Introduction

Neuroendocrine neoplasms (NENs) are a relatively rare and highly heterogeneous tumor derived from neuroendocrine cells. The incidence and prevalence of NENs have increased steadily over the past 40 years, with increasing awareness and emergence of better diagnostic tools [1]. This has been accompanied by a concomitant increase in the rate of NEN distant metastases, which negatively affects NEN treatment and survival [2]. Thus, effective NEN monitoring using sensitive imaging approaches is needed to detect progression and adapt treatment strategies.

NENs commonly express somatostatin receptors (SSTRs), making them amenable to molecular imaging with radionuclide-coupled somatostatin analogs as a diagnostic tool [3]. Currently, 68Ga-labeled somatostatin analogs (SSAs) for positron emission tomography/computed tomography (PET/CT) have been used in routine clinical practice [4, 5]. Relative to single-photon emission computed tomography (SPECT), PET has a higher spatial resolution, shorter imaging times, lower radiation exposure, and better lesion detection [6, 7]. Thus, PET scanning using 68Ga-labeled SSAs is critical for tumor detection rate, staging and restaging and post-therapy follow-up [4]. However, the use of 68Ga-labeled PET is limited by the high cost of 68Ge/68Ga generators [8] and the relatively short half-life of 68Ga (68 min). 18F-labeled (18F half-life: 106.9 min) SSAs have high tumor-to-background ratio (TBR) for NEN lesions; thus, these probes may be used as an alternative in NEN imaging and also allow for longer transport times [9, 10].

In recent years, there has been increased research on 18F-labeled agents [10,11,12]. 18F-AlF-NOTA-octreotide (18F-OC) exhibits satisfactory biodistribution and dosimetry profiles with a high NEN lesion detection rate [13]. A comparison of the imaging parameters of 18F-OC and 68Ga-DOTATATE for NENs in a small number of patients found that 18F-OC has excellent dynamics and imaging characteristics [14]. Here, we assessed the clinical applicability and efficacy of 18F-OC relative to those of 68Ga-DOTATATE in a larger group of patients.

Methods

Patients and study design

The research was approved by the institutional ethics review committee of Xiangya Hospital, Central South University for research purposes only (No. 20181001). All study participants gave written informed consent before the start of the study. Patients with clinically confirmed NENs were prospectively recruited into the study. Participants received an intravenous injection of 18F-OC and 68Ga-DOTATATE for PET/CT within 8 days (range: 1–8 days), except patients No. 1 (interval time: approximately 147 days without any treatment) and No. 15 (interval time: approximately 279 days without peptide receptor-radionuclide therapy (PRRT)).

68 Ga-DOTATATE and 18 F-OC preparation

68Ga-DOTATATE was synthesized using the acetone method on a fully automated Modular Lab system (Eckert & Ziegler, Germany), and quality control was performed as previously described [15]. The radiochemical purity of 68Ga-DOTATATE was > 90%. 18F-OC was produced as previously described [16] and under good manufacturing practice guidelines.

PET/CT image acquisition

The study was carried out with a General Electric PET/CT scanner (Discovery 690 Elite, General Electric Health care, Waukesha, Wis). 18F-OC PET/CT imaging was performed 60 min after the radiotracer was intravenously (IV) injected at a dose of 3.7–4.44 MBq (0.1–0.12 mCi) per kilogram of body weight. 68Ga-DOTATATE imaging was performed 50 min after an injection with a total activity of 194.4 ± 37.9 MBq. First, a low-dose CT scan (120 kV; automatic mAs; pitch, 1:1; slice thickness, 3.75 mm; matrix, 512 × 512) was performed from the head to mid-thigh for anatomical localization and attenuation correction. Next, PET scanning was performed, with 2 min per bed position. Finally, images were reconstructed using the 3-dimensional ordered-subsets expectation maximization algorithm with 2 iterations and 23 subsets.

