68Ga-PSMA and 68Ga-DOTA-RM2 PET/MRI in Recurrent Prostate Cancer: Diagnostic Performance and Association with Clinical and Histopathological Data

Simple Summary Prostate cancer (PCa) relapse occurs in up to 50% of patients after radical treatment. Once PCa recurrence is detected, a precise identification of the number and sites of recurrence is necessary to tailor the treatment on the patient’s needs. Positron emission tomography (PET) plays a pivotal role in this clinical setting and new radiotracers have been developed to improve its performance. While 68Ga-PSMA is a well-established radiotracer for PCa recurrence detection, 68Ga-DOTA-RM2 is a recently proposed tracer that targets the gastrin-releasing peptide receptors that are overexpressed in prostate cancer. In this work, the performance of 68Ga-PSMA and 68Ga-DOTA-RM2 PET/MRI in identifying recurrent disease were compared on the same cohort, using the same study protocol, as this is the only way to assess whether one outperforms the other and therefore should be preferred in clinical practice. Furthermore, the association between PET findings and clinical and histopathological characteristics was investigated to find potential biomarkers. Abstract The aim of the present study is to investigate and compare the performances of 68Ga-PSMA and 68Ga-DOTA-RM2 PET/MRI in identifying recurrent prostate cancer (PCa) after primary treatment and to explore the association of dual-tracer PET findings with clinical and histopathological characteristics. Thirty-five patients with biochemical relapse (BCR) of PCa underwent 68Ga PSMA PET/MRI for restaging purpose, with 31/35 also undergoing 68Ga-DOTA-RM2 PET/MRI scan within 16 days (mean: 3 days, range: 2–16 days). Qualitative and quantitative image analysis has been performed by comparing 68Ga-PSMA and 68Ga-DOTA-RM2 PET/MRI findings both on a patient and lesion basis. Clinical and instrumental follow-up was used to validate PET findings. Fisher’s exact test and Mann-Whitney U test were used to investigate the association between dual-tracer PET findings, clinical and histopathological data. p-value significance was defined below the 0.05 level. Patients’ mean age was 70 years (range: 49–84) and mean PSA at time of PET/MR scans was 1.88 ng/mL (range: 0.21–14.4). A higher detection rate was observed for 68Ga-PSMA PET/MRI, with more lesions being detected compared to 68Ga-DOTA-RM2 PET/MRI (26/35 patients, 95 lesions vs. 15/31 patients, 41 lesions; p = 0.016 and 0.002). 68Ga-PSMA and 68Ga-DOTA-RM2 PET/MRI findings were discordant in 11/31 patients; among these, 10 were 68Ga-PSMA positive (9/10 confirmed as true positive and 1/10 as false positive by follow-up examination). Patients with higher levels of PSA and shorter PSA doubling time (DT) presented more lesions on 68Ga-PSMA PET/MRI (p = 0.006 and 0.044), while no association was found between PET findings and Gleason score. 68Ga-PSMA has a higher detection rate than 68Ga-DOTA-RM2 in detecting PCa recurrence. The number of 68Ga-PSMA PET positive lesions is associated with higher levels of PSA and shorter PSA DT, thus representing potential prognostic factors.


Introduction
Prostate cancer (PCa) relapse affects up to 50% of patients treated with radical prostatectomy (RP) or external-beam radiotherapy (EBRT) as primary treatment for clinically localized disease [1].
When biochemical recurrence (BCR) is detected, an accurate identification of location and extent of cancer recurrence is mandatory in order to address patients to directed therapies with prolonged intervals of cancer-free survival, thus avoiding systemic treatments, including androgen deprivation therapy [2].
Prostate Specific Antigen (PSA) is the commonly used circulating biomarkers identifying BCR; however, it is unable to provide information regarding the exact location of disease recurrence.
