The combination of 13N-ammonia and 18F-FDG whole-body PET/CT on the same day for diagnosis of advanced prostate cancer.

PURPOSE: The aim of the study was to evaluate the efficacy of 13N-ammonia and 18F-fluorodeoxyglucose (18F-FDG) PET performed on the same day in the detection of advanced prostate cancer (PC) and its metastases. PATIENTS AND METHODS: Twenty-six patients with high-risk PC [Gleason score 8-10 or prostate-specific antigen (PSA)> 20 ng/ml or clinical tumor extension ≥ T2c] were recruited into the study. 13N-Ammonia and 18F-FDG PET/CT were performed on the same day (18F-FDG followed ammonia, with an interval of a minimum of 2 h). Lesions were interpreted as positive, negative, or equivocal. Patient-based and field-based performance characteristics for both imaging techniques were reported.
RESULTS: There was significant correlation between 13N-ammonia and 18F-FDG PET/CT in the detection of primary PC (κ=0.425, P=0.001) and no significant difference in sensitivity (60.2 vs. 54.5%) and specificity (100 vs. 83.3%). The maximum standard uptake values and corresponding target-to-background ratio values of the concordantly positive lesions in prostate glands in the two studies did not differ significantly (P=0.124 and 0.075, respectively). The sensitivity and specificity of PET imaging using 13N-ammonia for lymph node metastases were 77.5 and 96.3%, respectively, whereas the values were 75 and 44.4% using 18F-FDG. The two modalities were highly correlated with respect to the detection of lymph nodes and bone metastases.
CONCLUSION: The concordance between the two imaging modalities suggests a clinical impact of 13N-ammonia PET/CT in advanced PC patients as well as of 18F-FDG. 13N-Ammonia is a useful PET tracer and a complement to 18F-FDG for detecting primary focus and distant metastases in PC. The combination of these two tracers on the same day can accurately detect advanced PC.

Introduction

Prostate cancer (PC) is the most common malignancy among men in the USA, and its incidence has shown a growing trend worldwide 1. It is clinically a heterogeneous disease characterized by an overall long natural history in comparison with other solid tumors, showing a wide spectrum of biological behavior ranging from indolent to aggressive 2. Currently, PC is classified into three risk groups (low, intermediate, and high) on the basis of serum PSA level, Gleason score, and clinical stage. The treatment strategy and prognosis of PC is related to its risk level. Proper staging of PC is particularly important in high-risk primary disease before embarking on radical prostatectomy or radiation therapy 3. Therefore, it is important to accurately detect tumor and estimate tumor extent in PC.

Conventional imaging studies such as transrectal ultrasound, computed tomography (CT), and MRI are currently used for the diagnosis of PC. However, they are not completely adequate for this remarkably heterogeneous disease 4,5. In recent times, the use of PET/CT imaging in oncology has opened up a new role for molecular imaging in PC. As the most commonly used PET tracer, 18F-fluorodeoxyglucose (18F-FDG) has been regarded as limited in the diagnosis of PC because of 18F-FDG excretion into the urinary bladder, relatively little glucose metabolism in some PC cases, and high 18F-FDG accumulation in inflammatory tissues or benign prostatic hypertrophy 6,7. However, 18F-FDG PET/CT has been shown to have a relatively high sensitivity for detecting advanced PC lesions 8–11.

13N-Ammonia is a useful 13N-labeled PET imaging agent for assessing regional blood flow in tissues 12. Nevertheless, as one of the principal products of nitrogen metabolism, 13N-ammonia plays a significant role in glutamine synthesis 13 and has been used to detect some types of tumors 14. Our recent studies have found that 13N-ammonia can be taken up by some brain tumors and PC cells, which may be associated with its involvement in glutamine synthesis 15–17. Although the utility of 13N-ammonia PET/CT in the imaging of PC has been studied 17, the capability of 13N-ammonia whole-body PET/CT in advanced PC and the combination of 13N-ammonia and 18F-FDG PET/CT for the diagnosis of advanced PC have not been reported.

In this study, we attempt to determine the value of 13N-ammonia in comparison with 18F-FDG PET/CT for detecting local tumors, lymph nodes (LNs), and bone metastases in advanced PC patients.

