[⁸⁹Zr]Zr-PSMA-617 PET/CT in biochemical recurrence of prostate cancer: first clinical experience from a pilot study including biodistribution and dose estimates

[⁸⁹Zr]Zr-PSMA-617 PET/CT imaging was performed in n = 7 consecutive patients due to BCR of prostate cancer who revealed no or undetermined findings on [⁶⁸Ga]Ga-PSMA-11 PET/CT. In 4 patients, [⁶⁸Ga]Ga-PSMA-11 did not reveal suspicious findings (negative [⁶⁸Ga]Ga-PSMA-11 PET/CT). Three patients had undetermined findings on [⁶⁸Ga]Ga-PSMA-11 PET/CT with no definite assignment to pathological or physiological uptake, e.g., faintly accumulating or unusually located tracer uptake. The time interval between both, [⁸⁹Zr]Zr-PSMA-617 and [⁶⁸Ga]Ga-PSMA-11 PET/CT, was 31 ± 18 days (range 5–49 days), with no treatment performed in between. Prostate-specific antigen (PSA) serum level at time of imaging ranged from 0.43 to 1.92 ng/ml. All patients were initially treated with radical prostatectomy (RP). Initial Gleason score ranged from 7a to 9. Four patients underwent salvage treatments such as androgen-deprivation therapy (ADT), lymphadenectomy (LA) or radiation therapy (RT). Detailed patient characteristics including age, Gleason score, primary therapy, salvage therapies, time from initial diagnosis (ID) of prostate cancer, PSA and PSA doubling time (DT) are presented in Table 1. [⁸⁹Zr]Zr-PSMA-617 PET/CT imaging was performed on a compassionate use basis under the German Pharmaceutical Act §13 (2b). The medical indication for the examination and the labeling of the tracer were under the direct responsibility of the applying physician. Patients gave their written consent after being thoroughly informed about the general risks of both radiation exposure and application of a novel PET tracer including possible adverse effects of, e.g., therapeutic PSMA tracers. Furthermore, all patients agreed to the publication of the resulting data in accordance with the Declaration of Helsinki.

The radiolabeling of PSMA-617 with ⁸⁹Zr was based on our previously published procedure with further optimizations [23]. Briefly, [⁸⁹Zr]Zr-oxalate (PerkinElmer, Groningen, The Netherlands) was transformed into [⁸⁹Zr]ZrCl₄ using a QMA cartridge (Waters, Milford, USA), which was activated by each 10 mL of acetonitrile, saline, 1 M hydrochloric acid, and deionized water. [⁸⁹Zr]Zr-oxalate was then loaded onto the cartridge followed by a washing step with 60 mL of deionized water. The activity was recovered from the QMA cartridge by fractionated elution with two fractions of 700 µL and 800 µL of 0.1 M hydrochloric acid. The latter fraction contained ~ 95% of the initial activity and was used for radiolabeling. To this fraction, 1.5 µg (1.44 nmol) of PSMA-617 per MBq [⁸⁹Zr]ZrCl₄ in 50 µL water were added followed by the addition of 800 µL MES buffer (0.5 M, pH 5.5). The reaction mixture was then heated at 95 °C for 30 min. After cooling to room temperature, the mixture was loaded onto a C₁₈ Sep Pak cartridge (Waters, Milford, USA), which was pre-equilibrated with each 10 mL of ethanol and water for injection. The cartridge was washed with 5 mL of water for injection and the activity eluted with 2 mL 50% ethanol (v/v) and 8 mL of saline. The product was finally passed through a 0.22-µm filter for sterilization. This procedure is reliable for up to 800 MBq of initial ⁸⁹Zr activity providing the final product [⁸⁹Zr]Zr-PSMA-617 in radiochemical yields of 70 ± 5% and radiochemical purities of > 98%. Quality control of the final product was performed according to current Good Manufacturing Practice (cGMP) guidelines checking for pH, clarity and color, radiochemical purity (HPLC and iTLC), chemical purity (HPLC), radionuclidic purity, endotoxin content, and sterility.

Each patient underwent PET/CT imaging scans at 4 time points: 1 h, 24 h, 48 h, and 72 h post intravenous injection (p.i.) of 111 ± 11 MBq (mean ± standard deviation, range 97–129 MBq) [⁸⁹Zr]Zr-PSMA-617. All PET/CT imaging was performed on a Biograph mCT 40 scanner (Siemens Medical Solutions, Knoxville, TN, USA) comprising whole-body PET/CT scans extending from vertex to mid-femur in 3D mode. PET acquisition time duration was 4 min per bed position on the initial day of injection and extended up to 10 min on the last day. CT data were acquired in low-dose technique using an X-ray tube voltage of 120 keV and a modulation of the tube current (CARE Dose4D, Siemens Erlangen; maximal tube current 30 mA) followed by reconstruction with a soft tissue reconstruction kernel (B31f) to a slice thickness of 5 mm (increment 2–4 mm). PET emission data was corrected for decay, randoms, and scatter. PET image reconstruction was performed applying an iterative 3-dimensional ordered-subset expectation maximization (OSEM) algorithm (3 iterations; 24 subsets) with gaussian filtering to a transaxial resolution of 5 mm at full width half maximum (FWHM). The matrix and the pixel size were 200 × 200 and 3.0 mm, respectively. Attenuation correction was performed using the low-dose CT data.

