PSMA Theranostics: Review of the Current Status of PSMA-Targeted Imaging and Radioligand Therapy

Prostate-specific membrane antigen (PSMA) has been the subject of extensive investigation in the past two decades as a promising molecular target for prostate cancer (PCa). Its appealing molecular features have enabled the development of a novel diagnostic and therapeutic—thus “theranostic”—approach to PCa. There is now substantial evidence of the high sensitivity of PSMA-targeted imaging for PCa lesions and growing evidence of the therapeutic efficacy of PSMA radioligand therapy for metastatic castration-resistant prostate cancer. This article presents a broad overview of the current status of PSMA theranostics, including current evidence, potential clinical impact, and active areas of research.


Introduction
Also known as folate hydrolase I or glutamate carboxypeptidase II, prostate-specific membrane antigen (PSMA) is a type II, 750 amino acid transmembrane protein. In benign prostatic cells, it is localized to the cytoplasmic and apical side of the prostate epithelium. As malignant transformation occurs, PSMA is transferred from the cytoplasm to the luminal surface of the prostatic ducts, where it presents a large extracellular domain to ligands [1]. The biological function of PSMA remains unclear, but it is hypothesized to have a transport function because PSMA ligands are internalized through endocytosis.
PSMA ligand internalization theoretically enables specificity of synthetic PSMA radioligands for malignant prostatic tissue. Furthermore, research suggests a 100-to 1000-fold increase in PSMA expression in prostatic adenocarcinoma vs. benign prostatic tissue [2,3]. Although there is an increasing understanding of inter-and intra-patient heterogeneity of expression, PSMA expression generally increases with tumor dedifferentiation and in metastatic castration-resistant prostate cancer (mCRPC). Neuroendocrine PCa may be an exception to this rule, as case reports suggest that the PSMA gene (FOLH1) may be suppressed in neuroendocrine prostate cancer (PCa) [4].
Despite its name, PSMA is expressed in various benign and neoplastic tissues. Histopathological studies have confirmed PSMA expression in salivary glands, duodenal mucosa, proximal renal tubular cells, and neuroendocrine cells in the colonic crypts [5]. However, PSMA expression is substantially lower in these tissues than in PCa lesions [6,7]. Studies have observed PSMA radiotracer uptake in various non-prostatic tissues. High uptake of the novel PSMA radiotracer 64 copper ( 64 Cu)-PSMA has been observed in salivary glands, kidneys, and the liver [8]. High uptake of 68 gallium ( 68 Ga)-PSMA-11 single-photon-emission computed tomography (SPECT), PET/MRI, and PET/CT, but most studies have utilized PSMA-PET/CT.
Multiple novel PSMA radioligands have demonstrated sensitivity for BRPC comparable to 68 Ga-PSMA-11, as well as potential advantages. Fluorine-18 ( 18 F)-labeled PSMA-ligands have shown comparable sensitivity for BRPC lesions and enhanced image quality, suggesting potentially improved detection of small metastases. In a cohort study of 248 patients, Wondergem et al. reported comparable efficacy of fluorine-18 DCFPyL ( 18 F-DCFPyL) to 68 Ga-PSMA-11 and potentially increased efficacy for patients with PSA < 2.0 [21]. 18 F-PSMA-1007 has been found to have diagnostic accuracy comparable to 68 Ga-PSMA-11 for detection of BRPC and is only minimally excreted in the urinary tract, suggesting a potential advantage for pelvic imaging [22,23]. Additionally, logistical concerns surrounding 68 Ga-PSMA-11 PET/CT may be mitigated by novel radiotracers. 68 Ga-PSMA-11 PET/CT requires an onsite 68 germanium ( 68 Ge)/ 68 Ga generator, and, if available, 68 Ge/ 68 Ga generators may also have limited production. The longer half-life of 18 F-labeled compounds may facilitate production and permit longer-distance delivery. Novel radiotracers may also be useful to PET/CT centers lacking a 68 Ge/ 68 Ga generator.