Patients and methods

Patients

Twenty-six patients (mean age: 72.2 years, range: 60–88 years) with high-risk PC were recruited from our PET center between August 2010 and November 2014. The inclusion criteria were as follows: (i) patients had to have a Gleason score 8–10 or PSA more than 20 ng/ml or clinical tumor extension more than or equal to T2c; (ii) 13N-ammonia and 18F-FDG imaging studies should have been performed on the same day; (iii) biopsy and pathological results were available. Patients with a history of a second cancer and those without pathology results were excluded. Detailed information of all patients is presented in Table 1.

Table 1

Patient characteristics

The study was approved by the local ethics committee. All patients gave their written informed consent after receiving a detailed explanation of the study purpose and imaging procedure.

PET imaging

PET/CT imaging was performed using a Gemini GXL 16 scanner (Philips, Amsterdam, the Netherlands). All patients fasted for at least 8 h and urinated just before starting the PET/CT scan. Ammonia and 18F-FDG PET/CT studies were performed on the same day (18F-FDG followed ammonia, with an interval of a minimum of 2 h). The PET images were obtained from the top of the skull to the mid-thighs for 1.5 min/bed position in two-dimensional mode, reconstructed by the line of response algorithm and attenuation-corrected using CT. The scan protocol for CT was as follows: peak kilovoltage 140 kV, 180 mA/slice, thickness 5 mm, and rotation time 0.5 s.

Ten minutes after an intravenous injection of 13N-ammonia (555–740 MBq) or 45–60 min after an intravenous injection of 18F-FDG (5.18 MBq/kg), the PET/CT acquisition started. Images were interpreted using a Gemini workstation (Philips).

Image and data analysis

CT images without contrast enhancement were consistently available and allowed the identification of LNs and distant unrelated findings 18. CT images (from PET/CT) were assessed by an experienced radiologist blinded to all other data. The malignant lesions were divided into LN metastasis and sclerotic and osteolytic lesions according to morphological criteria (size, shape, and regional grouping). The interpretation of PET/CT was made as a consensus reading of two nuclear medicine physicians. Each site of abnormally increased 13N-ammonia and 18F-FDG uptake on PET images was interpreted as positive, negative, or equivocal. For PET images, hyperactivity above background was considered a positive lesion. Negative ammonia and 18F-FDG PET scans were those that did not show any activity or showed activity that was apparently lower than background. Equivocal was defined as any lesion with an activity between the two categories ‘positive’ and ‘negative’.

Patients were monitored for at least 4 months (median: 11 months, range: 4–20 months). The lesions were classified as ‘true positive’ if they were positive on 13N-ammonia and/or 18F-FDG PET/CT and finally confirmed by CT and/or by clinical data (i.e. biopsy). Ammonia and 18F-FDG PET lesions that primarily appeared to be benign and also benign on CT were considered as ‘true negative’ for metastases. The positive lesions on 13N-ammonia or 18F-FDG PET/CT but which were finally proved to be benign were considered as ‘false positive’. The malignant lesions clearly confirmed by CT but which were negative on 13N-ammonia or 18F-FDG PET were considered ‘false negative’.

The uptake of the lesion was evaluated by semiquantitative analysis using the maximum standard uptake values (SUVmax). To determine the SUVmax, regions of interest (ROIs) were drawn over the abnormal lesions in 13N-ammonia or 18F-FDG PET images. Thereafter, a reference ROI in iliac fossa fat was chosen as a reference background ROI in both imaging modalities. Finally, the SUVmax of all ROIs was used for the calculation of target-to-background ratios (TBRs).

Statistical analysis

Patient-based and lesion-based analyses were performed. Data were defined as mean±SD and were compared in different groups using the independent t-test. Sensitivity and specificity were calculated using data collected from PET studies. The Mann–Whitney U-test was used to compare quantitative variables in a paired group. The κ coefficient was calculated for comparison of two imaging modalities. Statistical analysis was performed with SPSS software (SPSS Inc., Chicago, Illinois, USA). A P value less than 0.05 was considered statistically significant.

Results

Of 26 patients, 12 (46%) had bone metastases, four (15%) had LN metastases, eight (31%) had both metastases, and two (8%) had no metastasis according to medical examinations (CT, PET/CT, biopsy, etc.) and clinical follow-up. In total, 218 bone lesions were assessed in 15 patients with a mean number of 15 bone lesions per patient (median: 11, range: 1–27). In those bone lesions, 115 were osteogenic, 98 were osteolytic, and five were hyperostosis (the benign lesions are not shown in Table 1). Five of the patients with positive PET/CT had extensive spread with countless bone metastases and therefore could not be included in the analysis. Meanwhile, 67 LNs were assessed in 18 patients, of which 40 LNs were metastases and 27 LNs were reactive lymphaden proliferation (the benign lesions are not shown in Table 1). Table 2 summarizes all PET results according to concordance for 13N-ammonia and 18F-FDG PET studies, and Fig. 1 shows the distribution of SUVmax and TBR values among lesions.