Any adverse event was recorded during and after examination. Vital parameters as heart rate, blood pressure, body temperature, and oxygen saturation were closely monitored. Within a time interval of 4 weeks after [⁸⁹Zr]Zr-PSMA-617 PET/CT, patients were additionally interviewed about potential side effects.

The biodistribution of [⁸⁹Zr]Zr-PSMA-617 was quantified by analyzing the standard uptake values (SUV) SUV_max, SUV_peak and SUV_mean at 1 h, 24 h, 48 h, and 72 h p.i.. Considering normal-organs, elliptical volumes of interest (VOI) were manually drawn enclosing regions of relatively homogenous uptake. Activity estimation was performed within the VOI applying a 40% or, in case of faintly accumulating organs, a 20% threshold using the SyngoVia software (Enterprise VB 60, Siemens, Erlangen, Germany). The following organs were included in this evaluation: the brain, salivary, and lacrimal glands, nasal mucosa, lung, liver, spleen, small and large intestine, and the kidneys. With regard to the large intestine, SUV was determined at the descending colon. Blood pool and background were evaluated in the descending aorta and the gluteal muscle, respectively. Tissue-to-background ratios (TiBR) were calculated by dividing the SUV_max of the organs by the SUV_mean of the background (gluteal muscle).

Three physicians with long-time experience in PSMA-targeted PET/CT (S.E., F.K., and F.R.) visually identified suspicious lesions taking into account all four imaging time points (1 h, 24 h, 48 h, and 72 h p.i.). Lesions that were visually considered as suggestive for prostate cancer were analyzed by measuring SUV_max and by calculating tumor-to-background ratios (TBR), defined as SUV_max divided by SUV_mean, of the background (gluteal muscle).

Mean absorbed radiation doses were estimated using the QDOSE program package (ABX-CRO, Dresden, Germany) considering the following organs as source organs: kidneys, liver, spleen, salivary glands (parotid gland and submandibular gland), and lacrimal gland. As a first step, volumetric co-registration of the different time points was performed by taking the first CT scan as reference. PET images, which were coupled to the CT images of the respective imaging session, were transformed according to the CT transformation matrix. Boundary VOIs which solely enclosed the source organs without interfering with the activity concentration of neighboring structures were manually drawn in the PET image that allowed the best organ delineation and then copied onto all other time-point scans. Manual adjustment of the boundary VOI for each time point was done when necessary, using the respective CT. Volume and activity estimation were performed within the boundary applying a fuzzy locally adaptive Bayesian (FLAB) segmentation algorithm for automatic volume delineation [24]. The next step comprised mono-exponential regression of the serial measured activities using weighted least squares method and estimation of both, the time integrated activities (TIA) and the time-integrated activity coefficients (TIAC) in the source regions. As an approximation, a linear increase from t = 0 to the first acquisition time point was assumed, and the integration method for the first time interval was based on the trapezoidal method. Between the first and the last time point, a trapezoidal integration was used, whereas the TIA from the last time-point to infinity was calculated analytically by applying the mono-exponential fitting curve. The respective TIACs (often termed residence time) of the source organs were calculated by normalizing the TIA to the amount of activity administered. The total body TIA was calculated approximately using the total FOV volume of interest and the injected activity. As a further approximation, constant activity between t = 0 and the first acquisition time point was assumed, and the integration method for the first-time interval was based on the trapezoidal method. Between the first time point and infinity, the TIA was calculated analytically by using the mono-exponential fitting curve. Estimations of absorbed organ dose and effective dose were performed by the IDAC 2.1 software which is implemented in QDOSE [25‐27]. The IDAC reference man was applied for the kidneys, the spleen, the liver, the heart and the intestine, and the sphere model for the salivary and lacrimal glands. Patient-specific organ mass adjustment, which is available in QDOSE, was applied for the kidneys, the liver and the spleen. Respective organ masses were determined using the volume of each organ delineated from the CT images (PACS software, SECTRA, Linköping, Sweden) and the respective biological tissue density. Organ masses for the salivary glands were taken from International Commission on Radiological Protection (ICRP) publication 89 with 25 g estimated weight for the parotid and 12.5 g for the submandibular gland [28].

The examination with [⁸⁹Zr]Zr-PSMA-617 was not associated with any side effects in all 7 patients. No acute adverse events or other drug-related pharmacologic effects occurred. Monitored vital parameters remained unchanged. No patient complained of any subjective symptoms during examination and follow-up.

Intense physiological uptake was observed in the salivary and lacrimal glands, liver, spleen, kidneys, intestine, and urinary tract. Figure 1 demonstrates a representative example of [⁸⁹Zr]Zr-PSMA-617 PET/CT at the predefined time points 1 h, 24 h, 48 h, and 72 h p.i., including the respective biodistribution data of selected organs assessed by SUV kinetics.

