Patients and PET data analyses
Between January 2020 and September 2020, we performed 109 PET/CT examinations using 18F-JK-PSMA-7 in patients referred for increased PSA blood level. We analysed clinical and imaging aspects in all 54 patients with BCR, which were followed in the Urology Department of Erasme hospital with no missing data. The retrospective analysis of data acquired on the PET/CT Vereos has been approved by the Ethics Committee of the institution, which waived the obligation of written informed consent.
The examinations were analysed by three experienced nuclear medicine physicians in a randomized order. Examinations were positive if there was at least one lesion which was considered as overexpressing PSMA based on the visual analysis of the images, in accordance to published guidelines (Ceci et al. 2021). SUVmax were calculated in all lesions detected. To calculate TBR (tumour/background ratio), the SUVmean in the left gluteal muscle was used as background value. Effect of PSA level on lesion detection was tested using a PSA threshold of 0.3 ng/ml as in a previous study performed with 18F-JK-PSMA-7 (Dietlein et al. 2021).
Quantitative analysis was performed using Philips software IntelliSpace Portal (ISP, Koninklijke Philips N.V., The Netherlands, version 11.1).
All statistical analyses on PET data were realized with JASP software (University of Amsterdam, The Netherlands, version 0.14.1). Group analyses were performed using unpaired t-tests and the results were considered significant at a p value < 0.05.
18F-JK-PSMA-7 tracer preparation
18F-JK-PSMA-7 was produced on an All-in-one automated synthesizer (Trasis SA, Belgium) configured with an HPLC purification system, as described before (Simaeys et al. 2021). All chemical reagents were provided in commercially available reagent kits by Trasis SA. The automated radiosynthesis of 18F-JK-PSMA-7 consists in a two-step reaction followed by an HPLC purification and a reformulation.
Briefly, 18F-fluoride was produced through the 18O(p,n)18F reaction in a Cyclone 30 cyclotron from IBA, Belgium. The irradiated enriched water was collected and passed through a QMA Sep-Pak carbonate cartridge where 18F-fluoride was trapped and [18O]H2O collected for recycling. 18F-fluoride is then eluted to the reactor. After 18F-fluoride drying at 125 °C under a stream of nitrogen, the precursor was added to the reactor and heated for 5 min at 70 °C. The reaction mixture was cooled down and hydrolysis of the protecting groups was realized. The crude mixture was diluted in saline and purification was carried out in a semi-preparative HPLC column. The resulting fraction was collected and diluted in saline and trapped for reformulation. 18F-JK-PSMA-7 was eluted with ethanol and formulated in a solution of sodium ascorbate in saline. Quality control of the final product was performed on each batch and complied with in-house specifications and with the principles of the general monograph on radiopharmaceutical preparations (EP 07/2016:0125). Quality controls included the assessment of appearance, pH, chemical purity, radiochemical purity and radionuclidic purity, residual solvents, bacterial endotoxins, and sterility.
18F-JK-PSMA-7 imaging
The PET/CT scans were all performed on a Philips Vereos digital PET/CT (Philips Medical Systems, Cleveland, Ohio, USA), the characteristics of which having been previously described (Zhang et al. 2018). PET acquisitions were performed 140 min post-injection based on data showing a 50% TBR increase between 100 and 140 min after tracer administration (Hohberg et al. 2019). The median injected dose of 18F-JK-PSMA-7 was 4 MBq/kg (range: 226-507 MBq). The injected dose was adapted from the dose applied for analog acquisition systems in our centre, corresponding to a reduction by 20% of the dose we have adopted for patients investigated on the Philips Gemini TF64 PET/CT (Philips Medical Systems, Cleveland, Ohio, USA). Our experience with PSMA imaging with an analog PET/CT system started in December 2017. A total of 674 examinations were performed in our centre between December 2017 and December 2019.
After the injection of the radiotracer all patients received intravenous furosemide in order to force diuresis.
A whole-body scan was performed in all patients, from skull vertex to feet. PET acquisitions were performed for approximately 14 min including a total of 17 bed positions (1 min per bed position for the 10 positions from skull to thighs and 30 s per bed position for the 7 positions from thighs to feet, adapted to the patient’s morphology). This mode of acquisition corresponds to a 50% reduction of imaging time in comparison to our protocol of acquisition on the Philips Gemini TF-64 PET/CT. PET acquisitions were combined with a low-dose non-contrast-enhanced CT (50 mAs, 120 kV). All images were corrected for radioactivity decay as well as for scatter and random coincidences; 511 keV photon attenuation correction was performed using CT images.
Images were reconstructed with the 3D ordered subset expectation maximization (OSEM) algorithm implemented on the Philips Vereos system and set up to 3 iterations, 15 subsets and a Gaussian 3D filter with a full width at half-maximum of 6 mm.