The Diagnostic Role of 18F-Choline, 18F-Fluciclovine and 18F-PSMA PET/CT in the Detection of Prostate Cancer With Biochemical Recurrence: A Meta-Analysis

Background Diagnosing the biochemical recurrence (BCR) of prostate cancer (PCa) is a clinical challenge, and early detection of BCR can help patients receive optimal treatment. We conducted a meta-analysis to define the diagnostic accuracy of PET/CT using 18F-labeled choline, fluciclovine, and prostate-specific membrane antigen (PSMA) in patients with BCR. Methods Multiple databases were searched until March 30, 2021. We included studies investigating the diagnostic accuracy of 18F-choline, 18F-fluciclovine, and 18F-PSMA PET/CT in patients with BCR. The pooled sensitivity, specificity, and detection rate of 18F-labeled tracers were calculated with a random-effects model. Results A total of 46 studies met the included criteria; 17, 16, and 13 studies focused on 18F-choline, fluciclovine, and PSMA, respectively. The pooled sensitivities of 18F-choline and 18F-fluciclovine were 0.93 (95% CI, 0.85–0.98) and 0.80 (95% CI, 0.65–0.897), and the specificities were 0.91 (95% CI, 0.73–0.97) and 0.66 (95% CI, 0.50–0.79), respectively. The pooled detection rates of 18F-labeled choline, fluciclovine and PSMA were 66, 74, and 83%, respectively. Moreover, the detection rates of 18F-labeled choline, fluciclovine, and PSMA were 35, 23, and 58% for a PSA level less than 0.5 ng/ml; 41, 46, and 75% for a PSA level of 0.5–0.99 ng/ml; 62, 57, and 86% for a PSA level of 1.0–1.99 ng/ml; 80, 92, and 94% for a PSA level more than 2.0 ng/ml. Conclusion These three 18F-labeled tracers are promising for detecting BCR in prostate cancer patients, with 18F-choline showing superior diagnostic accuracy. In addition, the much higher detection rates of 18F-PSMA showed its superiority over other tracers, particularly in low PSA levels. Systematic Review Registration PROSPERO, identifier CRD42020212531.