PSMA-Targeted Imaging of Metastatic Disease
PSMA-targeted imaging has demonstrated higher sensitivity for the detection of lymph node metastases (LNM) than conventional imaging. In a cohort of 20 patients, 68 Ga-PSMA-11 PET/CT demonstrated higher sensitivity and specificity than MRI but comparable efficacy to DW-MRI [24]. In a study of 38 patients planned to undergo salvage lymphadenectomy, 68 Ga-PSMA-11 PET/CT had significantly higher negative predictive value (NPV) and accuracy for detection of LNM than 18 F-fluoroethylcholine PET/CT [25]. In a study of 65 patients who underwent 68 Ga-PSMA-11 scanning prior to salvage lymph node dissection following biochemical recurrence, Abufaraj et al. reported sensitivity ranging from 72% to 100% and specificity ranging from 96% to 100% [26]. In a retrospective study with a larger cohort of 130 patients with intermediate-to-high risk PCa staged preoperatively with 68 Ga-PSMA-PET/CT, Maurer et al. reported sensitivity and specificity of 99.1% and 95.2%, respectively, outperforming CT and MRI [27]. Region-specific PPV and NPV in this study ranged from 95% to 100% and 93% to 100%, respectively. In the first prospective study of 23 patients, 64 Cu-PSMA PET/CT demonstrated similar efficacy for detection of LNM, with a reported sensitivity of 87.5% and specificity of 100% [28]. Despite promising results, a negative correlation between lymph node size and diagnostic accuracy of PSMA-PET/CT has been described, raising concerns about low sensitivity for micrometastatic nodal tumor deposits [29].
Although few studies have investigated PSMA localization of bone metastases, preliminary research suggests that PSMA-PET/CT may be superior to bone scanning. A recent systematic review of 31 case series suggested that 68 Ga-PSMA-PET/CT identified more lesions than bone scans but noted that the large majority of studies were retrospective and did not include a reference standard [30]. A study of 415 patients who underwent 68 Ga-PSMA-PET/CT observed detection rates for bone metastasis of 48.3%, 52.6%, 74.4%, 79.6%, and 93.9% for PSA values of < 0.2 ng/mL, 0.2-0.5 ng/mL, 0.5-1 ng/mL, 1-2 ng/mL, and >2 ng/mL, respectively [31]. PSMA-PET/CT detected 258 suspicious regions, 255 of which were metastatic and 3 of which were equivocal, whereas bone scanning detected only 223 suspicious regions, 203 of which were metastatic and 20 of which were equivocal.
Variables that may influence the performance of PSMA-targeted imaging include PSA, Gleason score, and the presence of ongoing androgen-deprivation therapy (ADT). As expected, there is a strong correlation between lesion-detection rate and increasing PSA level. However, a rising PSA level does not always correlate with an increased tumor-detection rate, as patients with PSA levels above 10 ng/mL have been noted to have negative 68 Ga-PSMA-11 PET/CT scans [32]. Possible explanations for this observation include tumor location adjacent to the urinary bladder and inter-patient heterogeneity in PSMA expression. Studies have reported conflicting results with regard to Gleason score and probability of a pathological scan [33]. Regarding ADT, preclinical studies suggest that ADT increases expression of PSMA in PCa cells [34][35][36][37]. The effect of ongoing ADT on 68 Ga-PSMA-11 PET/CT efficacy, however, is unclear, with studies reporting either a positive correlation or no significant association [32,33,38,39].
PSMA expression in various benign tissues and non-prostatic malignancies has led to concerns about the specificity of PSMA-targeted imaging for metastatic disease. Neoplastic tissues with PSMA expression have been described, including transitional cell carcinoma, hepatocellular carcinoma, renal cell carcinoma, and colorectal carcinoma. False positives in benign conditions are also increasingly noted, such as 68 Ga-PSMA-11 uptake in sarcoidosis and Paget's disease and 64 Cu-PSMA uptake in pneumonitis [10,40,41]. However, PSMA-targeted imaging has demonstrated higher specificity for PCa than conventional imaging. Moreover, as understanding of the physiological distribution of PSMA improves, unusual sites of tracer avidity are less likely to lead to false positive interpretation.