Of note, patients with low values of PSA might benefit the most by receiving early local therapies with curative intent; however, conventional imaging techniques have limited sensitivity in detecting tumour recurrences at PSA values lower than 1 ng/mL [1]. On the other hand, patients presenting metastatic disease should be candidate to systemic therapies. Therefore, the distinction between local or systemic disease is of utmost importance to plan the most appropriate treatment [3]. In this scenario, imaging plays a fundamental role in the identification of local and/or distant metastases.
In the last decade, Prostate Specific Membrane Antigen (PSMA), a transmembrane protein with a significantly increased expression in PCa cells, has gained particular attention as a radiotracer for positron emission tomography (PET) imaging in PCa, in view of its better sensitivity and specificity in detecting of metastatic disease, compared to conventional imaging, also at low PSA values [4][5][6].
Additionally, gastrin releasing peptide receptors (GRPR) are G-protein coupled receptors overexpressed in prostate tumours both at the mRNA and the protein level [7]. For this reason, GRPR analogues have been recently identified as alterative targets for molecular imaging in prostate cancer [8]. So far, among the available radiotracers targeting GRPR, 68 Ga-DOTA-RM2 has been the most investigated [9][10][11][12].
Very few studies have been reported on the combined use of both 68 Ga-PSMA and 68 Ga-DOTA-RM2 in prostate cancer setting, both in the staging [12,13] and restaging [10,14] phase of the disease, with the majority of them using PET/computed tomography (CT) scanners for imaging acquisition [10,13,14].
The recent introduction of fully hybrid PET/Magnetic Resonance Imaging (MRI) paves the way to a completely innovative imaging approach for PCa recurrence. The simultaneous acquisition of PET and multiparametric MR (mp-MR) provides metabolic, structural, and functional imaging information regarding PCa status in a whole-body single session examination, with better soft tissue contrast and reduced radiation exposure in comparison to PET/CT [15,16]. However, very limited data are currently available on its role in this specific clinical setting. In particular, studies using both 68 Ga-PSMA and 68 Ga-DOTA-RM2 with PET/MRI are still anecdotal [10,12,14] Therefore, the aim of our study was to compare the performances of 68 Ga-PSMA and 68 Ga-DOTA-RM2 PET/MRI in the identification of recurrent PCa after primary treatment.
Furthermore, we also investigated the association between 68 Ga-PSMA and 68 Ga-DOTA-RM2 PET/MRI positive findings and clinical, as well as histopathological characteristics.

Patients
In this prospective clinical study, 35 consecutive patients with biochemically recurrent PCa were enrolled from June 2020 to March 2021 at IRCCS San Raffaele Scientific Institute.
Inclusion criteria were age greater than 18 years at the time of PET/MRI scan, previous radical treatment for PCa and biochemical failure, defined as two consecutive PSA measurements ≥ 0.2 ng/mL after RP or an increase by ≥2 ng/mL above the nadir PSA for patients that underwent RT [17].
Exclusion criteria were medical condition possibly interfering and significantly affecting study compliance and contraindications to undergo MRI scan (i.e., severe claustrophobia).
All recruited patients underwent 68 Ga-PSMA PET/MRI and 31/35 also 68 Ga-DOTA-RM2 PET/MRI in two different scan sessions, with at least 48 h interval.
PSA level at different time points and Gleason Score (GS) after RP, whenever available, were collected for all patients.
This study was approved by the Institutional Ethics Committee of IRCCS San Raffaele Scientific Institute (EudraCT 2018-001036-21), and all patients gave written informed consent to participate to the study.
The data presented in the present manuscript are part of a single-centre prospective study still ongoing at IRCCS San Raffaele Scientific Institute (Project GR-2016-02363991).

68 Ga-PSMA PET/MRI Acquisition Protocol
68 Ga-PSMA was synthetized according to the procedure previously described. [12] Fasting condition was requested on the day of 68 Ga-PSMA PET/MRI scan. Images were acquired on a fully hybrid 3 Tesla PET/MRI system (SIGNA PET/MRI; General Electric Healthcare, Waukesha, WI, USA) from the skull base to mid-thigh in the following examined patients.
