[18F]-JK-PSMA-7 PET/CT Under Androgen Deprivation Therapy in Advanced Prostate Cancer

Purpose PSMA imaging is frequently used for monitoring of androgen deprivation therapy (ADT) in prostate cancer. In a previous study, [18F]-JK-PSMA-7 exhibited favorable properties for tumor localization after biochemical recurrence. In this retrospective study, we evaluated the performance of [18F]-JK-PSMA-7 under ADT. Procedures We examined the performance of [18F]-JK-PSMA-7 in 70 patients (first cohort) with increasing or detectable PSA values under ADT (PSA < 2 ng/ml for 21/70 patients). We further analyzed 58 independent patients with PSA levels < 2 ng/ml under ADT, who were imaged with [68Ga]PSMA-11 or [18F]DCFPyL (second cohort). Finally, we compared detection rates between [18F]-JK-PSMA-7, [68Ga]PSMA-11, and [18F]DCFPyL. Results In the first cohort, we detected [18F]-JK-PSMA-7-positive lesions in 63/70 patients. In patients with PSA levels ≥ 2 ng/ml, the detection rate was 100 % (49/49). In patients with PSA < 2 ng/ml, the detection rate was significantly lower (66.7 %, 14/21, p = 9.7 × 10−5) and dropped from 85.7 % (12/14, PSA levels between 0.3 and 2.0 ng/ml) to 28.6 % (2/7) for PSA levels < 0.3 ng/ml (p = 1.73 × 10−2). In the second cohort (PSA < 2 ng/ml), the detection rate was 79.3 % (46/58) for [68Ga]PSMA-11 or [18F]DCFPyL. Again, the detection rate was significantly higher (p = 1.1 × 10−2) for patients with PSA levels between 0.3 and 2.0 ng/ml (87.0 %, 40/46) relative to those with PSA levels < 0.3 ng/ml (50 %, 6/12). No significant difference was found between [18F]-JK-PSMA-7 and [68Ga]PSMA-11 or [18F]DCFPyL in patients with PSA levels < 2 ng/ml (p = 0.4295). Conclusion [18F]-JK-PSMA-7 PET showed a high detection rate in patients with PSA levels ≥ 0.3 ng/ml under ADT. The lower PSA threshold of 0.3 ng/ml for high detection rates was consistent across the three PSMA ligands. Thus, PSMA imaging is suitable for clinical follow-up of patients with increasing PSA levels under ADT.

PSMA PET is now included in international guidelines for imaging in biochemical recurrence (BCR) [8]. Additionally, PSMA PET scans are often in high demand for monitoring of patients with relapsed or metastasized prostate cancer. Under continuous androgen deprivation therapy (ADT) in castration-resistant prostate cancer (CRPC) with PSA values of 9 2 ng/ml, Fendler and colleagues reported PSMA positivity in 196/200 patients [9]. These patients had mainly received [ 68 Ga]PSMA-11 (n = 195) while some were examined with [ 18 F]DCFPyL (n = 5). When administering enzalutamide or abiraterone in addition to an LHRH analog in CRPC, expression levels of PSMA increase in metastases, as recently shown in a cohort of 7 patients [10]. Thus, PSMA-specific PET tracers are not negatively affected by ADT once metastases have become castrate resistant.
In contrast to castrate-resistant patients, the impact of ADT on PSMA expression in hormone-sensitive prostate cancer (HSPC) remains unclear. A comparison of PET scans before and after therapy with ADT revealed a heterogeneous pattern of change in the PSMA expression of metastases in small cohorts of 10, 7, and 5 patients, respectively [10][11][12]. In the majority of these patients, PSMA expression in tumor metastases decreased or stopped completely after the start of ADT. To date, the influence of ADT on PSMA expression has been investigated primarily for the tracer [ 68 Ga]PSMA-11.
Data on the probability of PSMA-positive lesions in patients with low PSA levels under continuous ADT are sparse. There is no consensus as to which PSA value for PMSA PET imaging is a sensitive tool under ADT. The definition of castration-resistant prostate cancer in the guidelines of the European Association of Urology includes a castrate serum testosterone G 50 ng/dl plus three consecutive rises in PSA, resulting in two 50 % increases over the nadir, and a PSA level 9 2 ng/ml [8]. Hence, PSA levels of both G 2 ng/ml and 9 2 ng/ml are among the criteria determining treatment plans.
Here, we report the results of the [ 18 F]-JK-PSMA-7 PET/ CT from the first year of clinical implementation in patients under ADT who were referred for PET imaging by their local urologists with a broad spectrum of PSA levels. The observed detection rate in patients with PSA levels G 2 ng/ml was re-examined in a second cohort investigated with [ 68 Ga]PSMA-11 or [ 18 F]DCFPyL.