INTRODUCTION
Prostate cancer (PCa) is one of the most common malignancy in men worldwide and is also the fifth major cause of cancer-related death in men. It is estimated that over 300,000 PCa-related deaths occur in 2018 (1). In addition to its high morbidity and mortality, the recurrence and metastasis of prostate cancer are also troublesome in clinical practice (2,3).
It is challenging to detect initial recurrence and metastasis after prior treatment because of few obvious characteristics on early recurrent or metastatic lesions. PCa recurrence is usually considered when observing a rise in the serum prostate-specific antigen (PSA) level. This is regarded as biochemical recurrence (BCR) of PCa, and the definition of BCR is a serum PSA level over a threshold of 0.2 ng/ml twice after radical prostatectomy (RP) or an absolute increase in PSA level of 2 ng/ml over the lowest posttreatment PSA level after radiation therapy (RT) (4,5).
The key issue for patients with BCR is the early and correct identification of recurrent or metastatic disease, which is essential for further devising treatment strategies since treatment varies based on the presence of local recurrence, regional lymph node and distant viscera or bone metastasis (6). Conventional imaging modalities consisting of CT, bone scan, and MRI have been used for patients with advanced PCa, but their roles in detecting minimal or occult lesions are limited (7,8). These conventional imaging modalities also have low sensitivity and specificity in detecting patients with BCR, especially those with a low PSA level. According to the 2020 American Society of Clinical Oncology (ASCO) guidelines, nextgeneration imaging (NGI) such as PET/CT, PET/MRI, and whole-body MRI is recommended for use in patients with rising PSA after prior treatment when conventional imaging findings are negative (9). Radioactive tracers such as choline and fluciclovine have been used for prostate cancer staging, restaging, and treatment response evaluation. Meanwhile, prostate-specific membrane antigen (PSMA), a new radiopharmaceutical that binds to prostate cancer-specific target, has demonstrated outstanding detection rate for recurrent or metastatic lesions among patients with BCR.
Choline is an essential element of phospholipids in the cellular wall, and the increased uptake of choline means increased metabolism of the cell membrane components of malignant tumors (10). 11 C-choline was approved by the Food and Drug Administration (FDA) in 2012, but its short half-life limits its widespread use in PET/CT centers without onsite cyclotrons. Later, 18 F-labeled choline was developed, and its longer half-life has solidified 18 F-choline PET/CT as a significant imaging modality in patients with suspected PCa recurrence (11,12). 18 F-Fluciclovine (anti-1-amino-3-18 Ffluorocyclobutane-1-carboxylic acid, 18 F-FACBC), as a synthetic amino acid that is upregulated in PCa, is an option for molecular imaging in patients with BCR, which was approved by the FDA in 2016 (13,14). The main advantage of 18 Ffluciclovine is its low urinary excretion, which allows for better detection and localization of PCa recurrence in patients with rising PSA level (15). PSMA is a type 2 transmembrane protein that is more highly expressed in the prostate cancer cell membrane than in normal tissues (16)(17)(18). Therefore, PSMA has become a promising target for imaging prostate cancer (19). 68 Ga-PSMA PET/CT has been proven to improve the detection of metastatic disease and the monitoring of treatment effects in patients with PCa (20). Most recently, 68 Ga-PSMA-11, as the first PSMA PET agent, has been approved by the FDA. 18 F-labeled PSMA has also begun to be used in clinical practice, and its long half-life and high resolution in PET/CT images have further increased the detection rate of PSMA-targeted imaging in subtle or occult metastases (21,22). Moreover, 18 F-PSMA-1007 PET/ CT can differentiate local recurrence from physical uptake in the urinary bladder or ureter due to non-urinary clearance (23,24).
Some previous studies have compared the diagnostic roles of 11 C-choline, 18 F-fluciclovine, and 68 Ga-PSMA PET/CT in patients with BCR, showing that 68 Ga-PSMA PET/CT has a superior detection rate (25). Even so, there have been some clinical challenges existing in 68 Ga-PSMA PET/CT due to certain shortcomings including its short half-life, non-ideal energies and the limited availability of 68 Ga. Compared with 68 Ga and 11 C, 18 F, as a longer half-life nuclide, has many advantages such centralized production in a cyclotron facility and more favorable positron energies for imaging, thereby motivating the development of 18 F-labeled analogs. Currently, growing clinical experience has revealed the high diagnostic accuracy of some 18 F-labeled tracers in PCa patients with BCR. However, the effectiveness of 18 F-labeled choline, fluciclovine, and PSMA remains unclear because of limited number of studies. Herein, we aimed to perform a meta-analysis to review and compare the diagnostic value of 18 F-labeled choline, fluciclovine, and PSMA PET/CT imaging for detecting BCR in patients with PCa, in order to provide better creditability for clinical practice.

METHODS
The Preferred Reporting Items for Systematic Reviews and Metaanalyses (PRISMA) guidelines was used for our study (26). Our review has registered on the international prospective register of systematic reviews (PROSPERO) (CRD 42020198861). responsible for disagreement and solved the controversy between two authors through discussion.

Selection of Studies
Studies using 18 F-labeled tracers such as 18 F-choline, 18 Ffluciclovine, and 18 F-PSMA were evaluated. Studies were included according to the following criteria: (a) sample size >10; (b) patients who had evidence of BCR underwent PET/CT; (c) studies evaluating the diagnostic accuracy of 18 Flabeled tracers in prostate cancer patients with BCR; (d) histological results, imaging, or clinical follow-up as a reference standard. Studies on other tracers were not included. Abstracts, reviews, and case reports were also not included. If the studies included duplicate patients, we reviewed and included the study with the largest sample or the most recent study performed. The included studies were limited to those published in English.

Quality Assessment
The quality of included studies was critically assessed by two independent authors according to the Quality Assessment of Diagnostic Accuracy Studies-2 (QUADAS-2) tool. This tool comprises four domains (patient selection, index test, reference standard, and flow and timing), and each domain was used to assess the risk of bias. Next, applicability was also considered according to patient selection, the index test, and the reference standard.

Data Extraction
Two authors collected various parameters and outcomes from each eligible study as follows: author, country, publication year, study design, number of patients, age, pre-PET PSA level, reference criteria, scanner model, ligands, and injection dose and the detection rate as well as true positive (TP), false positive (FP), false negative (FN), and true negative (TN) PET/CT with different tracers in patients with BCR. All discrepancies were resolved by consensus and ultimately based on the decision of the third author.