Clinical Impact of and Future Directions for PSMA-Targeted Imaging
Pending results from ongoing prospective trials (Table 1), PSMA-targeted imaging may eventually play multiple roles in the management of PCa. Given the substantial evidence of higher sensitivity of PSMA-PET/CT for BRPC than conventional imaging, particularly at low PSA levels, currently the best evidenced role for PSMA-PET/CT is restaging patients with BRPC. Given its high efficacy for detection of lymphatic metastases, PSMA-PET/CT may also become the standard of care in lymph node staging and preoperative planning prior to lymph node dissection. Other potential applications of PSMA-PET/CT currently under investigation include identification of the suspected primary site of PCa, primary staging of intermediate-to-high risk PCa, and targeted biopsy [42][43][44][45].
Preliminary evidence suggests that PSMA-targeted imaging significantly affects clinical practice. In a study of 118 patients with BRPC and high-risk (HR) PCa who underwent 68 Ga-THP-PSMA PET/CT at diagnosis, management changed in 34% of patients (9/26) in the BRPC group and in 24% of patients (12/50) in the HR group [46]. In a prospective Phase II/III study of initial staging with 18 F-DCFPyL PET/CT of 252 men with HR PCa who were planned for radical prostatectomy with lymphadenectomy, Pouliot et al. observed a PPV of 86.7% and reported that 22% of men (56/252) were upstaged to N1 or M1 disease by 18 F-DCFPyL PET/CT [47]. PSMA-PET/CT may also better stratify patients potentially eligible for early SRT, given that SRT is commonly initiated in patients with serum PSA levels below those at which conventional imaging is reliably sensitive. Thus, more accurate localization of target volumes prior to SRT initiation might improve clinical response and reduce off-target effects. The impact of 68 Ga-PSMA-11 PET/CT on the success rate of SRT for recurrent PCa after prostatectomy is currently being evaluated in a large randomized prospective trial (PSMA-SRT, NCT03582774) [48]. PSMA-PET/CT may also effectively identify patients for treatment with PSMA radioligand therapy (RLT). In a prospective Phase II trial that evaluated the efficacy of PSMA-PET/CT in predicting response to RLT, PSMA-PET/CT reliably predicted ≥30% PSA reduction, but no imaging parameters predicted ≥50% PSA reduction [49]. Improved localization of metastatic PCa using PSMA-targeted imaging may also increase the success rate of metastasis-directed therapy (MDT), including stereotactic body radiotherapy (SABR). MDT intends to postpone systemic treatment for patients with oligometastatic disease, thereby reducing the side-effects of hormonal therapy. Prospective data suggest that SABR is well tolerated and improves PFS in patients with oligometastatic PCa [50]. In the randomized STOMP study, which used choline PET/CT, stereotactic ablation of oligometastatic disease in 62 patients delayed the need for hormonal therapy from 13 to 21 months compared with surveillance [51]. Data from ORIOLE, a similar randomized Phase II study investigating the efficacy of SABR in forestalling metastases for hormone-sensitive PCa compared with observation, provide evidence for the value of PSMA PET/CT in controlling disease [52]. Patients randomized to the SABR arm of ORIOLE underwent 18 F-DCFPyl PET/CT, a urea-based PSMA radiotracer, prior to and 180 days after treatment. Patients with no additional untreated lesions detected by PSMA PET/CT at baseline were significantly less likely to develop new metastatic lesions at six months than those whose PSMA PET/CT showed at least one additional lesion at baseline (16% vs. 63%, respectively).
Currently, the only PSMA-targeted imaging agent approved by the U.S. Food and Drug Administration is 111 In-capromab pendetide (ProstaScint), which is approved for SPECT imaging of biopsy-proven PCa localized to the prostatic bed but at high risk for pelvic LNM. Regulatory approval of PSMA radiotracers in the United States has lagged behind other areas of the world, such as Europe and Australia, where much of the innovation in PSMA theranostics has occurred. Given the growing evidence of the clinical potential of PSMA, its availability is expanding in the United States, and regulatory approval of novel radiotracers, including 68 Ga-PSMA-11 and 18 F-DCFPyl, is expected within the next year. Ongoing clinical trials will help to better define the clinical role and impact of PSMA imaging and possibly strengthen the case for regulatory approval (Table 1). In sum, PSMA-targeted imaging has demonstrated clinical benefits through targeting stereotactic ablation in oligometastatic disease, but no studies have yet shown that PSMA-targeted imaging improves clinical outcomes for biochemically recurrent PCa patients.