The 68 Ga-PSMA PET/MR examination protocol included a high statistic (HS) scan of the pelvis (20 min), covering a single bed position, that was simultaneously acquired with the following MRI sequences. Following the single bed HS acquisition, a total-body (TB) PET scan (5-6 FOVs, 4 min/FOV) was then simultaneously acquired to a MRI TB T1 Lava Flex sequence for anatomical localization, and a TB DWI with b = 50 and b = 1000 s/mm 2 . Attenuation Correction (AC) of PET data was performed using MR-Based AC technique based on the processing of the LAVA-Flex sequences acquired simultaneously with the PET data.
2.3. 68 Ga-DOTA-RM2 PET/MRI Acquisition Protocol 68 Ga-DOTA-RM2 was synthetized according to the procedure previously described in [12]. 68 Ga-DOTA-RM2 PET/MRI scan was performed within sixteen days (mean: 3 days, range: 2-16 days) from 68 Ga-PSMA PET/MRI; as for 68 Ga-PSMA PET/MRI, fasting condition was requested on the day of the examination and the same PET/MRI scanner was used.
Following the single bed acquisition, a TB PET scan (5-6 FOVs, 4 min/FOV) was acquired from the base of the skull to mid-thigh simultaneously with a MRI TB Lava Flex sequence and TB DWI with b = 50 and b = 1500 s/mm 2 .
Reconstruction and attenuation correction of PET images were performed by using the same algorithms and parameters used for 68 Ga-PSMA PET images.

PET/MR Image Analysis
68 Ga-PSMA and 68 Ga-DOTA-RM2 image read-out was performed on the Advantage Workstation (AW, General Electric Healthcare, Waukesha, WI, USA), on which PET, MRI and fused PET/MR images could be visualized in axial, coronal, and sagittal planes. HS PET acquisition bed on the pelvic region and TB PET examination of both 68 Ga-PSMA and 68 Ga-DOTA-RM2 PET/MRI images were qualitatively interpreted by two experienced (more than 10 years of experience) Nuclear Medicine physicians, with knowledge of all the available patients' clinical and imaging information. MR examination being part of the patients' clinical work-up, MR images were evaluated by a Radiologist Physician and a report was provided to each patient. As the present study is specifically focused on the use of 68 Ga-PSMA and 68 Ga-DOTA-RM2 PET in recurrent PCa, MRI data were not considered for the purpose of this work and therefore were not included in the analysis. MRI was just used for the localization of the lesions.
The whole-body distribution pattern of both 68 Ga-PSMA and 68 Ga-DOTA-RM2 was qualitatively assessed, and the presence of increased uptake was considered positive for malignancy, with the exception of areas of physiologically increased uptake [18,19]. The anatomical sites were defined based on MR images simultaneously acquired with PET examinations.
Regions of interest (ROI) were semi-automatically defined on transaxial PET images on the single site of recurrence in those patients presenting only one pathological finding and on the most representative site of pathological uptake in those patients presenting multiple lesions. Furthermore, SUVmax and SUVmean, calculated at the 40% threshold of SUVmax [20], were calculated from each ROI drawn on TB PET images, for both radiotracers.
A qualitative comparison between 68 Ga-PSMA and 68 Ga-DOTA-RM2 pathological uptakes in the different anatomical sites localized on MR images was performed in order to describe the possible concordances and discrepancies between 68 Ga-PSMA and 68 Ga-DOTA-RM2 PET/MR imaging. Specifically, 68 Ga-PSMA and 68 Ga-DOTA-RM2 PET/MR findings were defined concordant if the same pathological lesions were detected by both modalities or if both scans resulted negative for the presence of pathological findings. Moreover, Cancers 2022, 14, 334 5 of 19 68 Ga-PSMA and 68 Ga-DOTA-RM2 findings were described as partially concordant if at least one lesion, but not all pathological findings, was detected by PET examination with both tracers. Finally, when pathological uptake was observed in one PET examination, but not in the other, the exams were considered discordant.