Table 2

PET results (the numbers of double-positive, double-equivocal, double-negative and discordant for each patients) of ammonia and 18F-FDG PET/CT

Fig. 1

The distribution of maximum standard uptake values (SUVmax) and target-to-background ratio (TBR) values among lesions.

Primary tumor

Patient-based analysis

Twenty-six patients with primary tumor were detected correctly on ammonia PET imaging, but one patient was negative on 18F-FDG PET imaging. Therefore, the sensitivity of ammonia was 100% and that of 18F-FDG was 96.2%.

Lesion-based analysis

A total of 106 segments of prostate glands in 14 patients were analyzed. Pathology evaluation showed that 88 segments were malignant. Ammonia PET was able to identify 53 positive segments correctly, whereas 48 segments were positive on 18F-FDG PET. According to the results of pathology, there were 18 true-negative results for ammonia PET and 15 true-negative results for 18F-FDG PET. Hence, the sensitivity and specificity of ammonia were 60.2 and 100%, respectively, and those for 18F-FDG were 54.5 and 83.3%. Figure 2 shows the 13N-ammonia and 18F-FDG images in one patient with primary PC and prostatitis, which were not consistent.

Fig. 2

A 75-year-old man (case 5), clinical tumor stage T2, Gleason score 3+4, PSA 68.13 ng/ml. HE staining images (a) and PET/CT images (b) of 13N-ammonia (upper row) and 18F-fluorodeoxyglucose (18F-FDG) (lower row) for prostatitis (red arrows) and primary prostate cancer (PC) tissues (yellow arrows). 13N-Ammonia PET showed increased tracer uptake for primary PC tissues and negative for tissues prostatitis; however, 18F-FDG PET were both positive. HE, hematoxylin & eosin.

The SUVmax of the concordantly positive lesions in prostate glands on ammonia and 18F-FDG PET/CT studies was 3.16±1.77 and 3.82±2.31, respectively. The values were not different between the two studies (P=0.124). The TBR values also did not differ significantly between the two studies (6.06±2.74 and 7.67±3.98 on ammonia and 18F-FDG PET/CT studies, respectively; P=0.075). In addition, moderate agreement was found between ammonia and 18F-FDG PET/CT for the detection of primary PC (κ=0.425, P=0.001).

Lymph node metastases

The sensitivity and specificity of PET/CT in the detection of LN metastases in PC were 83.3 and 92.9% for ammonia and 83.3 and 64.3% for 18F-FDG, respectively.

In total, 67 LNs were detected in 18 patients by PET and CT, of which 40 LNs were proven as malignant by PET/CT and clinical follow-up. Thirty-one LNs identified by ammonia PET were proven as true positive, compared with 30 LNs identified by 18F-FDG PET; thus, the sensitivity and specificity of ammonia and 18F-FDG were 77.5 and 96.3%, and 75 and 44.4%, respectively. Compared with 18F-FDG, 13N-ammonia PET/CT showed similar sensitivity but superior specificity. The SUVmax of positive LNs assessed by ammonia and 18F-FDG was 4.50±4.67 and 3.45±2.04, respectively, whereas the TBR for ammonia and 18F-FDG was 6.96±3.39 and 7.39±4.06, respectively. No significant difference was noted in SUVmax (P=0.419) or TBR (P=0.775) of the two imaging modalities for detecting LN metastases. However, fair agreement was found between these two imaging modalities (κ=0.326, P=0.003).

Bone metastases

Both PET/CTs were similarly positive for bone metastases in 26 patients. Only one hyperostosis patient had a positive 18F-FDG PET/CT scan, whereas the corresponding ammonia PET/CT scan was negative. The sensitivity and specificity of PET/CT for the detection of bone metastases in PC were both 100% for ammonia and 100 and 83.3% for 18F-FDG, respectively.

