[18F]Fluciclovine PET/CT Improves the Clinical Management of Early Recurrence Prostate Cancer Patients

Simple Summary In the challenge between increasingly sensitive PET radiopharmaceuticals for the evaluation of prostate cancer patient in biochemical relapse, the choice of the most accurate PET tracer must be guided by literature data, but above all tailored to the patient’s profile. In describing our single-center experience, we aimed to identify biochemical and clinical–histological factors to be considered in patient selection and the semiquantitative parameters that can help the interpretation of malignant from benign lesions, in order to optimize the performance of this imaging method. These data in combination with a significant impact on therapeutic decision making can be useful to further validate the [18F]Fluciclovine PET/CT clinical application. Abstract We investigated the [18F]Fluciclovine PET/CT reliability in the early detection of recurrent prostate cancer (PCa) and its impact on therapeutic decision making. We retrospectively analyzed 58 [18F]Fluciclovine PET/CT scans performed to identify early PCa recurrence. Detection rate (DR) and semiquantitative analysis were evaluated in relation to biochemical and clinical–histological features. Clinical follow-up data were collected and considered as gold standard to evaluate sensitivity, specificity, accuracy, positive and negative predictive value (PPV, NPV). The impact of [18F]Fluciclovine PET/CT on clinical management was also assessed. Overall DR resulted as 66%, while DR was 53%, 28%, and 7% in prostate/bed, lymph nodes, and bone, respectively. DR significantly increased with higher PSA values (p = 0.009) and 0.45 ng/mL was identified as the optimal cut-off value. Moreover, SUVmax and SUVmean resulted significant parameters in interpreting malignant from benign findings. [18F]Fluciclovine PET/CT reached a sensitivity, specificity, PPV, NPV, and accuracy of 87.10%, 80.00%, 87.10%, 80.00%, and 84.31%, respectively. Therapeutic strategy was changed in 51% of patients. Our results support [18F]Fluciclovine PET/CT as a reliable tool for early restaging of PCa patients, especially for local recurrence detection, leading to a significant impact on clinical management. Semiquantitative analysis could improve specificity in interpreting malignant from benign lesions.


Introduction
Despite the improvement in surgical and radiation techniques for clinically localized prostate cancer (PCa), between 27% and 53% of all patients experience a rising prostatespecific antigen (PSA) level after curative-intent primary treatment [1].
Nowadays, PCa relapse remains a diagnostic and therapeutic dilemma, since conventional imaging, including contrast-enhanced computed tomography (CT), Magnetic

Imaging Protocol and Analysis
[ 18 F]Fluciclovine PET/CT was performed according to standard protocols. Patients did not undertake any significant exercise for the day before the exam, fasted for at least 4 h before imaging, and were asked not to void 30-60 min prior to injection [15]. All scans were obtained using a hybrid PET/CT scanner (Discovery 710, GE, General Electrics, Milwaukee, Cancers 2022, 14, 1461 4 of 18 WI, USA). Immediately after the intravenous [ 18 F]Fluciclovine administration (370 MBq), the CT scan was performed, followed by the PET scan at 3-5 min post-injection from the mid-thigh to the base of the skull (5-6 bed positions). For clinical reasons, the scan was extended to the top of the skull. A 3D acquisition mode PET scan for the same longitudinal coverage, 2.5 min per bed position, was performed. Coregistered CT parameters were: pitch 0.98, gantry rotation speed of 0.5 s/rot, 120 kVp, and modulated tube current of 140 mA. CT images were used both for image fusion and anatomical localization and for attenuation correction of emission data.
Image analysis was carried out using a dedicated console (AW Server 4.7, General Electrics, Milwaukee, WI, USA). All [ 18 F]Fluciclovine PET/CT scans were assessed visually, using Maximum Intensity Projection (MIP), transaxial, sagittal and coronal images, and interpreted in consensus by two nuclear medicine physicians (C.F. and V.L.) with at least 3 years' experience in [ 18 F]Fluciclovine PET/CT reading and awareness of clinical data.
Examinations were considered positive if there was any [ 18 F]Fluciclovine uptake visually clearly higher than the surrounding background activity that did not correlate with physiological tracer uptake. Bone marrow uptake of vertebra L3 and abdominal aortic blood pool were used as references for lesions larger than 1 cm longest dimension or lesions smaller than 1 cm longest dimension, respectively [16]. In the case of doubtful lesions, a third nuclear medicine physician (A.G.N.) was consulted, and the final diagnosis was reached in consensus.

