Phase III Study of 18F-PSMA-1007 Versus 18F-Fluorocholine PET/CT for Localization of Prostate Cancer Biochemical Recurrence: A Prospective, Randomized, Crossover Multicenter Study

Visual Abstract

Pr ostate cancer (PCa) is the most prevalent cancer in men, with approximately 473,000 new diagnoses and more than 108,000 deaths in Europe in 2020 (1). Although long-term outcomes are good for most men, recurrence after definitive therapy is common. One study found that 37% of patients with radical prostatectomy and 48% of patients with radiation therapy had biochemical recurrence within 15 y of the initial definitive treatment; for both, most relapses occurred within the first 5 y (2). The diagnosis of PCa recurrence after prior definitive therapy is based on an increase in serum prostate-specific antigen (PSA); the threshold level varies by treatment, being higher for patients treated with radiation than for those treated by radical prostatectomy (3). PET/CT imaging is the recommended modality for the localization of PCa recurrence (3). 18 F-fluorocholine has been recommended for PET/CT imaging of PCa recurrence since a marketing authorization was granted in France in 2010. The development of radiopharmaceuticals that directly target the extracellular domain of the prostate-specific membrane antigen (PSMA) resulted in an improvement in the detection of PCa lesions, and this approach is now recommended for PET/CT imaging of PCa recurrence.
A recent metaanalysis evaluated the diagnostic efficacy of all PSMA-directed PET agents and reported an overall detection rate of 74.1% with no notable differences among the various tracers (4). The authors concluded that PSMA-directed PET agents were preferable to choline PET, particularly in patients with a serum PSA level of less than 1 ng/mL (4). However, to date, only singlecenter studies with a limited number of patients have compared 18 F-fluorocholine with a radiolabeled PSMA ligand for PET/CT imaging of PCa recurrence localization, and large randomized controlled trials are lacking (5). 18 F-PSMA-1007 was developed by Cardinale et al. (6) and Giesel et al. (7) in Heidelberg in 2016 as a PSMA-targeting ligand with low urinary excretion. The compound showed a high detection rate at low PSA values and high sensitivity and specificity in both biochemical recurrence and primary staging (8,9). In some patients, nonspecific bone uptake may be a confounding factor (10).
The ABX-CT-301 study (NCT04102553) aimed to compare 18 F-PSMA-1007 PET/CT and 18 F-fluorocholine PET/CT for the detection of PCa lesions in patients with biochemical recurrence. Secondary objectives were to compare the detection rates of the clinical investigators for both radiopharmaceuticals in a patientbased analysis; to assess the diagnostic performance of both radiotracers for PCa lesions in a region-based analysis; to assess the impact on diagnostic thinking, therapeutic decision making, and adequacy of therapy changes for both radiotracers; and to assess the safety profile of 18 F-PSMA-1007.

Population
This study was a prospective, open-label, randomized, 2-armed crossover study conducted in 6 study centers in France. Men who were at least 18 y old, who were diagnosed with PCa, and who had prior definitive therapy were considered for enrollment. Eligible patients presented with suspected PCa recurrence, defined as 3 consecutive PSA increases or a PSA rise of greater than or equal to 2.0 ng/mL above the nadir after radiotherapy (external-beam radiation therapy or brachytherapy) or cryotherapy or a PSA rise of greater than or equal to 0.2 ng/mL after prostatectomy. The main exclusion criteria were participation in another therapeutic clinical trial within 5 d of enrollment in the present study and a life expectancy of less than 6 mo.
The study protocol was approved by a national ethics committee certified by the French Ministry of Health (Institutional Review Board: IORG0009855). All patients gave written informed consent before randomization.