Patient Characteristics and Study Design
Hypotheses and Study Design We hypothesized that the detection rate of PSMA-positive metastases under continuous ADT was higher in patients with PSA levels ≥ 2 ng/ml relative to patients with PSA levels G 2 ng/ml (hypothesis 1). Our first cohort was imaged with [ 18 F]-JK-PSMA-7 in 2017 as a routine clinical procedure (cohort 1). As stated in the "Results" section, hypothesis 1 was confirmed. Based on this finding, we next hypothesized that high detection rates could be achieved for patients with PSA levels G 2 ng/ml under ADT (hypothesis 2). To test this hypothesis, we retrospectively analyzed an independent cohort that consisted exclusively of patients with PSA levels G 2 ng/ml under ADT. These patients were imaged with [ 68 Ga]PSMA-11 or [ 18 F]DCFPyL as a routine clinical procedure in 2015 and 2016 (cohort 2). The advantage of cohort 2 was that more patients could be included in the group used to calculate the lower PSA threshold for beneficial imaging with PSMA agents. To determine the lower PSA threshold, we continuously adjusted the PSA threshold and divided the number of PSMA-positive patients with PSA levels above this threshold by the total number of patients with PSMApositive lesions. Finally, we compared the detection rates for PSMA-positive lesions in patients with PSA G 2 ng/ml between cohorts 1 and 2, assuming the non-inferiority of the PSMA ligands [ 18 F]-JK-PSMA-7, [ 68 Ga]PSMA-11, and [ 18 F]DCFPyL to one another.
In this IRB-approved study , all patients (cohorts 1 and 2) gave their written, informed consent to PET imaging and inclusion of their data in a retrospective analysis. All procedures were carried out in compliance with the regulations of the local authorities responsible (District Administration of Cologne, Germany). Nine patients also received enzalutamide or abiraterone, and 11 patients had been treated with docetaxel. As a PSA level 9 2 ng/ml is one of several criteria to determine therapy decisions in advanced prostate cancer under ADT [8], we chose a PSA threshold of 2.0 ng/ml to differentiate between two patient groups under ADT. The PSA level was G 2 ng/ml in 21 patients (mean PSA 0.72 ± 0.52 ng/ml) and ≥ 2.0 ng/ml in 49 patients (158.53 ± 575.47 ng/ml). In the subcohort of 21 patients with PSA levels G 2.0 ng/ml, we retrospectively identified a subgroup of 7 patients with PSA levels G 0.3 ng/ml (PSA 0.11 ± 0.11 ng/ml). The PSA level in the remaining 14 patients was 1.03 ± 0.33 ng/ml. Data on the PSA nadir were available in 10 of the 21 patients in whom the PSA level was G 2.0 ng/ml: Four patients showed an increase in PSA level by a factor 9 5, while in another 5, the PSA level increased by a factor of 2-5, and in one patient an increase of G 50 % was observed.
The distribution of PET findings for different ranges of PSA is reported in Table 1. In this cohort, we identified 58 patients, who were 70.9 ± 6.8 years old and presented with PSA levels G 2 ng/ml under continuous ADT. One patient had received docetaxel. The average PSA value of these 58 patients was 0.94 ± 0.59 ng/ml under ADT. In 12 patients, PSA levels were G 0.3 ng/ml (average PSA 0.07 ± 0.09 ng/ml), and in the remaining 46 patients, PSA levels were between ≥ 0.3 and 2 ng/ml in (average PSA 1.16 ± 0.35 ng/ml). Data on PSA nadir were available for 25 patients. In 16 patients, PSA levels had increased 5-fold and in 9 patients, PSA levels had remained stable or increased by no more than 50 %. Gleason scores were available in 41 patients and were G 8 in 18 patients, = 8 in 12 patients, and 9 8 in 11 patients.
The distribution of PET findings for different ranges of PSA is reported in Table 2.
Comparison of the PSMA ligands The detection rate for PSMA-positive lesions in cohort 1, the subgroup with PSA G 2 ng/ml ([ 18 F]-JK-PSMA-7) was compared with the detection rate in cohort 2 ([ 68 Ga]PSMA-11 or [ 18 F]DCFPyL) by a chi-square test.

Imaging
PET imaging was performed as previously described [1-3, 6, 7]. In brief, all images were acquired on a Biograph mCT 128 Flow PET/CT scanner (Siemens Healthineers, Erlangen, Germany) and reconstructed using an ultra-highdefinition algorithm. The same filters and acquisition times (flow motion bed speed of 1.5 mm/s) were used for the different PSMA ligands. PET imaging started from the middle of the thighs to the tip of the skull. PET imaging was started 2 h after the injection of  ADT, androgen deprivation therapy; PSMA, prostate-specific membrane antigen; T+, PSMA-positive tissue within the prostate fossa; N+, PSMA-positive lymph node; M+, PSMA-positive lesion in the bone, lung, or liver ADT, androgen deprivation therapy; PSMA, prostate-specific membrane antigen; T+, PSMA-positive tissue within the prostate fossa; N+, PSMA-positive lymph node; M+, PSMA-positive lesion in the bone, lung, or liver 14] by a team of two specialists in nuclear medicine and one radiologist. Any disagreement was resolved in consensus. The same team interpreted PET/CT scans from cohorts 1 and 2.

