC/PGL range from 74% to 100%,⁶ and it particularly has a higher detection rate in patients with clinically aggressive and SDHB-associated metastatic tumors.^9,10 In the present study, ¹⁸F-FDG PET/CT was discordant with anatomic imaging before 4 operations in patients with an SDHB mutation, 1 patient with an SDHD mutation, 2 patients with a VHL mutation, and 1 patient with an FH mutation; furthermore, it was similarly discordant before 7 operations in patients with no known germline mutation in PC/PGL susceptibility genes. One of the patient with SDHB, 2 patients with VHL, and 2 with no known germline mutation presented with PC, whereas the rest had PGL.

The accuracy of ¹⁸F-FDG PET/CT in SDHx- and VHL-associated PC/PGLs is hypothesized to be related to the tumor cells’ metabolic shift to hypoxic and angiogenic pathways and, eventually, aerobic glycolysis, which in turn leads to increased expression of glucose transporters and glucose phosphorylation by hexokinases. With recent advances in nuclear medicine, new functional imaging modalities such as ⁶⁸Ga-DOTATATE, ¹⁸F-FDOPA, and ¹⁸F-FDopamine have been evaluated in patients with PC/PGL. Some studies report higher accuracy of these functional imaging modalities than ¹⁸F-FDG PET/CT for metastatic tumors.^11,12 However, higher sensitivities have been reported for ¹⁸F-FDG PET/CT in aggressive tumors than other functional imaging modalities and have been attributed to tumor dedifferentiation resulting in a shift to aerobic glycolysis in the tumor cells.¹³ The validity of this hypothesis has not been proved in PC/PGL, as benign differentiated tumors are frequently ¹⁸F-FDG–avid, which indicates that ¹⁸F-FDG avidity is not necessarily a marker of only aggressive disease.^9,11 The sensitivity of ¹⁸F-FDG PET for patients with a RET mutation and PC/PGL has been reported to be 40%, which is thought to be related to the underlying mechanism of tumorigenesis, the activation of the RET/RAS/RAF/MAPK pathway, which is not thought to greatly affect glucose metabolism.¹⁴ Although ⁶⁸Ga-DOTATATE, ¹⁸F-FDOPA, and ¹⁸F-FDopamine imaging may have a role in some patients with PC/PGL, they are not widely available and not all surgical patients have had these imaging modalities for a head-to-head comparison of their effect on the surgical management of patients. Moreover, ¹⁸F-FDG PET/CT imaging is widely available and can be used at most medical centers.

Our study has several limitations. Because we are a tertiary referral center for patients with PC/PGL, our study may have a selection bias of patients with more aggressive disease and inherited PC/PGL. However, 55% of our cohort had no known germline mutation, and 37% of the patients with a germline mutation had no family history to suggest that the referral of patients with a higher predisposition to inherited PC/PGL accounts for our study findings. Although we found that ¹⁸F-FDG PET/CT detected additional lesions not detected on CT scan or MRI, we cannot be certain that this approach leads to higher biochemical cure without a randomized control trial. Furthermore, false-positive results do occur with ¹⁸F-FDG PET/CT imaging (4 cases in our cohort), so treatment decisions should be discussed by a multidisciplinary care team to maximize optimal treatment approaches.

CONCLUSIONS

In summary, we found 18F-FDG PET/CT detected additional lesions in patients with PC/PGL during routine preoperative imaging that affected the type of surgical approach selected. Thus, we believe it is important to incorporate 18F-FDG PET/CT in the preoperative workup of all patients with PC/PGL. We found no reliable clinical factors that would enable a more selective use of 18F-FDG-PET/CT imaging in these patients.