Image analysis

Regions of interest (ROIs) were drawn on fused PET/CT images on a dedicated nuclear medicine AW 4.6 workstation (General Electric Healthcare) to obtain standardized uptake values (SUVs). 18F-OC and 68Ga-DOTATATE images were independently assessed by 2 experienced nuclear medicine physicians who were blinded to the patients and their medical information. The ROIs for measuring the maximum standard uptake value (SUVmax) and mean standard uptake value (SUVmean) in normal organs and tissues and the ROIs for measuring the SUVmax of NEN lesions were drawn on serial images. The mean SUVmax and SUVmean in the reference organs were evaluated by placing 3 consecutive ROIs (including the area with the highest uptake and that on the upper and lower slices based on visual assessment) inside the organ of interest, including pituitary, cerebral cortex, adrenal gland, uncinate process of the pancreas (PU), pancreas (except the PU), stomach, spleen, thyroid, salivary glands, liver, bone, renal parenchyma, small intestine, uterus (female), prostate (male), colon, lung, fat, myocardium, muscle, bladder wall, and blood pool, on both scans. Candidate lesions with activities greater than the physiologic uptake in the involved organs were considered lesions. These lesions were divided into 5 regions or groups: primary tumor, liver metastases, bone metastases, lymph node metastases, and metastases in other organs (lung, muscle, stomach, rectum, peritoneum, soft tissue, and thyroid). For patients with multiple lesions, at most 5 lesions with the highest uptake per organ were included in the uptake analysis. The tumor-to-liver ratio (TLR) and tumor-to-spleen ratio (TSR) were calculated by dividing the SUVmax of different tumor lesions by the SUVmean of the liver and spleen in each patients, respectively. All ratios on corresponding 18F-OC and 68Ga-DOTATATE scans were computed from the same layer on the 2 scans. All discrepant lesions between the images of the 2 radiotracers were identified by other imaging or patient follow-up (computed tomography (CT), magnetic resonance imaging (MRI), and PET/CT) and then classified as true- or false-positive findings.

Statistical analysis

Data analysis was performed using GraphPad Prism 6 (Version 6.01, 2012). Data are expressed as the median (interquartile range). Nonparametric data were compared using the Wilcoxon signed-rank test. P < 0.05 indicates statistical significance.

Results

Twenty patients were prospectively enrolled in the study, and their clinical characteristics are summarized in Table 1. No patients received PRRT treatment between 68Ga-DOTATATE and 18F-OC PET/CT scans. Both radiotracers were tolerated well by all patients, and no adverse events were reported. The physiological uptake of 68Ga-DOTATATE and 18F-OC is shown in Fig. 1. 68Ga-DOTATATE and 18F-OC PET/CT scans were compared at the lesion and region levels and based on SUV.

Table 1 Patient clinical characteristics
Full size table
Fig. 1
figure 1

Uptake of 68Ga-DOTATATE and 18F-OC in normal organs was calculated in patients based on the mean SUVmax (a) and SUVmean (b). Significant differences between 68Ga-DOTATATE and 18F-OC are indicated. **: p =  < 0.01, *: p =  < 0.05

Full size image

Biodistribution of 68 Ga-DOTATATE and 18 F-OC

Similar to that of 68Ga-DOTATATE, the highest SUVmax values for 18F-OC were recorded in the spleen, adrenal gland, renal parenchyma, pituitary gland, liver, and PU. Lower SUVmax and SUVmean values were observed in the salivary glands, myocardium, bone, lung muscle, fat, and cerebral cortex. In most organs, the biodistribution of 68Ga-DOTATATE was not significantly different from that of 18F-OC. Relative to 68Ga-DOTATATE, 18F-OC had significantly lower uptake in organs such as the salivary glands, liver, pancreas, bone, renal parenchyma, and prostate (Fig. 1).

Comparison of tumor detection rates between 68 Ga-DOTATATE and 18 F-OC PET/CT

This study included 20 NEN patients. Examinations using 68Ga-DOTATATE and 18F-OC PET/CT revealed that 19 patients had lesions and 1 patient had no lesions. Follow-up examination confirmed lesions in the 19 patients and that patient No. 1 had no lesions. Table 2 shows the discordant lesions examined by 68Ga-DOTATATE and 18F-OC PET/CT.

