Introduction
Neuroendocrine tumors (NETs) are a diverse group of neoplasms characterized by their endocrine secretion and histologic features and are often well-differentiated and slow-growing (1). They most commonly occur in the gastroenteropancreatic tract, but are also found in the lung and the rest of the body (1,2). Pancreatic neuroendocrine tumors (PNETs) are rare neoplams; in 2004, their incidence was 0.22 per 100,000 in the United States (3). PNETs are classified clinically as either functioning or non-functioning based on their clinical endocrine manifestations (4). Most PNETs (90%) are non-functioning (3). Unlike functioning PNETs such as insulinomas and gastrinomas, non-functioning PNETs are often malignant and usually present at a late stage, often including a huge mass, local invasion, and distant metastases (5–7).
Modern cancer care highly depends on imaging technologies. For general imaging, positron emission tomography (PET) using 18F-fluorodeoxyglucose (18F-FDG) is a powerful and common functional imaging technique that can provide information about the functional and metabolic characteristics of malignancies, tumor stage, therapeutic response, and tumor recurrence (8). The advent of combined 18F-FDG PET with computed tomography (18F-FDG PET/CT) allows anatomic correlation and exact localization of lesions (9). Unfortunately, FDG PET has limited value for imaging gastroenteropancreatic NETs and only tumors with low differentiation and high proliferative activity have demonstrated an elevated FDG uptake (10–13). In contrast to functioning PNETs, which are usually well differentiated and slow growing, non-functioning PNETs are often found incidentally or through symptoms caused by their enlargement or metastatic spread; they are often malignant and usually present at an advanced stage (14). Therefore, it is reasonable to infer that non-functioning PNETs could be detected by 18F-FDG PET. However, there is currently no published evidence of 18F-FDG PET/CT in evaluating non-functioning PNETs, and generally 18F-FDG PET/CT is considered to have limited value in assessing non-functioning PNETs (10–13). Moreover, 18F-FDG PET/CT is not recommended to detect non-functioning pancreatic neuroendocrine tumors by the newest edition of NCCN guideline (15).
In the present study, we evaluated the utility of whole-body 18F-FDG PET/CT in the detection and staging of non-functioning PNETs.
Materials and methods
Patients
This study retrospectively reviewed 31 patients with PNETs referred to Shanghai Cancer Center, Shanghai, China (single institution) who underwent 18F-FDG PET/CT from January 2010 to February 2014. One patient underwent transcatheter arterial chemoembolization before PET/CT scan. Tumor size was measured intraoperatively or by contrast-enhanced CT scan for patients without operation. The resected or biopsy tissues were prepared for histological and immunocytochemistry examination (cell cycle-associated Ki-67 antigen, insulin and gastrin). The functioning category of PNETs was determined by clinical symptoms and immunocytochemistry examination. The TNM staging of PNETs was determined according to the newest edition of the AJCC Cancer Staging Handbook (15). Tumor grade was defined as low grade (G1), intermediate grade (G2), and high grade (G3) according to the NCCN guidelines (15). The study protocol conforms to the ethical guidelines of the World Medical Association Declaration of Helsinki, and was approved by the ethics boards of Fudan University Shanghai Cancer Center. Written informed consent was obtained from all the patients participating in the study.
The whole-body <sup>18</sup>F-FDG PET/CT protocol
The whole-body FDG PET/CT was performed as previously described (16). Briefly, 18F-FDG was made automatically by cyclotron (Siemens CTI RDS Eclipse ST; Knoxville, TN, USA) using an Explora FDG4 module. All patients had been fasting for ≥6 h and at the time of the tracer injection, normal glucose plasma levels were confirmed. Scanning was started 60 min after intravenous administration of the tracer (7.4 MBq/kg). The images were acquired on a Siemens biograph 16HR PET/CT scanner with a transaxial intrinsic spatial resolution of 4.1 mm. CT scanning was first initiated from the proximal thighs to the head, with 120 kV, 80–250 mA, pitch 3.6, and rotation time 0.5 sec. A PET emission scan that covered the identical transverse field of view was carried out immediately after CT scanning. PET image data were reconstructed iteratively by using the CT data to attenuate correction.
Image interpretation
The image interpretation was performed as previously described (16). Two experienced nuclear medicine physicians independently assessed the data for each patient. The reviewers reached a consensus in cases of discrepancy. Image review and manipulation was carried out on a multimodality computer platform (Syngo; Siemens). Quantification of metabolic activity was acquired using the standardized uptake value (SUV) normalized to bodyweight, and the maximum SUV (SUVmax) for each lesion was calculated. A PET lesion was considered as positive if the SUVmax was ≥2.5 (16). For the detection of distant metastases, we adopted the same functional criteria (SUVmax ≥2.5). Suspicious lesions detected by CT scans, such as nodules in the liver or the peritoneum, were also included for follow-up in order to confirm whether they were false-negative cases of PET/CT. Images were independently interpreted and compared with the histopathologic findings at surgery (5 patients) or with serial imaging and clinical follow-up findings (7 patients).
Statistical analysis
Demographics and clinical characteristics of patients with or without SUVmax of the primary tumor >2.5 were compared using Fisher’s exact probability test. The statistical software used was StataSE11.0 (StataCorp, College Station, TX, USA). A P-value of <0.05 was considered statistically significant.