Tracer Synthesis
The detailed procedure for the radiosynthesis of [ 18 F]-JK-PSMA-7 using the "minimalist light" protocol was described previously [5]. Two batches per week were produced at the Forschungszentrum Juelich in accordance with applicable good manufacturing practice (GMP). The GMP-based quality control measures included radiochemical purity, endotoxin testing, pH value, and the determination of residual content of solvents such as acetonitrile, acetone, tertiary butanol, and tetra-ethylammonium-hydrogen-carbonate [TEAHC] [5][6][7].  %) (Figs. 1, 2, and 3). This detection rate varied with the PSA level under ADT and was significantly (p = 9.7 × 10 −5 , two-tailed Fisher's exact test) higher in patients with PSA levels ≥ 2 ng/ml (100 %, 49/49) compared with PSA levels G 2 ng/ml (66.7 %, 14/21). However, individual patients with very low PSA levels still showed PSMA-positive lesions (Fig. 3). We therefore set out to derive a lower PSA threshold for imaging of prostate cancer patients under ADT that avoids unnecessary PSMA imaging but captures most PSMApositive patients. In agreement with Ceci and colleagues [15], whose best nomogram-derived probability threshold was associated with a sensitivity of 84.7 %, we determined a PSA threshold for which ≥ 85 % of the patients with . This detection rate dropped significantly to 28.6 % (2/7) for PSA levels G 0.3 ng/ml (p G 1.73 × 10 −2 , two-tailed Fisher's exact test) (Fig. 4a, b).

Comparison of the PSMA Ligands
The detection rates did not differ significantly between [

Discussion
This study revealed the following three findings on the PSMA imaging of prostate cancer under ADT: Using [ 18 F]-JK-PSMA-7, PET/CT detection rates were statistically higher in prostate patients under ADT with PSA levels ≥ 2 ng/ml compared with patients with PSA level G 2 ng/ml. PSMA-positive lesions could be detected in many patients with PSA levels G 2 ng/ml under continuous ADT. Detection rates were high for PSA levels as low as 0.3 ng/ml. However, detection rates were significantly higher in patients with PSA levels ≥ 0.3 ng/ml than in patients with PSA levels G 0.3 ng/ml. Detection rates did not differ significantly between   [5,7]. We now perform PSMA imaging frequently for monitoring of androgen deprivation therapy (ADT) in routine diagnostics. Substantially, fewer studies have examined the performance of PSMA imaging under ADT [9][10][11]15] and our calculation of the lower PSA threshold for this clinical scenario is helpful for the accurate selection of patients under ADT for PSMA PET/CT imaging.
Consistent with a recent study by Fendler and colleagues [9], we found a robust overexpression of PSMA in the metastases of patients with PSA levels of ≥ 2 ng/ml under long-term ADT. Moreover, our study demonstrates that [ 18 F]-JK-PSMA-7 PET has a high detection rate for patients under ADT with PSA levels as low as 0.3 ng/ml. This PSA threshold is as low as the thresholds derived from studies examining the performance of PSMA imaging based on biochemical recurrence after a curatively intended prostatectomy. Potential explanations of this observation include the possibility that tumor cells might increase their PSMA expression in response to long-term ADT.
An overview of PSMA PET studies under ADT is presented in Table 3. The largest cohort included 200 patients and demonstrated that the sensitivity of PSMA PET is high in CRPC under ADT [9]. Most patients in the largest cohort had PSA levels 9 2 ng/ml (198/200). The present study demonstrates the potential of PSMA PET once PSA levels have exceeded 0.3 ng/ml under ADT. The association between detection rate and PSA level holds for  [20]. In clinical practice, PSMA imaging is feasible under long-term ADT as soon as the PSA level exceeds 0.3 ng/ml. Besides the PSA level, further predictors of a positive scan are PSA doubling time and grading [15,20]. The statistically calculated PSA threshold of 0.3 ng/ml under long-term ADT is conservative, as 42.1 % of the patients with PSA levels G 0.3 ng/ml had a PSMApositive scan, when taking the results of the 3 PSMA ligands together. We found that in patients with unfavorable Gleason scores, PSMA PET can reveal PSMA-positive lesions even at lower PSA levels under ADT.

Limitations
Our work demonstrates that [ 18 F]-JK-PSMA-7 PET can detect PSMA expression in metastases and relapses under continuous ADT. Some patients were heavily pretreated. However, we have no information on castrate serum testosterone or on consecutive rises in PSA resulting in two 50 % increases over the nadir, which are defining criteria for hormone-sensitive or castration-resistant prostate cancer [8]. We therefore used the PSA level at the time of PET to differentiate between a low PSA level cohort and a high PSA level cohort. We further emphasize that we did not investigate PSMA expression on the commencement of ADT. On the other hand, the ENZAMET phase 3 trial has shown that the androgen-receptor inhibitor enzalutamide will improve survival in men with metastatic, hormone-sensitive prostate cancer [21]. In the future, a greater variety of therapeutic agents for ADT will be used as first-line therapy of patients with polytope metastatic prostate cancer.
Comparison of our [ 18 F]-JK-PSMA-7 PET cohort with the independent [ 68 Ga]PSMA-11 or [ 18 F]DCFPyL PET cohort was hindered by the fact that we did not have details of the Gleason score in each patient and could not therefore carry out a matched-pair analysis.