Lesion Validation
All PET findings were validated by using clinical and instrumental follow-up. Specifically, 68 Ga-PSMA and 68 Ga-DOTA-RM2 PET/MR imaging findings were considered true positive when at least one of the following criteria was met: (a) histological confirmation on surgically resected specimen; (b) progression (increase in number of pathological 68 Ga-PSMA uptake sites or increase in uptake intensity) on follow-up 68 Ga-PSMA PET/CT or PET/MR studies associated with an increase in PSA level; (c) confirmation on conventional imaging either at baseline (including the diagnostic MR exam performed simultaneously to the 68 Ga-PSMA scan) or during follow-up; (d) disappearance or considerable reduction (i.e., number and intensity) of 68 Ga-PSMA uptake on follow-up PET/CT or PET/MR scans after local or systemic treatment associated with a decrease in PSA level greater than 50%; (e) a decrease in PSA level greater than 50% after selective irradiation of the site of pathological 68 Ga-PSMA or 68 Ga-DOTA-RM2 uptake.
Patients with negative 68 Ga-PSMA and/or 68 Ga-DOTA-RM2 PET/MR were considered true negative in the absence of evidence of disease on conventional imaging or 68 Ga-PSMA PET/CT or PET/MR acquired during the follow-up period (mean follow-up duration for both positive and negative findings: 13.8 months, range: 8.2-17.9 months).

Statistical Analyses
Statistical analyses were performed with R statistical software [21], and Python version 3.7, Python Software Foundation, available at http://www.python.org. (accessed on 30 November 2021). McNemar's test was used to test for the differences in detection rate of recurrent PCa according to 68 Ga-PSMA and 68 Ga-DOTA-RM2 PET/MRI uptake on a patient basis in those patients examined with both radiotracers. Moreover, in the same patients, Wilcoxon-signed rank test was performed to compare the positive findings of 68 Ga-PSMA and 68 Ga-DOTA-RM2 PET/MRI on a lesion basis.
In order to study the association between 68 Ga-PSMA and 68 Ga-DOTA-RM2 PET/MRI findings and clinical data, the sample was subdivided based on PSA concentration in three groups (PSA ≤ 0.5 ng/mL, 0.5-2 ng/mL, ≥2 ng/mL); based on PSA doubling time (PSA DT) in two groups (PSA DT <6 months and ≥ 6 months); and based on GS of the primary tumour (prior to radical therapy) in two groups (GS ≤ 3 + 4 and ≥4 + 3). Fisher's exact test was used to correlate clinical and histopathological data with dual-tracer PET positive findings on a patient basis. Furthermore, Receiver Operating Characteristics (ROC) analysis was computed to investigate the predictive value of clinical and histopathological data (PSA, PSA DT, and GS) to predict 68 Ga-PSMA and 68 Ga-DOTA-RM2 PET/MRI positivity. Corresponding area under the curve (AUC) with 95% confidence interval (CI) were derived.
The Mann-Whitney U test was used to investigate whether higher levels of PSA, GS, and lower PSA DT were related to a higher number of PET positive lesions. All p-values were corrected for multiple testing by false discovery rate and significance was defined below the 0.05 level.