Table 2 Patients with discordant lesions on 18F-OC and 68Ga-DOTATATE PET/CT
Full size table

In the region-based comparison, 9 patients had primary tumors on both 18F-OC and 68Ga-DOTATATE images. In addition, there were 4 patients staged with unknown primary lesions. Sixteen patients had metastases on 68Ga-DOTATATE PET/CT, and 17 patients had metastases on 18F-OC PET/CT. 18F-OC demonstrated a higher ability to detect liver lesions (Fig. 2). In 11 patients with liver metastases, 100% (11/11) and 90.9% (10/11) of patients showed liver metastases on 18F-OC and 68Ga-DOTATATE scans, respectively. 18F-OC also detected peritoneal lesions more effectively than 68Ga-DOTATATE in 1 patient (No. 9).

In the lesion-based examination, 68Ga-DOTATATE and 18F-OC PET/CT detected 152 and 177 focal lesions, respectively (p = 0.54). A total of 149 tumor lesions (9 in the primary sites, 93 in the liver, 20 in the lymph node, 8 in the bone, and 19 in other sites) were concordantly detected on both 18F-OC and 68Ga-DOTATATE PET/CT scans. An additional 30 lesions were detected by one of the scans only (Table 2). Both 18F-OC and 68Ga-DOTATATE had lesions that could not be detected by another imaging agent (Fig. 3). 18F-OC detected 28 lesions (23 in the liver, 2 in the lymph node, and 3 in the peritoneum) not visualized with 68Ga-DOTATATE. 68Ga-DOTATATE identified 1 lymph node lesion and 1 retroperitoneal lesion not seen with 18F-OC. 18F-OC detected significantly more liver lesions (116 vs. 93, p < 0.01). There was a difference of 10 liver metastases detected by the 2 radiotracers in patient No. 14 (Fig. 2), which were confirmed as true lesions by follow-up CT and MR. Additionally, 18F-OC detected 3 peritoneal lesions in patient No. 9. Regarding lymph node lesions, both 18F-OC and 68Ga-DOTATATE detected 1 lesion that was not clearly detected by the other imaging agent. In addition, 18F-OC and 68Ga-DOTATATE PET/CT had comparable effectiveness in detecting primary tumors and bone metastases.

Fig. 2
figure 2

More liver lesions were detected by 18F-OC (d–f) than by 68Ga-DOTATATE (a–c). Maximum intensity projection image (a, f) shows more liver metastases present in a 50-year-old patient with a grade II primary pancreatic neuroendocrine tumor. Although transaxial fused PET/CT (b, d) and PET images (c, e) acquired with 68Ga-DOTATATE and 18F-OC show an equal number of liver lesions in this slice, the lower background level of 18F-OC uptake by normal liver (d–f) better delineates liver lesions by making the lesions appear to have more obvious uptake and sharper edges relative to background levels of 68Ga-DOTATATE uptake (a–c)

Full size image
Fig. 3
figure 3

Both 18F-OC and 68Ga-DOTATATE had lesions that could not be detected by another imaging agent. PET/CT images acquired with 68Ga-DOTATATE and 18F-OC for patient No. 19 (a, b) and No. 20 (c, d). The 18F-OC fusion image (b) found an increased uptake in a retroperitoneal lymph node, while the uptake of this lymph node was not significantly increased on 68Ga-DOTATATE (a) image. However, in another patient, the 18F-OC fusion image (d) shows that the uptake in one retroperitoneal lesion was not obvious, while the uptake of 68Ga-DOTATAE was significant (c)