A total of 218 bone lesions were studied, which were divided into 115 osteogenic, 98 osteolytic, and five hyperostosis. Out of 115 osteogenic lesions, 105 (91.3%) were detected positively by 18F-FDG PET/CT, compared with 112 (97.4%) by ammonia PET/CT. The SUVmax of positive osteogenic lesions assessed by ammonia and 18F-FDG was 4.24±2.56 and 5.17±3.07 (P=0.294), respectively, whereas the TBR for ammonia and 18F-FDG were 7.54±2.85 and 9.35±4.48 (P=0.034), respectively. For 98 osteolytic lesions, both ammonia and 18F-FDG were 100% positive. The SUVmax for ammonia and 18F-FDG was 3.89±3.16 and 4.19±1.98 (P=0.548), respectively, and the TBR was 6.96±1.34 and 8.57±3.01 (P=0.049), respectively. Of five hyperostosis lesions, one was positively identified (SUVmax: 3.15) by ammonia PET/CT and two (SUVmax: 2.45, 3.06) by 18F-FDG PET/CT. Furthermore, a relatively close agreement was found between ammonia and 18F-FDG PET/CT for the detection of metastatic bone disease in PC patients (κ=0.589, P<0.001).

Discussion

Molecular imaging has been adopted in recent PC studies as it is a noninvasive diagnostic modality that allows accurate management of the disease 5–10. PET/CT as one of the most important molecular imaging modalities is fundamentally suited for the imaging evaluation of biologic targets and events 5, which has been widely applied in clinical practice for examination of PC biology. As the most common PET tracer, 18F-FDG may enhance the staging of advanced PC 6–9. Generally, high uptake of 18F-FDG is expected in prostate tumors that are poorly differentiated, are hypoxic, and have a high Gleason score 7. Many different PET radiotracers are likely to be suited to various clinical states of PC, such as 11C-methionine, 11C-choline or 18F-choline, and 11C-acetate 10,18–20. Furthermore, we found that 13N-ammonia uptake in PC segments is significantly higher than that in benign segments 17, which was related to the expression of glutamine synthetase (GS) in PC. However, the previous research only studied the uptake of 13N-ammonia in the pelvis. To our knowledge, the present study provided a systematic comparison of 13N-ammonia and 18F-FDG whole-body PET/CT for advanced PC.

The biologic mechanisms for the accumulation of ammonia are not yet clear, but there is a reasonable explanation for its use in PC. Generally, glutamine is conditionally essential in cancer cells, being utilized as an alternative fuel source to glucose for the tricarboxylic acid cycle, and as a source of fatty acid production through reductive carboxylation 21–23. Some authors have reported that abnormal glutamine metabolism has been found in PC 24,25. According to Cooper 13, ammonia can act as a source of glutamine in the glutamine cycle. Nevertheless, our recent studies found that 13N-ammonia could also be obviously taken up by PCs; the mechanism might be associated with upregulation of de-novo glutamine synthesis in tumors 17. Therefore, ammonia might play a significant role in glutamine synthesis.

At the time of 18F-FDG PET, nearly 3 h had passed since the intravenous injection of 13N-ammonia. Meanwhile, because of the short half-life of 13N and urine excretion, 13N-ammonia had already cleared from the body. Therefore, it would have had no effect on the two PET studies, and it enabled the study of 13N-ammonia and 18F-FDG PET on the same day. Ultimately, it was beneficial for the patients, as they did not have to travel to our PET center twice.

In this study, ammonia PET revealed all pathological lesions in the prostate in 26 patients, whereas 18F-FDG PET failed in one patient. However, ammonia PET gave 53 true-positive results in 106 segments and 18F-FDG PET gave 48 true-positive results. Ammonia PET had a sensitivity and specificity of 60.2 and 100%, respectively, for detecting primary PC. The sensitivity and specificity of 18F-FDG PET were 54.5 and 83.3%, respectively. Shiiba et al. 10 reported that the sensitivity and specificity of 11C-methionine for distinguishing between patients with no Gleason score and those with low-to-high Gleason scores were 78.7 and 75.6%, respectively. The sensitivity was not high for ammonia and 18F-FDG for the diagnosis of primary PC. It has been reported that the usefulness of 18F-FDG PET for locally prostatic neoplasms and pelvic LN metastases is limited because of bladder urine activity 7,8. 13N-Ammonia was cleared from the body primarily through the renal system 26, which also affected its interpretation of the pelvic lesions.

However, 13N-ammonia could be more effective than 18F-FDG in distinguishing PC from prostatitis. In a 75-year-old patient with both prostatitis and PC, both positive 18F-FDG could be found in those segments, whereas prostatitis showed an absence or lower uptake of 13N-ammonia (Fig. 2). One possible reason is that GS is more active in PC than in prostatitis. The combination of ammonia and 18F-FDG could be helpful for the detection of PC.