Patients
Thirty-five men (mean age: 70 years; range: 49-84) with BCR of PCa were enrolled in this prospective study. Mean PSA level at time of scanning was 1.88 ng/mL (range: 0.21-14.4). All patients underwent 68 Ga PSMA PET/MRI, with 31/35 also undergoing 68 Ga-DOTA-RM2 PET/MRI because of reduced compliance to the study protocol. Patients' characteristics are reported in Table 1.  68 Ga-PSMA and 68 Ga-DOTA-RM2 PET/MRI findings were concordant in 12/31 patients (6 patients with the same positive findings on both examinations, 6 patients with negative examinations; Figure 1 for an exemplar image), partially concordant in 8/31 patients and discordant in 11/31 patients. Specifically, regarding the discordant findings, 68 Ga-PSMA detected 7 lesions in 5 patients having local recurrences, 5 lesions in 5 patients experiencing bone disease, 4 lesions in 3 patients with lymph nodal recurrence and 2 lesions in one patient with visceral metastases. Conversely, 68 Ga-DOTA-RM2 PET/MRI detected only one lesion in one patient presenting bone involvement. 68 Ga-PSMA and 68 Ga-DOTA-RM2 68 68 Ga-PSMA and 68 Ga-DOTA-RM2 PET/MRI findings were concordant in 12/31 patients (6 patients with the same positive findings on both examinations, 6 patients with negative examinations; Figure 1 for an exemplar image), partially concordant in 8/31 patients and discordant in 11/31 patients. Specifically, regarding the discordant findings, 68 Ga-PSMA detected 7 lesions in 5 patients having local recurrences, 5 lesions in 5 patients experiencing bone disease, 4 lesions in 3 patients with lymph nodal recurrence and 2 lesions in one patient with visceral metastases. Conversely, 68 Ga-DOTA-RM2 PET/MRI detected only one lesion in one patient presenting bone involvement. Among the 12 concordant cases, 6/12 were confirmed as true positive and 4/12 true negative, whilst 2/12 patients were found to be false negative. Of note, regarding patient n. 9 who resulted to be true positive, further examination on conventional imaging confirmed only a minority of lesions detected by PET with both radiotracers.

PET/MRI Findings and Comparison between
Among patients presenting partially concordant uptake findings of the two radiotracers, in only one subject 68 Ga-DOTA-RM2 detected more lesions compared to 68 Ga-PSMA; notably, these additional findings were not confirmed by available evidence gathered in the follow-up period. Conversely, in 5/7 patients for whom 68 Ga-PSMA identified more locations of PCa recurrence than 68 Ga-DOTA-RM2, the results were confirmed by follow-up examination.
Among the 11/31 discordant findings, 10/11 were 68 Ga-PSMA positive and 68 Ga-DOTA-RM2 negative (Figures 2 and 3 for exemplar images). Among the 12 concordant cases, 6/12 were confirmed as true positive and 4/12 true negative, whilst 2/12 patients were found to be false negative. Of note, regarding patient n. 9 who resulted to be true positive, further examination on conventional imaging confirmed only a minority of lesions detected by PET with both radiotracers.
Among patients presenting partially concordant uptake findings of the two radiotracers, in only one subject 68 Ga-DOTA-RM2 detected more lesions compared to 68 Ga-PSMA; notably, these additional findings were not confirmed by available evidence gathered in the follow-up period. Conversely, in 5/7 patients for whom 68 Ga-PSMA identified more locations of PCa recurrence than 68 Ga-DOTA-RM2, the results were confirmed by follow-up examination.
Among the 11/31 discordant findings, 10/11 were 68 Ga-PSMA positive and 68 Ga-DOTA-RM2 negative (Figures 2 and 3 for exemplar images).    Nine/11 of these patients were confirmed to be true positive by follow-up examination, while 1/11 68 Ga-PSMA positive finding was not supported by further evidence, thus  Table 2 for detailed information on the follow-up examination for each patient. Evidence of disease on conventional imaging at baseline (vesical-urethral anastomosis and right iliac bone) and increase in PSA level during follow-up   68 Ga-DOTA-RM2. Fifteen lesions in 3/31 patients were visible only on 68 Ga-DOTA-RM2 PET/MRI. The patient-based detection rates (Figure 4a) and the average number of detected lesions (Figure 5a) were significantly different between the two imaging modalities (p = 0.016 and 0.002). The detection rates of these radiotracers stratified by PSA serum concentration, PSA DT and GS prior to radical treatment are depicted in Figures 4b, 4c and 4d, respectively. Moreover, in Figure 5b is shown the average number of lesions identified by dual-tracer PET stratified by site of recurrence.