[ 18 F]Fluciclovine PET/CT Detection Rate
[ 18 F]Fluciclovine PET/CT-positive findings were reported as "detection rate" (DR) since histological confirmation was not available or feasible.
The DR on the per-patient and the per-lesion (prostate/prostate bed, lymph node, bone) analyses were investigated in our cohort in relation to biochemical parameters (PSA value < 0.5, 0.5-1, or >1 ng/mL; PSA dt < 12 or ≥12 months), Gleason Score (GS < 8, GS ≥ 8), EAU BCR risk group (low-and high-risk), the interval time from primary treatment to PSA relapse (TTR), primary treatment (RP alone, RT alone or combined), and ongoing hormonal therapy (yes vs. no). The PSA and TTR cut-off values predictive of a positive and negative scan were evaluated.

Follow-Up Data Analysis
At least 3 months clinical follow-up data were collected for the analysis, which included post-PET PSA values, diagnostic investigations, and start/change treatment. These data were considered as gold standard in the evaluation of sensitivity, specificity, accuracy, positive and negative predictive value (PPV, NPV) of [ 18  All positive findings were analyzed by calculating the following PET-based semiquantitative and volumetric parameters: maximum Standardized Uptake Value (SUVmax), mean Standardized Uptake Value (SUVmean), Metabolic Tumor Volume (MTV), Total Lesion Activity (TLA, defined as SUVmean × MTV), Tumor-to-Background ratio (T/Bratio), resulting in the ratio of the SUVmax of the lesion to the SUVmean of abdominal aorta or bone marrow as background.
The reference lesion (RL) was identified as the highest [ 18 F]Fluciclovine uptake lesion and considered for the per-patient analysis. Per-lesion analysis included the highest uptake lesion of prostate/prostatectomy bed, lymph nodes and bone, respectively.
The different distribution of the semiquantitative PET parameters was evaluated in our sample stratified according to biochemical (PSA level, PSAdt) and clinical-histological features (GS, EAU BCR risk group, TTR, primary and ongoing treatments).

Subpopulation Analysis
A further analysis was conducted dividing our sample into subpopulations, based on PET-disease extent, as follows: exclusively prostate/prostatectomy bed disease or extraprostatic progression, with nodal and/or skeletal involvement. In addition, oligometastatic disease was defined as the presence of one to three lesions regardless of site, while polymetastatic as the presence of four or more lesions [17].
The different distribution of biochemical (PSA level, PSAdt) and clinical-histological (GS, EAU BCR risk group, TTR, primary and ongoing treatments) variables and semiquantitative parameters was evaluated in these subpopulations.

Statistical Analysis
Quantitative variables were expressed as mean with standard deviations (SD) and/or as median with range. Categorical variables were presented with absolute and relative frequencies. The Chi-squared and Fisher exact test were employed to analyze differences in categorical variables: PSA value (<0.5, 0.5-1, >1), PSA dt (<12 months, ≥12 months), GS (<8 vs. ≥8), EAU BCR risk group (low-, high-risk), primary treatment (RP alone, RT alone, combined RP+RT) and ongoing hormonal therapy (yes vs. no). The Mann-Whitney U test and Pearson correlation test were used to compare differences between continuous nonnormally distributed variables (PSA level, TTR, semiquantitative parameters).
The performance of [ 18 F]Fluciclovine PET/CT in relation to the continuous variables was assessed by receiving operating characteristic (ROC) curves generated by plotting sensitivity versus 1− specificity. The best cut-off values of continuous variables for predicting between positive and negative [ 18 F]Fluciclovine PET/CT scan and for distinguishing between malignant from benign findings were determined using Youden's index. Statistical significance was assumed for p values less than 0.05.
All statistical analyses were performed using SPSS statistical software, version 28 (IBM Corporation, Armonk, NY, USA).
The per-patient semiquantitative analysis showed no significant difference neither according to biochemical variables nor to clinical-histological ones (p > 0.05). However, a trend towards increased SUVmaxRL and SUVmeanRL with higher PSA levels (especially greater than 1.0 ng/mL) was observed, as presented in Table 3. This trend was further investigated with Pearson correlation, confirming a statistically significant linear correlation between PSA level and SUVmaxRL (p = 0.006) and SUVmeanRL (p = 0.003).
The per-patient semiquantitative analysis showed no significant difference neither according to biochemical variables nor to clinical-histological ones (p > 0.05). However, a trend towards increased SUVmax RL and SUVmean RL with higher PSA levels (especially greater than 1.0 ng/mL) was observed, as presented in Table 3. This trend was further investigated with Pearson correlation, confirming a statistically significant linear correlation between PSA level and SUVmax RL (p = 0.006) and SUVmean RL (p = 0.003).
The type of curative-intent treatment showed a significantly difference in the two subpopulations. Particularly, patients treated with combined prostatectomy and radiotherapy developed more likely extraprostatic extension on [ 18 F]Fluciclovine PET/CT scan (p = 0.005) compared with only local relapse. No other significant differences were found between two subpopulations in relation to other biochemical and clinical-histological factors included in the analysis, as shown in Table 4. As regards semiquantitative evaluation, a significantly higher T/Bratio RL value (p = 0.014) was observed in the extraprostatic disease subpopulation (Table 5). ROC analysis identified a T/Bratio RL optimal cut-off value of 1.12 (sensitivity 88%; specificity 43%) with an AUC of 0.732 (95% CI 0.573-0.892), associated with more likely extraprostatic disease ( Figure 6).