When assessing LN metastasis, 13N-ammonia PET showed 77.5% sensitivity versus 75% sensitivity for 18F-FDG PET, but it exhibited 96.3% specificity compared with 44.4% for 18F-FDG PET. (Figure 3 shows the hilar LN inflammation detected by both modalities, and Fig. 4 shows LNs and bone metastases.) Likewise, 13N-ammonia PET had higher specificity than 18F-FDG PET in detecting primary PC (100 vs. 83.3%). These results were consistent with our previous studies 17. GS could be inactivated by reactive oxygen species in the macrophages 27. Therefore, an absence or lower uptake of 13N-ammonia was seen in the setting of inflammation and infection.

Fig. 3

Case 24, transaxial images of 13N-ammonia PET/CT (a–c) and 18F-fluorodeoxyglucose (18F-FDG) PET/CT (d–f) for hilar lymph node inflammation. 13N-Ammonia was negative while 18F-FDG PET showed increased uptake (red arrows).

Fig. 4

A 64-year-old prostate cancer patient (case 20) with lymph nodes and bone metastases: (a) whole-body anterior maximal intensity projection images of 13N-ammonia PET/CT (left column) and 18F-fluorodeoxyglucose (18F-FDG) PET/CT (right column); (b) transaxial images of ammonia PET/CT (upper row) and 18F-FDG PET/CT (lower row) for pelvic lymph node metastasis, ammonia showed increased tracer uptake while 18F-FDG was negative (red arrows); (c) transaxial images of ammonia PET/CT (upper row) and 18F-FDG PET/CT (lower row) for osteogenic metastases, they both showed high uptake.

In the patient-based analysis, both imaging modalities were able to detect bone metastases. (Figure 5 shows a 60-year-old PC patient with bone metastases.) Meanwhile, the results were similar in the lesion-based analysis. Both for osteogenic and osteolytic lesions, ammonia and 18F-FDG had similar positive rates (97.4 vs. 91.3% for osteogenic lesions; both 100% for osteolytic lesions). Some other PET tracers also suggested for the assessment of PC with bone metastases include 18F-fluoride and 18F-fluorocholine 20,28, but they were not used to make a distinction between the osteogenic and osteolytic lesions. We found that 13N-ammonia and 18F-FDG are more sensitive in osteolytic lesions than in osteogenic lesions, and 13N-ammonia is more sensitive than 18F-FDG in osteogenic lesions, although there was no significant difference between the two methods. Nevertheless, whether in osteogenic lesions or in osteolytic lesions, the TBR values for ammonia were lower than those for 18F-FDG. This might indicate that the radiation-absorbed doses in tumors for 13N-ammonia are less than those of 18F-FDG. This might be associated with abnormal glutamine metabolism after bone destruction. However, the mechanism about the usefulness of 13N-ammonia in bone metastases is still not yet fully understood and further study needs to be done.

Fig. 5

A 60-year-old prostate cancer patient (case 1) with bone metastases: (a) whole-body anterior maximal intensity projection images of 13N-ammonia PET/CT (left column) and 18F-fluorodeoxyglucose (18F-FDG) PET/CT (right column); (b) transaxial images of ammonia PET/CT (upper row) and 18F-FDG PET/CT (lower row) for osteolytic metastasis, although they were both positive (red arrows), ammonia had more accumulation than 18F-FDG.

The current results demonstrated close agreement between 18F-FDG and ammonia PET/CT. They showed us that 13N-ammonia and 18F-FDG have similar uptake in PC cells either on glutamine metabolism or on glucose metabolism. It is suggested that 13N-ammonia is a potential PET tracer for detecting distant metastases in PC and is a complement to 18F-FDG for detecting advanced primary PC. However, the diagnosis of primary PC using PET imaging remains a dilemma that warrants further research.

Our study has several limitations. First of all, clinical follow-up instead of histopathology was used in patients as reference for the patient’s LN and bone status. In addition, we had collected a small sample of advanced PC cases, and these results need further validation by prospective studies with larger sample size. Further studies are needed to confirm the clinical utility of 13N-ammonia imaging in PC.

Conclusion

The data obtained in this preliminary investigation suggest a clinical impact of 13N-ammonia PET/CT in advanced PC patients as well as of 18F-FDG. 13N-Ammonia is a useful PET tracer for detecting distant metastases in PC, and is a complement to 18F-FDG for detecting advanced primary PC. The combination of these two tracers on the same day can accurately detect advanced PC.