Statistical Analysis
For studies reporting the diagnostic performance of 18 F-PSMA, 18 F-choline, and 18 F-fluciclovine PET/CT in patients with BCR, 2 × 2 table was used to calculate TP, FP, TN, and FN. The pooled sensitivity and specificity were calculated by a random-effects model. We developed a hierarchical summary receiver operating curve and calculated the area under the curve. We presented forest plots with 95% confidence intervals (CIs) for the sensitivity and specificity of each study. In addition, the detection rates of PET tracers were extracted and pooled using a random-effects model. If possible, subgroup analysis was considered based on different PSA serum values.
Heterogeneity within studies was evaluated using Cochran's Q test and the I² statistic (27). An I² value greater than 50% was indicative of substantial heterogeneity. The funnel plot test and Egger's test were used to assess the publication bias. All statistical analyses were performed using Stata 15.0 and RevMan 5.3. Pvalue <0.05 was considered to be statistically significant (28).

RESULTS
The flow chart demonstrates an overview of the search and selection process ( Figure 1). The initial search yielded 480 studies, of which 95 were duplicates. Subsequently, after reviewing the titles and abstracts, we excluded 238 studies for the following reasons: 170 studies were case reports, reviews, and academic meeting abstracts, 11 were basic studies, five applications in other diseases, and 52 studies used different radiotracers and imaging modalities. Of the remaining studies, 75 studies were not relevant to our aims, and most of them investigated the impact of novel PET/CT tracers in treatment management for patients with PCa or focused on evaluating metastatic disease. In addition, 26 studies did not provide sufficient information and were excluded. Thus, only 46 studies were finally included. Of these, 17 studies focused on the role of 18 F-choline PET/CT in prostate cancer patients with BCR (29)(30)(31)(32)(33)(34)(35)(36)(37)(38)(39)(40)(41)(42)(43)(44)(45). The numbers of included studies regarding 18 F-fluciclovine and 18 F-PSMA PET/CT were 16 and 13, respectively . Tables 1-3 outline the characteristics of each eligible study.

Quality Assessment
Figures 2A-C show the results of the quality assessment of each eligible study for 18 F-choline, 18 F-fluciclovine, and 18 F-PSMA, respectively. Patient selection was not considered the source of bias because all studies had qualified patient selection criteria. For the index test and reference standard, some studies did not adopt the blinding method when interpreting the positive scan of the PET/CT findings, and we rated these studies as high or unclear levels regarding the risk of bias and applicability concern. Similarly, unclear or high levels were displayed on the applicability concern of flow and timing because of the different follow-up times and multiple reference standards.