Detailed descriptive statistics of tracer distribution of all patients, including SUV_max, SUV_peak, SUV_mean and TiBR (tissue-to-background-ratio) of various organs, are presented in Fig. 2. The highest average SUV_max at 1 h p.i. was observed for the kidneys with 21.15 ± 12.31, consequently decreasing to 7.43 ± 1.56 at 72 h p.i.. For salivary glands and lacrimal glands, the maximum SUV_max was at 24 h p.i., with the highest value for submandibular gland (17.30 ± 4.69 at 24 h p.i., decreasing to 6.29 ± 1.61 at 72 h p.i.). Only colonic tracer uptake increased continuously (from 3.50 ± 2.21 at 1 h p.i. to 13.57 ± 5.00 at 72 h p.i.). TiBR increased between 1 and 72 h p.i. for the salivary glands, nasal mucosa, liver, colon, and kidneys. The highest TiBR values (≥ 50) were observed in salivary glands, kidneys, and colon at 72 h p.i.

The monoexponential curve-fitting parameters, the time-integrated activity coefficients (TIAC) for each source organ, and the respective estimated absorbed doses according to IDAC 2.1 are summarized in Table 2. Among the normal tissues, the parotid gland received the highest absorbed dose of [⁸⁹Zr]Zr-PSMA-617 with 0.601 ± 0.185 mGy/MBq followed by the kidneys with 0.517 ± 0.125 mGy/MBq and the submandibular gland with 0.468 ± 0.136 mGy/MBq. These values resulted in an average effective dose of 0.0913 ± 0.0118 mSv/MBq. Thus, administration of 111 MBq of [⁸⁹Zr]Zr-PSMA-617 (mean injected activity) induced a total effective dose of 10.1 mSv.

Among 7 patients with either negative (4 patients) or with undetermined findings on [⁶⁸Ga]Ga-PSMA-11 PET/CT (3 patients), at least one lesion (range 1–5 per patient, in total n = 14) that was suggestive for prostate cancer was detected in 6 patients with [⁸⁹Zr]Zr-PSMA-617 (Table 1). Out of all identified lesions (n = 14), 10 were lymph node metastases (in 4 patients), 3 were local recurrence (in 3 patients), and 1 was peritoneal metastasis. The majority of tumor lesions (n = 11) were not visible on PET/CT imaging at 1 h p.i., but were delineated at later time points (9 of them first visible at 24 h p.i., and 2 at 48 h p.i.). All tumor lesions were detectable at 48 h p.i. and 72 h p.i.. No additional tumor lesions were found at 72 h p.i.. Three tumor lesions (in 3 patients), which were identified at 1 h p.i., were also visible in [⁶⁸Ga]Ga-PSMA-11 PET/CT. An exemplary tumor lesion and its uptake over time is shown in Fig. 3. A graphical representation of the SUV_max and TBR values of all lesions in all patients is provided in Fig. 4. In those lesions, which could be identified from early imaging (1 h p.i), SUV_max further increased substantially from 1 to 24 h p.i.. In late imaging (≥ 24 h p.i.), SUV_max plateaued in all lesions and TBR increased continuously over time.

In 3 of 4 (75%) patients with previously negative [⁶⁸Ga]Ga-PSMA-11 PET/CT, lesions that were visually considered as suggestive for prostate cancer were identified using [⁸⁹Zr]Zr-PSMA-617. Two patients were found to have focal uptake in the prostate bed and one of these two and another patient revealed uptake in lymph nodes suggesting local recurrence and lymph node metastases, respectively (Table 1). Figure 5 exemplifies a local recurrence and a lymph node metastasis visible on [⁸⁹Zr]Zr-PSMA-617 PET/CT at 48 h p.i., which could not be identified by [⁶⁸Ga]Ga-PSMA-11 PET/CT. In 2 of 3 patients with undetermined findings in [⁶⁸Ga]Ga-PSMA-11 PET/CT, the respective findings were confirmed by [⁸⁹Zr]Zr-PSMA-617 PET/CT, whereas in 1 of 3 patients, no corresponding uptake was identified (Fig. 6). Furthermore, in 2 of these 3 patients, additional lesions were detected in [⁸⁹Zr]Zr-PSMA-617 PET/CT, which were unidentified by [⁶⁸Ga]Ga-PSMA-11 PET/CT (Table 1).

Subsequently, five patients received radiotherapy adjusted according to the results of [⁸⁹Zr]Zr-PSMA-617-PET/CT and in the other two patients, ADT was initiated. There were minor (e.g., additional boost) and also major adjustments on radiation treatment planning (e.g., extension of the radiation field) by the findings on [⁸⁹Zr]Zr-PSMA-617-PET/CT. In all 5 patients who received radiation therapy, a PSA decrease was achieved, in 4 patients by more than 70% and in 1 patient of 30%.