Intervention
All men underwent both 18 F-PSMA-1007 PET/CT and 18 F-fluorocholine PET/CT using a standardized imaging protocol (supplemental materials, available at http://jnm.snmjournals.org). Patients were randomized using a computer-generated block-randomized sequence, stratified by center, to receive either 18 F-PSMA-1007 PET/CT or 18 Ffluorocholine PET/CT first. Patients underwent both PET/CT examinations within a minimum of 24 h and a maximum of 240 h; depending on the individual site preferences, either low-dose or diagnostic CT could be used, but the use of contrast agents was not permitted. At each PET/CT visit, vital signs were recorded before and after injection of the study drug and again at the end of the PET/CT examination. Samples for laboratory tests, including serum PSA levels, were obtained before the administration of each study drug. Patients were monitored for adverse events for 24 h after the second PET/CT examination (supplemental materials). Then we collected all data related to the treatments, additional diagnostic methods (including biopsy confirmation of detected foci, if feasible) and PSA values obtained in the 6 months during which the patients were followed.

Image Reading and Standard of Truth
PET images were read on-site on the day of acquisition by investigators who were not masked regarding clinical data and were transferred to a core imaging laboratory where they were evaluated by 3 independent masked readers (supplemental materials). 18 F-PSMA-1007 and 18 Ffluorocholine images were read on separate days, at least 1 wk apart.
The results of an "average reader" were determined statistically from the 3 independent reader results and not by a consensus read.
The composite standard of truth (recurrence, no recurrence, or undetermined) at the time of imaging was determined by an independent expert panel that considered all available clinical patient data from before inclusion to the end of the follow-up period, excluding all information from PET/CT (supplemental materials). The expert panel consisted of a urologist, a radiologist, and a nuclear physician; they reached their conclusions by consensus.

Outcomes
The primary objective was to compare 18 F-PSMA-1007 with 18 Ffluorocholine with regard to the "correct detection rate" for recurrent PCa lesions on a patient basis, as determined by 3 independent readers and confirmed by an independent expert panel on the basis of a composite standard of truth (supplemental materials).
Secondary objectives were to compare the correct detection rates of the clinical investigators; to assess the correct detection rates of both radiotracers for PCa lesions in a region-based analysis; to report the impact on diagnostic thinking (impact of PET/CT result on posttest vs. pretest probability of a correct diagnosis), therapeutic decision making (impact on the comprehensive process by which physicians make decisions about the PCa response), and adequacy of management changes by the Investigators at 3 time points (before PET, immediately after both PET studies, and at the end of follow-up) and by the expert panel (only at the end of follow-up), using 3 dedicated forms (supplemental materials); to compare the masked intra-and interreader agreements; and to assess the safety profile of 18 F-PSMA-1007.

Statistical Analysis
Analysis was performed using SAS 9.4 or higher (SAS Institute), detailed in the supplemental materials. A P value of less than or equal to 0.05 was considered statistically significant. Descriptive statistics were calculated for quantitative variables; frequency counts by category were given for qualitative variables. Ninety-five percent CIs or interquartile ranges were given when appropriate. The intention-to-treat (ITT) population was the primary population for the analyses of efficacy endpoints and all baseline characteristics. The correct detection rate was determined for each reader individually. Generalized estimation equations were used to account for the correlations between readers' assessments and to summarize the overall reader results (average reader). Patients for whom the expert panel could assess the true disease state on a patient level but for whom the affected region could not be identified by the expert panel were included as correct assessments on a patient basis. If, on a region basis, there was no region with a correct detection of recurrence compared with the standard of truth, then the patient was regarded as having a false-negative result. If at least in 1 region the reader and the expert panel detected a recurrence, independent of the other regions, then this result was classified as a truepositive result. Subgroup analyses were performed on the basis of the PSA level at baseline: less than 0.5 ng/mL, 0.5 ng/mL to less than 1.0 ng/mL, greater than or equal to 1.0 ng/mL to less than 2.0 ng/mL, and greater than or equal to 2.0 ng/mL. For the primary analysis on a patient level, 2 distinct analyses were performed by considering undetermined results as positive or negative for PCa recurrence. For subgroup and secondary analyses, undetermined results were considered negative for PCa recurrence. Intra-and interreader agreements were evaluated using pairwise and multiple Cohen k-statistics. Each masked reader read 10% of the images twice (on separate occasions). Interreader agreement was assessed pairwise and across all 3 readers. The degree of agreement was defined as described by Landis and Koch (11).The sample size was selected to provide a power of at least 80% to detect a 10% difference in correct detection rates between the 2 products.