DISCUSSION
Our meta-analysis included studies investigating the diagnostic roles of three novel 18 F-labeled tracers applied in prostate cancer patients with BCR. From our study, the summary sensitivity and specificity of 18  Multiple PET/CT radiotracers have been developed and experimented in recent years, motivating the wide use of PET/ CT or PET/MRI in patients with PCa for staging, restaging, and response evaluation (75,76). 18 F-fluciclovine PET/CT showed a superior advantage over 11 C-choline PET/CT in patients with BCR and further aid guiding decision-making in regard to patients' treatment strategy (48,77). In addition, PSMA PET/ CT has shown superior diagnostic accuracy for recurrence and metastases of prostate cancer than fluciclovine and choline. A meta-analysis defined the diagnostic accuracy of PET/CT imaging using 11 C-choline, 18 F-fluciclovine, or 68 Ga-PSMA, showing that 68 Ga-PSMA PET/CT has a nearly equal sensitivity but the highest specificity among these tracers for PET/CT imaging in detecting biochemically recurrent PCa (25).
In contrast, our meta-analysis focused on only long-half radionuclides as 18 F-labeled tracers and summarized the diagnostic accuracy of 18 F-labeled choline, fluciclovine, and PSMA in detecting patients with BCR. Our study revealed that 18 F-PSMA had the highest detection rate at different PSA levels, and the detection rate was related to the PSA level. These results were consistent with another meta-analysis that compared the detection rate of biochemically recurrent PCa between PSMAtargeted radiotracers and 18 F-fluciclovine, finding that PSMAtargeted radiotracers demonstrate a greater detection rate than 18 F-fluciclovine (78). A study compared prospectively paired 18 Ffluciclovine and PSMA PET/CT scans for localizing recurrence of PCa after prostatectomy in patients with a PSA level <2.0 ng/ ml (55). They found that PSMA PET/CT showed higher detection rates and should be the tracer choice when PET/CT imaging is considered for patients with biochemical recurrence after radical prostatectomy with low PSA concentrations (≤2.0 ng/ml). The same conclusion was drawn from another prospective study paired that compared 18 F-PSMA and 18 Ffluorocholine PET/CT in patients with BCR (42). The advantage of PSMA-targeted PET/CT imaging could be attributed to the high expression of PSMA in PCa and its metastases.
In late 2020, 68 Ga-PSMA-11 became the first PSMA PET tracer to be approved by the FDA, which may facilitate widespread adaptation. Despite this, there also have been some limitations related to 68 Ga-PSMA PET/CT because of the short half-life, non-     ideal energies, and limited availability of 68 Ga, limiting its clinical application in detecting occult or metastatic lesions in the prostate bed (62,79). However, 18 F-PSMA analogs seemed to be more favorable due to their longer half-life and a higher physical spatial resolution (23), and 18 F-PSMA-1007, as a second-generation 18 Flabeled PSMA tracer, demonstrated high labeling yields, better tumor uptake, and hepatobiliary excretion, making it an ideal PSMA-target tracer for diagnostic imaging in patients with BCR (21,23). Our meta-analysis found the pooled detection rate with 18 F-PSMA of 58% for a PSA level of less than 0.5 ng/ml, 75% for a PSA level of 0.5 to 0.99 ng/ml, and 86% for a PSA level of 1.0 to 1.99 ng/ml. These detection rates are equal or higher than those in recent studies involving 68 Ga PSMA PET/CT (80,81).
Compared with FDA approval of 68 Ga-PSMA-11 in late 2020, 11 C-choline and 18 F-fluciclovine PET/CT have one temporary advantage as they have been granted FDA approval early. They were more accessible and used in the US and Europe. Many studies compared the diagnostic utility of 18 F-fluciclovine with 11 C-choline PET/CT imaging, showing a better performance in terms of lesion detection rate (48,82). A recent meta-analysis demonstrated that 18 F-fluciclovine had the similar sensitivity and detection rate compared with 11 C-choline, but lower specificity than 11 C-choline (83). Unlike the short physical half-life of 11 Ccholine, the radiofluorine of 18 F-choline provides a long physical half-life (109.8 min), allowing for centralized manufacture and distribution. These intrinsic advantages of 18 F labeling has made 18 F-choline PET/CT valuable in staging patients with PCa and detecting recurrently PCa metastases after initial treatment (33,84). There were limited studies in comparing directly to determine which imaging modality has a better diagnostic efficiency between 18 F-choline and 18 F-fluciclovine. In our meta-analysis, 18 F-choline had a higher sensitivity and specificity than 18 F-fluciclovine through assessing the summary sensitivity and specificity. 18 F-choline also has better detection rates than 18 F-fluciclovine at PSA levels under 0.5 ng/ml and 1.0-1.99 ng/ml, but the pooled detection rate of 18 F-fluciclovine was higher than that of 18 F-choline in biochemically recurrent PCa. This difference could be interpreted by different biological processes between amino acid transport and choline expression.

LIMITATIONS
There were several limitations to this study that should be mentioned. First, we only evaluated the diagnostic accuracy of both 18 F-choline and 18 F-fluciclovine PET/CT in patients with BCR, and the pooled sensitivity and specificity for 18 F-PSMA PET/CT were not feasible because of insufficient published data. Second, there were significant heterogeneities among institutions, PET/CT scanners, radiotracers, and prior treatment of patients, which increased the risk of bias and led to significant heterogeneity among 18 F-choline, 18 Ffluciclovine and 18 F-PSMA PET/CT. Third, most of the included studies were retrospective analyses, had small sample sizes, had limited reference standards, and lacked prospective, large sample, and interagent comparison studies. Fourth, there was publication bias according to Egger's test regarding the included studies of 18 F-choline,

CONCLUSION
PET/CT imaging with 18 F-choline, 18 F-fluciclovine, and 18 F-PSMA is promising in detecting prostate cancer patients with BCR. 18 F-PSMA PET/CT demonstrated a significantly higher detection rate over 18 F-choline and 18 F-fluciclovine for different PSA levels, particularly in PSA level less than 2.0 ng/ml.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/supplementary material. Further inquiries can be directed to the corresponding author.

FUNDING
This study was supported by the National Natural Science Foundation of China (81971653).