Results
The study group consisted of 31 cases with non-functioning PNETs and 18F-FDG PET/CT detected 28 (90.3%) primary lesions (Figs. 1 and 2). The study cohort included 16 women and 15 men, age range 25–77 years (median age, 48 years) (Table I). In the subgroup of patients aged ≤50 years, 100% patients had SUVmax ≥2.5, which was similar to that of the subgroup of patients aged >50 years (78.6%) (Table II). The percentage of patients with SUVmax ≥2.5 was similar for women and men (93.8 and 86.7%, respectively). Thirteen lesions were located at the head of the pancreas and 18 at the body and tail of the pancreas. Size of lesions ranged from 1.4 to 10.0 cm (mean ± SD, 4.9±2.3 cm). In 8 cases of tumor size ≤3.0 cm, 62.5% of patients had SUVmax ≥2.5. However, in 23 cases of tumor size >3.0 cm, all the patients (100.0%) had SUVmax ≥2.5 (Fig. 2B). According to the AJCC TNM staging system, 6 patients had stage I tumors, 12 had stage II, and 13 had stage IV tumors. All 25 patients with stage II and IV tumors had SUVmax ≥2.5, compared with only 50.0% of patients with stage I tumors (Fig. 2C). Of all 26 cases with grade information, 3 cases had G1 tumors, 16 had G2 tumors, and 7 cases had G3 tumors. The percentage of SUVmax ≥2.5 in the cases with G1 tumors was lower than that of the cases with G2 and G3 tumors (66.7 and 95.7%, respectively) (Fig. 2D).
For all 12 patients with stage IV tumors (liver metastases of case 15 were resected before PET/CT examination), 42 distant metastatic lesions were found and confirmed by biopsy or serial imaging and clinical follow-up findings (Table III). Thirty-eight (90.5%) lesions had a SUVmax ≥2.5 during PET/CT scans and only 4 (9.5%) lesions demonstrated normal 18F-FDG distribution. Eight of the 31 (25.8%) cases of PNETs had changed their clinical management based on results of the 18F-FDG PET/CT examinations. Among these 8 cases, 4 cases underwent one-stage resection for both the primary tumor and distant metastases and 4 cases abandoned curative surgery and underwent systematic treatments.
Discussion
Imaging technologies are essential to detect non-functioning PNETs and guide their treatment. At present, because of the high expression of somatostatin receptors on NET cells, somatostatin receptor scintigraphy with 111In octreotide is a standard functional imaging technique in the detection and staging of NETs (11,17). However, poorly differentiated or aggressive NETs with high proliferative rates may not be localized on 111In octreotide scanning (17). Kisker et al (17) found that somatostatin receptor scintigraphy visualized only 4 of 10 cases with non-functioning PNETs, which are often poorly differentiated and aggressive. In another study, Rickes et al (18) showed that the sensitivity and specificity of somatostatin receptor scintigraphy for diagnosing PNETs was 54 and 81%, respectively. Therefore, additional functional imaging modalities are needed to detect non-functioning PNETs.
The application of 18F-FDG PET in the detection of NETs depends on the grade of tumor differentiation and the biological characteristics of the tumor (10). NETs are mostly well-differentiated and low-malignancy tumors and 18F-FDG PET frequently fails to visualize them (10). However, 18F-FDG accumulation is often observed in less-differentiated NETs with high proliferative activity (10,19). It has been reported that tumors with increased FDG accumulation appear more aggressive and correlate with poor prognosis (20). Because a great proportion of non-functioning PNETs are poorly differentiated with high proliferative activity, it is reasonable to infer that they can be detected by 18F-FDG PET. However, there has been no published evidence of 18F-FDG PET in detecting non-functioning PNETs.
In the present study, we showed that 90.3% of non-functioning PNETs can be visualized by 18F-FDG PET/CT, indicating that 18F-FDG PET/CT can be employed to detect non-functioning PNETs. The implication 18F-FDG PET/CT in PNETs changed the clinical management of 8 cases. In addition, 18F-FDG PET/CT was found to accumulate in 90.5% of distant metastatic lesions, suggesting that 18F-FDG PET/CT can be used to stage non-functioning PNETs. Other PET imaging agents, such as 11C-5-hydroxytryptophan (11C-5-HTP) and 68Ga-DOTA-Tyr3-octreotide, have been used to detect NETs with promising results (13,21–23). For example, in 2011, Naswa et al (16) showed that 68Ga-labeled [1, 4, 7, 10-tetraazacyclododecane-1, 4, 7, 10-tetraacetic acid]-1-NaI(3)-octreotide (68Ga-DOTA-NOC) PET/CT had a sensitivity of 78.3 for primary tumor and 97.4% for metastases in gastroenteropancreatic NETs, findings that are similar to our study using 18F-FDG PET/CT. However, these imaging agents are still not widely applied in clinical practice compared with 18F-FDG PET/CT.
Because the uptake of 18F-FDG in NETs is closely related to differentiation status and tumor stage (10,12,13,19), we also examined whether tumor size, tumor grade, and TNM stage affect the uptake of 18F-FDG. We found that the uptake of 18F-FDG in PNETs was significantly associated with tumor size and TNM stage, indicating that 18F-FDG PET can be used to detect and stage non-functioning PNETs. In addition, although there was no significant association between increased 18F-FDG uptake and tumor grade, we found that FDG-PET/CT may aid in the diagnosis and staging of PNETs with G2 and G3 tumors (SUVmax ≥2.5, G1, 66.7%, G2 and G3, 95.7%, respectively, P=0.222).
In conclusion, 18F-FDG PET/CT can be used to detect, stage, and conduct surveillance of non-functioning PNETs. In the detection stage, 18F-FDG PET/CT can be applied to a fully evaluated tumor burden to guide treatments, especially for curative and cytoreductive surgery. It also can be used to assess therapeutic response including cytotoxic chemotherapy and targeted therapy and monitor disease recurrence. However, further systematic analyses with even larger study populations will be needed to confirm the role of 18F-FDG PET/CT in the clinical management of non-functioning PNETs.