Associations between Semi-Quantitative Imaging Parameters and Clinical Data
Correlations between clinical data and 68 Ga-PSMA and 68 Ga-DOTA-RM2 detection rates on a patient basis are reported in Table 3. PSA level and GS were not significantly associated with 68 Ga-PSMA PET/MRI detection rate, while patients with higher PSA DT, were less likely to be 68 Ga-PSMA PET positive compared to those showing PSA DT < 6 months (OR < 0.01, p = 0.022). However, this association was lost after correction for multiple testing (adjusted p = 0.065). None of the investigated clinical variables was significantly associated with 68 Ga-DOTA-RM2 PET/MRI detection rate on a patient basis. different between the two imaging modalities (p = 0.016 and 0.002). The detection rates of these radiotracers stratified by PSA serum concentration, PSA DT and GS prior to radical treatment are depicted in Figure 4b, Figure 4c and Figure 4d, respectively. Moreover, in Figure 5b is shown the average number of lesions identified by dual-tracer PET stratified by site of recurrence.

Associations between Semi-Quantitative Imaging Parameters and Clinical Data
Correlations between clinical data and 68 Ga-PSMA and 68 Ga-DOTA-RM2 detection rates on a patient basis are reported in Table 3. PSA level and GS were not significantly associated with 68 Ga-PSMA PET/MRI detection rate, while patients with higher PSA DT, were less likely to be 68 Ga-PSMA PET positive compared to those showing PSA DT < 6    Table 4 reports the number of lesions detected by PET/MRI stratified according to GS, PSA levels and kinetic.     A significant difference in the number of lesions detected by 68 Ga-PSMA PET/MRI was observed in patients with PSA value < 0.5 ng/mL vs. PSA value ≥ 2 ng/mL (Mann-Whitney U = 21.5, p = 0.001, Figure 7a); this difference remained significant after correction for multiple testing (adjusted p = 0.006). A significant difference was also observed in patients with PSA DT < 6 months vs. PSA DT ≥ 6 months (Mann-Whitney U = 119, adjusted p = 0.044; Figure 7c).

Discussion
The present study investigated the diagnostic performance of 68 Ga-PSMA and 68 Ga-DOTA-RM2 PET/MRI for detection of recurrent PCa after primary treatment. Furthermore, we also studied the association between 68 Ga-PSMA and 68 Ga-DOTA-RM2 PET uptake and clinical as well as histopathological data.
So far, only few studies have investigated the combined use of both 68 Ga-PSMA and 68 Ga-DOTA-RM2 PET in PCa both in the staging and restaging setting, using PET/MRI and PET/CT alternatively [10,[12][13][14].
Recently, Baratto et al. conducted the largest study to date investigating the use of 68 Ga-PSMA and 68 Ga-DOTA-RM2 in patients with recurrent PCa [10]. However, despite the lower number of patients included in the analysis, our study presents several methodological advantages. In fact, in the present clinical trial, all patients have been examined with both 68 Ga-PSMA and 68 Ga-DOTA-RM2 PET/MRI, thus not alternatively using either 68 Ga-PSMA or 18 F-PSMA as previous groups have done; patients' recruitment and data analysis have been prospectively performed and all findings were validated by clinical and/or instrumental follow-up.
In our cohort, 68 Ga-PSMA identified PCa recurrent lesions in 74% of patients, while 68 Ga-DOTA-RM2 PET/MRI was positive in only 48% of patients. Interestingly, 9/10 of the patients presenting 68 Ga-PSMA, but not 68 Ga-DOTA RM2 uptake were confirmed to be true positive by follow-up examination. This is in contrast with previous findings where PSMA and 68 Ga-DOTA-RM2 PET had similar detection rates on a patient basis [10].