Discussion
Molecular imaging is increasingly integrated in the diagnostic work-up of PCa patients in PSA relapse, yielding higher diagnostic performance compared to conventional imaging and facilitating earlier localization of recurrent disease [18]. For this purpose, new and more sensitive radiopharmaceuticals are challenging in nuclear medicine scenarios, even if there is still no consensus on the optimal imaging modality in the case of low PSA levels.
Current EAU guidelines recommend to perform PSMA-labeled PET/CT for imaging PCa with biochemical recurrence if PSA level is greater than 0.2 ng/mL and if results will influence subsequent treatment decisions; whilst [ 18 F]Fluciclovine or Choline-radiolabeled PET/CT are recommended in the case of PSMA-labeled PET/CT is not available and PSA is greater than 1 ng/mL [1]. On the other hand, National Comprehensive Cancer Network guidelines do not indicate a reference PSA value, suggesting [ 18 F]Fluciclovine PET/CT after conventional imaging modalities for further evaluation of equivocal findings [19].
Identifying the most suitable radiopharmaceutical based on the PCa patient characteristics, as well as available in clinical practice, would ensure a personalized diagnostic approach for guiding therapeutic decision making. For this purpose, we investigated the biochemical and clinical-histological parameters that could impact on the [ 18 F]Fluciclovine PET/CT positivity rate in a cohort of PCa patients with low PSA values in order to optimize its diagnostic performance and patient selection.