Full size image
Fig. 4
figure 4

Bar chart representing the maximum standardized uptake value (SUVmax, a), tumor-to-liver ratio (TLR, b) and tumor-to-spleen ratio (TSR, c) of 18F-OC. TLR (b) and TSR (c) were calculated by dividing the SUVmax of tumor lesions by the patient-specific SUVmean of the liver and spleen, respectively. The TLRs for lesions in the liver and lymph nodes were significantly higher with 18F-OC (p = 0.02 and p = 0.02, respectively). All other SUVmax, TLR and TLR calculations did not show significant differences

Full size image

Lesion uptake analysis found that 18F-OC uptake was slightly higher than 68Ga-DOTATATE uptake in primary tumors and metastases, but there were no significant differences (primary tumor: 25.01 (16.52–38.58) versus 19.08 (16.37–34.28), p = 0.80; metastases: 18.24 (11.25–37.48) versus 17.13 (8.99–28.72), p = 0.33). However, some lesions had higher 18F-OC uptake and others lesions had higher 68Ga-DOTATATE uptake, even among lesions of the same patient. For example, on PET/CT for suspected retroperitoneal pheochromocytoma in patient No. 20, 3 lesions were seen in the neck region, with higher 68Ga-DOTATATE uptake than 18F-OC uptake in 2 of the neck lesions (SUVmax 56.59 vs. 53.11 and 27.19 vs. 15.13), but higher 18F-OC uptake in the other lesion (SUVmax 115.14 vs. 111.45) (Additional file 1: Fig. S1). Furthermore, in liver and lymph node lesions, the 18F-OC TLR was higher than that with 68Ga-DOTATATE (p = 0.02, Fig. 4). However, the 18F-OC TSR was not significantly higher than that of 68Ga-DOTATATE for primary tumor or metastases (Fig. 4).

In our study, we found that despite physiological uptake in the PU (mentioned above), three cases of nodules with abnormal density or signal in the PU on CT or MRI showed abnormal uptake in the PU on 68Ga-DOTATATE and 18F-OC PET (Additional file 2: Fig. S2). The SUVmax were as follows: 41.1, 86.7, 16.4, respectively in 68Ga-DOTATATE and 27.5, 94.3, 15.3, respectively in 18F-OC.

Discussion

Here, we prospectively assessed the performance of 18F-OC PET/CT relative to 68Ga-DOTATATE PET/CT in 20 NEN patients. The 18F-OC had a favorable biodistribution profile and was not inferior to 68Ga-DOTATATE in tumor uptake, TLR and TSR.

Our data showed that the 18F-OC distribution in organs was similar to that of 68Ga-DOTATATE. 18F-OC accumulation was very high in the spleen, which was similar to that of 68Ga-labeled DOTA-SSAs. Because both radiotracers were mainly excreted by the urinary system, higher uptake was seen in the kidneys. However, the overall uptake of 18F-OC in organs was lower than that of 68Ga-DOTATATE, especially in the liver, where the background 68Ga-DOTATATE uptake was 1.5 times greater than that of 18F-OC. We found that the salivary glands showed visible differences between the 2 radiotracers, which was consistent with past findings that 68Ga-DOTATATE uptake by salivary glands was four–sixfold higher than that of 18F-OC, mainly because of different radiotracer clearance times [14].

Because of high physiological uptake due to high SSTR2 expression in the PU and artifacts caused by respiratory movement, focal pancreatic lesions and lesions around the head of the pancreas may be obscured. Here, we found that both 18F-OC and 68Ga-DOTATATE had high uptake nodules in the PU, and other imaging examinations (CT or MRI) showed changes in the shape, signal or density of these nodules (Additional file 2: Fig. S2). Other imaging agents based on 68Ga-labeled radionuclides also demonstrated high sensitivity and specificity for detecting lesions (93.6% and 90%, respectively) in the PU [17]. We considered that results of a previous study [17] and ours could indicate SSA- PET, combined with morphological information (CT or MRI), especially if performed with enhanced CT or MRI, will improve the accuracy of lesions in the PU. But our number of cases was relatively small (n = 3). Thus, larger studies are needed to confirm these findings.