Population
From March 5, 2019, to October 8, 2020, 200 patients consented to this study; 195 were randomized, and 189 completed all follow-up assessments (Fig. 1). One patient ended study participation prematurely because he died 3.5 mo after PET imaging but was included in the ITT population. Most of the patients had previously undergone prostatectomy, and the median serum PSA level was 1.7 ng/mL. Study population characteristics are summarized in Table 1.
The median follow-up period for the ITT population was 8.3 mo (range, 2.9-16.1 mo), as its duration was extended because of the COVID-19 pandemic.

Primary Objective
Per-patient and per-region PET findings are detailed in the supplemental materials. At the patient level, the masked readers found evidence for PCa recurrence in 145-162 patients (76.3%-85.3%) with 18      examined PSA-level subgroups, the correct detection rate was statistically higher for 18 F-PSMA-1007 (Table 3). Masked intra-and interreader agreements for the detection of metastases (patient level) ranged from 0.24 to 0.73 and from 0.    Comparison of Region-Based Correct Detection Rates According to Masked Readers' Findings. In the 72 patients for whom 1 or more regions could be assessed by the expert panel, there were 78 regions with confirmed PCa lesions. The most common sites for PCa lesions were the pelvis (59 patients) and the spine (6 patients) (supplemental materials). In the patients considered to have a positive PET result, more suggestive lesions were detected with 18 F-PSMA-1007 than with 18 F-fluorocholine, especially in patients with 3 or more lesions.
Impact on Diagnostic Thinking, Therapeutic Decision Making, and Adequacy of Therapy Changes. The investigator assessments of the changes in diagnostic thinking after both PET/CT examinations and at the end of follow-up are summarized in Tables 4 and  5 and the supplemental materials. Treatment plans before and after PET/CT examinations were available in 187 patients. Treatment plans were changed in 100 patients; 89 of the changes were major (supplemental materials). Changes in diagnostic thinking due to PET/CT were reported in 149 patients. Diagnostic thinking was unchanged in 41 patients (including 3 with no reported answer in the questionnaire). In the 149 patients for whom there were changes in diagnostic thinking, 18 F-PSMA-1007 contributed more in 93 patients (62%), both tracers contributed equally in 49 patients (33%), and 18 F-fluorocholine contributed more in 4 patients (3%).
In 122 patients, PET/CT led to a more accurate diagnosis that benefited them. In 11 patients, PET/CT was not to the benefit of the patient, and in 45 patients, PET/CT did not exert a positive or negative influence. In the 122 patients with a more accurate diagnosis after PET/CT that benefited them, 18 F-PSMA-1007 contributed more in 88 patients, 18 F-fluorocholine contributed more in only 6 patients, and both contributed equally in 27 patients.
Safety Profile of 18 F-PSMA-1007. There were no serious adverse events. No patient discontinued study participation because of an adverse event. Four patients had 4 events (toothache, diarrhea, chest discomfort, and arterial hypertension) after the administration of 18 F-PSMA-1007, and 1 patient had 1 event (shoulder pain) after the administration of 18 F-fluorocholine. None of the adverse events was considered to be attributable to the study drug.