On a lesion-based analysis, our results are coherent with the ones reported by Baratto et al., which directly compared the performances of 68 Ga-PSMA and 68 Ga-DOTA-RM2 PET and found that PSMA detected 36 additional lesions in 13 patients, with 66 additional lesions in 17 patients observed in our study.
The ability of PET to detect BCR at low PSA levels allows a personalized treatment planning at an early stage of recurrent disease by permitting early local therapies with curative intent. Therefore, it is particularly relevant to investigate the performance of different PET tracers by stratifying their detection rate according to PSA concentration. For this purpose, most studies divided their sample according to PSA in 5 sub-groups (<0.5, 0.5-1, 1-2, 2-5, >5) [10,22,23]. Differently from previous published data, in the present work, in order to allow statistical analysis, the cohort was stratified according to PSA level Similarly, higher PSA serum levels were associated with more PET positive lesions in 68 Ga-DOTA-RM2 PET/MRI, but significance was lost after correction for multiple testing (Mann-Whitney U = 46, adjusted p = 0.17; Figure 7b). The number of PET positive lesions did not differ significantly across GS levels both on 68 Ga-PSMA and 68 Ga-DOTA-RM2 PET/MRI (Mann-Whitney U = 56, adjusted p = 0.43 and U = 66.5, adjusted p = 0.65, respectively; Figure 7d).

Discussion
The present study investigated the diagnostic performance of 68 Ga-PSMA and 68 Ga-DOTA-RM2 PET/MRI for detection of recurrent PCa after primary treatment. Furthermore, we also studied the association between 68 Ga-PSMA and 68 Ga-DOTA-RM2 PET uptake and clinical as well as histopathological data.
So far, only few studies have investigated the combined use of both 68 Ga-PSMA and 68 Ga-DOTA-RM2 PET in PCa both in the staging and restaging setting, using PET/MRI and PET/CT alternatively [10,[12][13][14].
Recently, Baratto et al. conducted the largest study to date investigating the use of 68 Ga-PSMA and 68 Ga-DOTA-RM2 in patients with recurrent PCa [10]. However, despite the lower number of patients included in the analysis, our study presents several methodological advantages. In fact, in the present clinical trial, all patients have been examined with both 68 Ga-PSMA and 68 Ga-DOTA-RM2 PET/MRI, thus not alternatively using either 68 Ga-PSMA or 18 F-PSMA as previous groups have done; patients' recruitment and data analysis have been prospectively performed and all findings were validated by clinical and/or instrumental follow-up.
In our cohort, 68 Ga-PSMA identified PCa recurrent lesions in 74% of patients, while 68 Ga-DOTA-RM2 PET/MRI was positive in only 48% of patients. Interestingly, 9/10 of the patients presenting 68 Ga-PSMA, but not 68 Ga-DOTA RM2 uptake were confirmed to be true positive by follow-up examination. This is in contrast with previous findings where PSMA and 68 Ga-DOTA-RM2 PET had similar detection rates on a patient basis [10].
On a lesion-based analysis, our results are coherent with the ones reported by Baratto et al., which directly compared the performances of 68 Ga-PSMA and 68 Ga-DOTA-RM2 PET and found that PSMA detected 36 additional lesions in 13 patients, with 66 additional lesions in 17 patients observed in our study.
The ability of PET to detect BCR at low PSA levels allows a personalized treatment planning at an early stage of recurrent disease by permitting early local therapies with curative intent. Therefore, it is particularly relevant to investigate the performance of different PET tracers by stratifying their detection rate according to PSA concentration. For this purpose, most studies divided their sample according to PSA in 5 sub-groups (<0.5, 0.5-1, 1-2, 2-5, >5) [10,22,23]. Differently from previous published data, in the present work, in order to allow statistical analysis, the cohort was stratified according to PSA level in 3 groups (<0.5, 0.5-2, >2) because of the numerosity of the cohort. The missing association between 68 Ga-PSMA PET/MRI positivity rate and PSA level found in our cohort is likely due to the limited numerosity of the investigated sample combined with the high detection rate of 68 Ga-PSMA for low PSA values (66.7% for PSA < 0.5 ng/mL). These findings are in line with previous works assessing the performance of PSMA in detecting recurrent PCa in larger cohorts [24][25][26], but higher compared to the detection rates reported in similar settings [10,22,27,28].