18F-OC and 68Ga-DOTATATE were highly sensitive in detecting lesions, and there were no differences in their overall diagnostic efficacy. Relative to 68Ga-DOTATATE, 18F-OC can detect lesions better (177 vs. 152), especially lesions in the liver (116 vs. 93), probably due to the lower background level of 18F-OC uptake in the liver. This finding is of great clinical significance, as it may affect treatment methods. For example, in patient No. 3, liver lesions were detected with 18F-OC but not 68Ga-DOTATATE. In patient No. 10, only one lesion was detected in the left lobe with 68Ga-DOTATATE, while 18F-OC detected another lesion in the right lobe of the liver, which was confirmed to be NEN metastases through pathology. These data are consistent with findings by Pauwels et al. [14] that 18F-OC detects more liver lesions.

Our data did not uncover differences between 68Ga-DOTATATE and 18F-OC in the detection of bone lesions (8 vs. 8). However, Pauwels et al. [14] found that 18F-OC detects more bone lesions. The differences between the 2 studies may be due to the small number of bone lesions in our study. Regarding lymph node lesions, both imaging radiotracers detected unique lesions. Additionally, 18F-OC detected 3 relatively small peritoneal metastases (diameter: < 5 mm) in patient No. 9, which were missed by 68Ga-DOTATATE. This is attributable to 18F being a typical short-distance positron emitter with better spatial resolution [18], which may be better suited for detecting small lesions. Thus, the capacity of 18F-OC to detect lesions is similar to that of 68Ga-DOTATATE, and 18F-OC may detect liver lesions more efficiently.

In this study, the SUVmax of 18F-OC was higher than that of 68Ga-DOTATATE, but the difference was not statistically significant. Interestingly, relative to 68Ga-DOTATATE, 18F-OC had a better target-to-background ratio. In this study, using liver and spleen for background comparisons, we found the 18F-OC TLRs for lesions in the liver and lymph node were significantly higher than those of 68Ga-DOTATATE, probably due to low liver background with 18F-OC. However, this finding differs from the results from Pauwels et al. [14] that the 18F-OC SUVmax for all lesions were significantly lower than those of 68Ga-DOTATATE, but there was no difference in TBR, which may be attributable to the small sample size. We also found that some patients exhibited higher 68Ga-DOTATATE uptake in lesions while others had higher 18F-OC uptake, and even within the same patient, some lesions had higher 68Ga-DOTATATE uptake while others had greater 18F-OC uptake; this is probably because of NEN heterogeneity [19]. The reason for the difference between the two radiotracers still needs further study.

Taken together, we found that 18F-OC had similar characteristics to 68Ga-DOTATATE in terms of physiological distribution, lesion detection, and lesion uptake. However, 18F-OC was relatively better in detecting liver lesions than 68Ga-DOTATATE. The two radiotracers had significantly difference TLRs, which is an important parameter for lesion detection.

Limitations

The most significant limitation of this study was the lack of pathological confirmation of most lesions, which was not performed due to the ethical implications of pathologically examining all patient lesions. Thus, all lesions found with 18F-OC and 68Ga-DOTATATE were confirmed using alternative imaging approaches such as 18F-FDG PET, CT, or MRI. In addition, due to the small size of the study group and the small number of patients with higher-grade NENs, we could not evaluate the correlation between uptake and tumor grade. Future studies will involve a larger sample size.

Conclusion

Overall, 18F-OC shows a favorable biodistribution, in which the uptake in various organs is similar to or even lower than that of 68Ga-DOTATATE. 18F-OC can detect liver lesions better than 68Ga-DOTATATE, with a better tumor-to-liver ratio. However, both 18F-OC and 68Ga-DOTATATE have similar detection rates for lesions in other organs. In general, 18F-OC has great potential as an alternative to 68Ga-DOTATATE in the absence of a 68Ge/68Ga generator. In the future, more patients are needed for comparisons between 68Ga-DOTATATE and 18F-OC to verify the value of 18F-OC in clinical applications.