DISCUSSION
To our knowledge, the present study is the first multicenter, crossover randomized study to compare 18 F-PSMA-1007 and 18 Ffluorocholine for the localization of biochemical recurrence of PCa. The correct detection rate was significantly higher with 18 F-PSMA-1007 PET/CT than with 18 Ffluorocholine PET/CT. Results were similar for the population as a whole and for the 72 patients for whom individual lesions could be verified, either by biopsy or response to local treatment. The difference was especially pronounced in patients with lower serum PSA levels, allowing earlier, targeted salvage treatment. Our results for 18 F-PSMA-1007 are in agreement with those reported in the literature for 68 Ga-PSMA-11 (4,(12)(13)(14)(15)(16)(17)(18). In a metaanalysis, Treglia et al. (19) also found similar results when comparing a PSMA tracer ( 68 Ga-PSMA-11 or 64 Cu-PSMA-617) and radiocholine. Because the results obtained with the various PSMA ligands are generally similar, it is generally accepted that the PSMA ligands are interchangeable for this indication. PSMA ligands radiolabeled with 18 F have wider accessibility than those radiolabeled with 68 Ga.
The investigator assessments after PET/CT and at the end of follow-up demonstrated the superiority of 18 F-PSMA-1007 over 18 F-fluorocholine in identifying sites of recurrence. The impact of 18 F-PSMA-1007 was to benefit the patient in most cases. These results are likely linked to the higher correct detection rate for 18 F-PSMA-1007. In previous studies with 18 F-fluorocholine (13,20) and 68 Ga-PSMA-11 (5,21), rates of impact on patient management of 39%-58% were reported. Our study is consistent with published data on 18 F-fluorocholine and demonstrates a higher impact of 18 F-PSMA-1007 on patient management.
The strengths of our study are its prospective, multicenter, randomized crossover design. Limitations include the lack of histopathology for most lesions. Because obtaining histopathology is often ethically questionable or medically impractical, our standard of truth was a composite based on biopsy, response to local therapy, imaging, and changes in serum PSA levels during 6 mo of follow-up, as established by an independent panel of experts. The use of the independent panel removed potential bias in determining "truth" while modeling what is done in "real-life" practice.
In this work, we found that 18 F-PSMA-1007 had an impact on diagnostic thinking and therapeutic decision making and that therapy changes were more beneficial when based on 18 F-PSMA-1007 FIGURE 2. 62-y-old patient with history of PCa (International Society of Urological Pathology grade 3; PSA level of 5.7 ng/mL), initially treated with prostatectomy (pT3N0R0), prostate bed radiation therapy, and 6 mo of androgen deprivation therapy (ADT), presenting with PSA recurrence (0.72 ng/mL). 18 F-PSMA-1007 PET/CT detected pelvic lymph nodes (red arrow) and bone metastases (green arrow) that were not detected by 18 F-fluorocholine PET/CT. Therapeutic management changed from targeted radiation therapy before PET to ADT after PET, leading to drop in PSA level to 0.1 ng/mL at 6 mo. BW 5 body weight.
PET/CT findings than when based on 18 F-fluorocholine PET/CT findings (Fig. 2). However, we did not make a statistical comparison of these data because the study was not powered for this purpose.

CONCLUSION
This prospective, multicenter, open-label, crossover randomized study demonstrated that 18 F-PSMA-1007 PET/CT localizes PCa recurrence in significantly more patients than 18 F-fluorocholine PET/CT, especially in patients with low PSA serum levels. 18 F-PSMA-1007 PET/CT also had a higher impact on diagnostic thinking, therapeutic decision making, and therapy changes than 18 F-fluorocholine PET/CT.

DISCLOSURE
This study was funded and supported by ABX Advanced Biochemical Compounds, Radeberg, Germany. Ines Rudolph and Alexander Hoepping work for ABX Advanced Biochemical Compounds. Charles Merlin has received personal fees and nonfinancial support from Curium Pharma and Novartis. Mathieu Gauth e has received personal fees and nonfinancial support from Astellas, Curium Pharma, and Novartis. The manuscript was edited for English language by Dr. Jay R Wiggins, PhD, and Dr. Kelsey L. Pomykala, MD, supported by ABX Advanced Biochemical Compounds. The ABX-CT-301 ClinicalTrials.gov number is NCT04102553. No other potential conflict of interest relevant to this article was reported.