However, similarly to previous studies [22,27], we observed that patients presenting PSA concentration ≥2 ng/mL at the time of scans showed a significantly higher number of 68 Ga-PSMA PET/MRI positive lesions compared to patients with PSA level <0.5 ng/mL.
Although the association between 68 Ga-DOTA-RM2 PET/MRI positivity rate and PSA concentration both on a patient and lesion basis was not significant, we observed a trend towards a higher rate of 68 Ga-DOTA-RM2 PET/MRI positive findings at higher PSA values, confirmed also by ROC analysis, that will surely be an interesting matter of study in future investigations.
The link between PSA kinetics and 68 Ga-PSMA PET findings is well established, with patients having slow PSA kinetics being less likely to be 68 Ga-PSMA PET positive [29,30]. Accordingly, we reported that patients having a PSA DT ≥ 6 months present a significantly lower number of lesions compared to those having a faster rise in PSA levels after primary treatment; we observed the same trend on a patient-based analysis, although after adjustment for multiple testing, the p-value only approached the level of significance (p = 0.065). Furthermore, ROC analysis showed that PSA DT has a moderate predictive value in the identification of 68 Ga-PSMA PET positivity, in line with the correlation previously investigated. Conversely, we did not detect any association between 68 Ga-DOTA-RM2 PET/MRI positivity rate and PSA kinetics.
Currently, the possible association between GS of the primary tumour and 68 Ga-PSMA PET positive findings in recurrent PCa has not been definitively demonstrated, with different studies reporting conflicting results [10,24]. In our study, we did not observe any association between the positivity rate of PET/MRI with either radiotracer and GS.
A constraint of the present study might be the relatively limited number of included patients. However, as mentioned above, our cohort is remarkably homogeneous as all patients have been examined on the same scanner and applying the same methodology (all PET/MRI studies and not both PET/MRI and PET/CT; exclusive use of 68 Ga-PSMA and 68 Ga-DOTA-RM2).
Another limitation might be the lack of post-hoc correlation of imaging findings with histopathological results for all patients; however, we consider this a relative limitation related to patients' disease management, as a number of findings detected on PET/MRI are located in sites that are not usually considered for bioptic confirmation (i.e., bone) and/or are directly treated without any further confirmation (i.e., radiotherapy on positive lymphnodes, ADT). Additionally, all PET findings were validated using clinical and instrumental follow-up. All images were acquired using the same Signa 3 Tesla PET/MRI scanner, but since this study is part of a single-centre prospective trial that is still ongoing at IRCCS San Raffaele Scientific Institute, and it was specifically focused on the use of 68 Ga-PSMA and 68 Ga-DOTA-RM2 PET in recurrent PCa, MRI data were not considered for the purpose of this work and therefore were not included in the analysis. The inclusion of MR data will surely constitute an improvement of the present work and will be a matter of investigation in future studies.
Finally, the two PET examinations were read separately, with the Nuclear Medicine physicians not being blind to the clinical history of the patients, thus resulting in a potential bias.

Conclusions
The possibility to identify different sites of recurrence of PCa after radical treatment by using a dual-tracer approach certainly improves the disease characterization and therefore it may ultimately have an impact on patients' management and follow-up. Although in our cohort, 68 Ga-PSMA had a higher detection rate than 68 Ga-DOTA-RM2 in detecting recurrent PCa. Furthermore, the number of 68 Ga-PSMA PET/MRI positive lesions was associated with higher levels of PSA and a shorter PSA DT, thus representing potential prognostic factors. GS, on the other hand, was not related to PET findings.