Discussion
Molecular imaging is increasingly integrated in the diagnostic work-up of PCa patients in PSA relapse, yielding higher diagnostic performance compared to conventional imaging and facilitating earlier localization of recurrent disease [18]. For this purpose, new and more sensitive radiopharmaceuticals are challenging in nuclear medicine scenarios, even if there is still no consensus on the optimal imaging modality in the case of low PSA levels.
Current EAU guidelines recommend to perform PSMA-labeled PET/CT for imaging PCa with biochemical recurrence if PSA level is greater than 0.2 ng/mL and if results will influence subsequent treatment decisions; whilst [ 18 F]Fluciclovine or Choline-radiolabeled PET/CT are recommended in the case of PSMA-labeled PET/CT is not available and PSA is greater than 1 ng/mL [1]. On the other hand, National Comprehensive Cancer Network guidelines do not indicate a reference PSA value, suggesting [ 18 F]Fluciclovine PET/CT after conventional imaging modalities for further evaluation of equivocal findings [19].
Identifying the most suitable radiopharmaceutical based on the PCa patient characteristics, as well as available in clinical practice, would ensure a personalized diagnostic approach for guiding therapeutic decision making. For this purpose, we investigated the biochemical and clinical-histological parameters that could impact on the [ 18 F]Fluciclovine PET/CT positivity rate in a cohort of PCa patients with low PSA values in order to optimize its diagnostic performance and patient selection.
The impact of PSA value on [ 18 F]Fluciclovine PET/CT DR is already well-known in the literature. A multicenter study enrolling 600 PCa patients demonstrated an overall positivity rate of 67.7% (38.7% in prostate/prostatectomy bed, 32.6% in lymph nodes, 26.2% in extrapelvic sites), and of 41.4% in patients with PSA less than 0.79 ng/mL [20]. In a prospective multicenter study by Scarsbrook et al., an overall DR of 56% was registered in 104 recurrent PCa patients and of 33% among patients with pre-scan PSA less than 0.2 ng/mL, suggesting a good performance also at very low PSA levels [7]. Higher positivity rate (65%) resulted in the Dreyfuss et al. retrospective study with 328 PCa patients, preserving an optimal DR (58%) also among 26 patients with pre-scan PSA value less then 0.2 ng/mL [21]. Consistent with literature, our results confirmed an overall DR of 66% with the pre-scan PSA value as the main predictive parameter of a positive exam, with increasing DR in PSA greater than 0.5 ng/mL (40% in PSA < 0.5 ng/mL; 87% in PSA = 0.5-1 ng/mL; 74% in PSA > 1 ng/mL), especially in prostate bed.
However, data regarding the most reliable PSA threshold for recurrence detection on imaging are conflicting. Several PSA values, variable from 0.3 to 1 ng/mL, are suggested in the literature as a reliable threshold for a [ 18 F]Fluciclovine positive scan [22,23]. In our analysis a PSA cut-off value of 0.45 ng/mL was identified as adequate for patient selection.
Most of positive findings were observed in prostate/prostatectomy bed (53%), thanks to irrelevant urinary interference, compared to lymph node (28%) and bone involvement (7%). Differently, the Bulbul et al. study reported a similar prostatic and extraprostatic detection rate of 35% and 37% in patients with PSA less than 1 ng/mL [11].
In relation to cancer biological behavior, a higher [ 18 F]Fluciclovine PET/CT DR was documented in more aggressive PCa [24,25], particularly in case of GS≥8, T3-T4 stages, and castration-resistant neoplasm [26]. To the best of our knowledge, this is the first study thar investigated the role of TTR in [ 18 F]Fluciclovine PET/CT DR, already evaluated for other radiopharmaceuticals [17]. In our analysis, TTR resulted as the only clinical parameter related to tumor aggressiveness, showing a significant inverse correlation with DR, with an optimal cut-off of 20 months. Differently, statistical significance was not reached for the recently validated EAU BCR risk groups [15], investigated only by Selnaes et al., who reported a lower [ 18 F]Fluciclovine PET/MR DR in the EAU low-risk BCR group [27].
Type of primary treatment with curative intention could influence the site of disease recurrence observed on [ 18 F]Fluciclovine PET scan. Across all patients enrolled in the FALCON trial, DR was significantly lower in those who underwent prostatectomy (32%) compared with those with an intact prostate (95%), due to prostate/bed findings [18]. Our analysis demonstrated that patients treated with combined surgical and irradiation therapy developed more likely extraprostatic progression (p = 0.005). We can speculate that combined therapy reaches a better prostatic disease control rather than extraprostatic one while an inadequate PCa staging performed with poor sensitive imaging methods could cause primary treatment failure with disease recurrence. Furthermore, consistent with preclinical results, our data demonstrated the absence of a negative impact of ADT on [ 18 F]Fluciclovine PET/CT positivity rate [28].
According to the EANM procedure guidelines [16], the interpretation of [ 18 F]Fluciclovine PET images should be based on the visual analysis. Although in the absence of standardized criteria, the semi-quantitative analysis could be helpful in the interpretation of PET images, especially in doubtful or equivocal cases. Zanoni et al. defined SUVmax as an imaging biomarker predictor of disease progression, being significantly correlated with biological aggressiveness of prostate cancer [29]. Confirming this data, our results demonstrated a statistically significant linear correlation between SUVmax RL (p = 0.006) and SUVmean RL (p = 0.003) and PSA level.
