Patients
For this retrospective, single-center study, patients with pathologically confirmed NET who underwent [18F]FDOPA PET scanning at the department of Nuclear Medicine and Molecular in the UMCG as part of their routine follow-up were considered. The scans were performed within a period of 5 years (4 years prior and up to one year after the transition date of the radiopharmaceutical production method). If more than two scans of the same patients were available, the two scans closest to the transition date (one with the old and one with the new production method) were selected for this analysis. Patients’ clinical records were checked for current medication use and surgical intervention. Patients were allowed to have had minor surgery (resection of the gallbladder or a unilateral adrenal gland resection) in between scans. Patients who received major surgery (tumor resections from the gastrointestinal tract or resection of one of the kidneys) between scans were excluded from further analysis. Also excluded from the analyses were patients who showed progressive disease between the scans, according to RECIST criteria on the diagnostic CT, showed more than 5 new lesions on the second PET scan, or showed more than 10 tumor lesions in total on the baseline PET scan. To control the potential influence of the course of the disease on the uptake of [18F]FDOPA, all remaining subjects were subdivided into three groups. The first group consisted of patients without visible lesions on both [18F]FDOPA PET/CT scans (no metabolically active disease; follow-up scans after an earlier detected and resected NET or subjects at risk for NET without any found) to allow for measurements of background organ uptake. The second group were patients with no more than 2 lesions visible on both scans and with stable disease, whereas the third group consisted of the patients with 3–10 lesions on both scans, with no more than 5 new lesions on the second PET and who did not fulfil the criteria of PD according to RECIST. In the latter groups, both uptake in background and tumor tissue was evaluated. This approach was chosen to minimize the influence of disease progression on the uptake of the radiopharmaceutical.
[18F]FDOPA production
Both [18F]FDOPA-L and H were produced via electrophilic fluorination of the trimethylstannyl precursor with [18F]F2. However, the used radionuclide method to produce [18F]F2 gas was different. [18F]FDOPA-H was produced from [18O]O2 via a double-shoot approach, according to a previously published protocol [5]. This new GMP compliant multi-dose production of [18F]FDOPA-H resulted in a higher activity yield and improved molar activity, resulting in a lower administered mass compared to [18F]FDOPA-L (Table 2). The conventional [18F]FDOPA-L was synthesized using [18F]F2 produced from 20Ne using deuterons [6]. This radionuclide production proves lower activity yield and lower molar activity, because more cold F2 gas is needed for optimal recovery of [18F]F2 gas from the target. The average activity yield of [18F]FDOPA-H was 4 ± 0.4 GBq with a molar activity of 107 (40–163) GBq/mmol. For [18F]FDOPA-L, the average activity yield was 0.5 ± 0.1 GBq with a molar activity of 8.23 (4–61) GBq/mmol.
PET imaging and analysis procedure
All [18F]FDOPA PET/CT scans were performed from the top of the skull through mid-thigh 60 ± 6 min after intravenous administration of a standard dose of approximately 200 MBq [18F]FDOPA on a Biograph mCT camera (Siemens Medical Systems, Knoxville, TN, USA). All patients fasted for 6 h and were allowed to continue medication. All subjects were pretreated with 2.0 mg/kg of carbidopa orally with a maximum of 150 mg. The acquisition was performed in seven bed positions of 2 min emission time for patients between 60 and 90 kg. Patients with a bodyweight less than 60 kg and more than 90 kg body weight were scanned with 1 min and 3 min per bed position, respectively. Raw data were reconstructed according to guideline-based standardized NEDPAS or EARL algorithms for SUV calculations [7]. The uptake of [18F]FDOPA in target and background lesions was measured using manually drawn spherical volumes of interests (VOIs) on EARL reconstructed PET images, using Siemens Syngo.via software version VB10. Tumor lesions were defined as visual uptake higher than the background (bloodpool) and not due to a known physiological uptake pattern of [18F]FDOPA. [18F]FDOPA uptake in tumor lesions was expressed as maximum SUV (SUVmax). SUVmean values were obtained for the following background organs: striatum, heart, thoracic aorta, liver, gallbladder, pancreas (divided in head, body and tail because of considerable physiological variation in these regions), kidneys, and adrenal glands. Scans were divided into two categories based on the use of a low molar activity ligand ([18F]FDOPA-L) or a high molar activity ligand ([18F]FDOPA-H).
Statistics
Group comparisons between [18F]FDOPA-L and [18F]FDOPA-H in background regions and tumor lesions were performed using Wilcoxon Signed-Rank tests (p < 0.05 considered significant for tumors, p < 0.004 in organs after correcting for multiple comparisons (0.05/12)). When comparing tumors, the same tumor was measured at baseline and follow-up. The effect size r was calculated for each comparison using Rosenthal’s formula. In the Wilcoxon Signed-Rank test, negative ranks indicate an increase in SUV from [18F]DOPA-L to [18F]FDOPA-H, whereas positive ranks indicate a decrease from [18F]FDOPA-H to [18F]FDOPA-L. Independent sample t tests were used for group comparisons in injected activity and mass, with p < 0.05 considered significant. For statistical analysis, SPSS was used (IBM Corp. Released 2015. IBM SPSS Statistics for Windows, Version 23.0. Armonk, NY: IBM Corp).