However, the role of SUVmax appears controversial in differentiating malignant from benign lesions, thus compromising the [ 18 F]Fluciclovine PET/CT specificity [29]. In the same study, the measuring of T/Bratio-AORTA and T/Bratio-L3 significantly improved the specificity and sensitivity of this imaging method with the optimal cut-off values of 2.7-3.75 and 1.35-1.55, respectively [29]. In our cohort, SUVmax and SUVmean resulted helpful in interpreting malignant/benign lesions, with optimal cut-off values of 2.05 and 1.75, respectively. In addition, higher T/Bratio values were found in extraprostatic, more advanced, disease. However, further study is needed to validate the role of semiquantitative analysis and identify standardized cut-off values in PET images' interpretation.
Consistent with the LOCATE [9] and FALCON trials [7], a significant change (51%) in therapeutic strategy was observed after the [ 18 F]Fluciclovine PET/CT scan. In the LOCATE and FALCON trials, the most common pre-scan management was salvage radiotherapy. In the LOCATE study, the treatment strategy was frequently changed from salvage radiotherapy or systemic therapy to watchful waiting in [ 18 F]Fluciclovine-negative patients (25%), while in positive ones from ADT to radiotherapy (24%), and less commonly the contrary (9%) [9]. In the FALCON trial, salvage radiotherapy was frequently changed to systemic therapy (24%) after a positive scan or similarly to watchful waiting after a negative one (24%) [7]. In our analysis, the most frequent changes were from clinical surveillance to systemic hormone therapy (38%) or radiotherapy (31%) following a positive scan, avoiding an undertreatment.
The clinical management is closely influenced by an accurate imaging definition of the real disease extent and particularly by the emerging importance to differentiate oligometastatic from polymetastatic disease. While current data suggest promising sensitivity of PSMA-based PET radiopharmaceuticals, the growing evidence for the good performance of [ 18 F]Fluciclovine at low PSA levels indicates that amino acid tracers should play an important role for the imaging of oligometastatic disease [14,17]. In our cohort, a great percentage (95%) of positive patients was defined as oligometastatic disease, with influence on therapeutic decision making.
Regarding [ 18 F]Fluciclovine PET/CT performance in detecting PCa recurrence, literature reports heterogeneous data [30]. A recent meta-analysis registered a pooled sensitivity and specificity of 86.3% and 75.9%, respectively [25]. On the other hand, a lower specificity of 40% was documented in the Schuster et al. analysis with a higher sensitivity of 90.2% [31]. Similar diagnostic performance was found in a larger cohort of 596 PCa patients (sensitivity 88.1%, specificity 32.6%) [20]. Our data confirmed a good [ 18 F]Fluciclovine PET/CT sensitivity (87%); however, a higher specificity (80%) emerged in our study. The lack of tissue sampling as gold standard in our analysis could have impacted on this result.
In the nuclear medicine challenge between more sensitive PET radiopharmaceuticals, the trump card remains to ensure an increasingly personalized diagnostic approach to guide a tailored therapeutic choice. Even if the comparison of the different radiopharmaceuticals does not fit the goal of this study, this issue deserves a consideration. In the last 5 years, literature provided comparison data between PET tracers. [ 18 F]Fluciclovine demonstrated a greater detection rate compared to 18 F/ 11 C-Choline radiopharmaceuticals, that remains suboptimal especially for low PSA levels [6,32,33]. Conversely, the head-to-head comparisons between [ 18 F]Fluciclovine and 68 Ga-PSMA PET/CT resulted controversial [34]. Pernhtaler et al. registered a more accurate detection of local recurrence of [ 18 F]Fluciclovine than 68Ga-PSMA (37.9% vs. 27.6%), especially when located in close relation to the urinary bladder [10]; while, Calais et al. suggested PSMA as the tracer of choice for patients in BCR after radical prostatectomy, since an overall significantly higher detection of PSMA-PET was observed (56% vs. 26%) [35]. On the other hand, it should be noted that limited availability of PSMA-labeled tracer restricted its clinical application, unlike [ 18 F]Fluciclovine is more accessible in the US and Europe [6]. Consequently, the best PET target remains a dilemma and it needs to be further investigated. Knowing strengths and weaknesses of each radiopharmaceutical could help to choose the best imaging method, optimized on PCa patient profile, in order to improve the clinical management.
There are some limitations to the present analysis: first, the retrospective nature and the small sample size of this single-center study; second, the lack of tissue sampling as gold standard to assess the diagnostic accuracy of this imaging technique; lastly, the need of long-term follow-up data to investigate patients' outcomes after PET-induced change in therapeutic strategy.

Conclusions
[ 18 F]Fluciclovine PET/CT is a reliable tool for early restaging PCa patients, with an optimal PSA cut-off value of 0.45 ng/mL to be considered for patient selection. The signifi-cant [ 18 F]Fluciclovine detection of recurrent lesions (66%) at an earlier time, especially in prostate bed (53%), due to its irrelevant urinary interference, lead to a significant impact on clinical management (51%). Semiquantitative analysis, especially SUVmax and SUVmean, could improve specificity in helping the interpretation of malignant from benign lesions. Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.