PSMA Radioligand Therapy (RLT)
PSMA has also emerged as a promising therapeutic molecular target. Although various therapies are now approved for mCRPC, their survival benefit is generally limited to less than 6 months. There is thus a clinical need for novel therapies leading to a sustained response. Progress in the development of synthetic PSMA radioligands has led to an emerging body of research indicating significant therapeutic efficacy of PSMA RLT.
PSMA RLT studies have mostly utilized small-molecule inhibitors of PSMA as radioligands, which have been shown to be less hematotoxic than monoclonal antibodies. For instance, a study of MEDI3726 (a PSMA-targeted antibody-drug conjugate) in patients with mCRPC after failure of abiraterone or enzalumatide observed significant responses at higher doses of MEDI3726, although responses were not durable because patients discontinued therapy due to drug-related adverse events [53]. Small-molecule PSMA inhibitors have been labeled with both beta-and alpha-emitting radioisotopes, which have variable energy levels and path lengths. Beta-emitting radioisotopes, such as lutetium-177 ( 177 Lu), are the favored radioisotopes given their short maximal tissue penetration and relatively long half-life, permitting delivery of a high degree of radiation to PCa lesions. Advantages of alpha-emitting radioisotopes include reduced red-marrow infiltration, leading to less hematotoxicity. In a proof-of-concept paper, application of the alpha-emitting 225 actinium ( 225 Ac) to two patients with diffuse bone marrow involvement led to undetectable PSA levels in both patients and no relevant hematotoxicity [54].
Among alpha-emitting radioisotopes, preliminary investigations of 225 Ac-PSMA-617 suggest that it has substantial therapeutic efficacy for mCRPC. In a study of 17 chemotherapy-naive patients with advanced metastatic PCa treated with 225 Ac-PSMA-617, 82% of patients had a PSA decline of ≥90%, and 41% of patients had undetectable serum PSA 12 months after therapy [55]. Targeted alpha-therapy may also benefit patients resistant to beta-emitting therapy and patients for whom beta-emitting therapy is contraindicated [54].
The few studies that have examined overall survival (OS) of 177 Lu-PSMA RLT have observed OS rates comparable to currently available third-line therapies. In a cohort of 59 patients with CRPC who were previously treated with second-generation ADT and chemotherapy, Brauer et al. reported a median progression-free survival (PFS) of 4.5 months and median OS of 8 months [59]. A retrospective study with a larger (104 patients) and more homogenous cohort of patients who were heavily pretreated reported a median OS of 14 months [70]. Investigators determined that a decline of at least 20.87% was the optimal parameter for predicting improved OS, but no specific level of PSA decline has been established as a surrogate for OS. Any PSA decline after the first cycle of 177 Lu-RLT has been reported as a significant prognosticator of survival [59,71]. In the first prospective Phase II study of 30 patients with CRPC, investigators reported median PSA PFS and OS of 7.6 months and 13.5 months, respectively [63]. Compared to conventional third-line therapies for CRPC, a recent systematic review found that median OS was longer with 177 Lu-PSMA RLT than with third-line treatment, but the difference was not statistically significant (mean of 14 months vs. 12 months, respectively, p = 0.32) [72].
Inter-and intra-patient heterogeneity of PSMA expression has been cited as a potential pitfall of PSMA RLT and may limit its clinical application. Studies of primary PCa suggest high homogeneity of PSMA expression [73]. However, immunohistochemistry studies of mCRPC lesions have noted significant inter-and intra-patient heterogeneity of PSMA expression [7,74]. Preclinical research has suggested that despite an overall increase in PSMA expression during progression of PCa from androgen sensitivity to androgen independence, some metastatic cell lines lose PSMA expression [75]. A significant proportion of liver metastases may also lack PSMA expression, although most liver metastases highly overexpress PSMA [76]. Heterogeneity of PSMA expression may partly explain why about 30% of patients do not respond to 177 Lu-PSMA RLT [77]. In contrast, low PSMA expression in patients with mCRPC who progress after conventional therapies may be a negative prognostic indicator [78].
PSMA has also been the subject of novel immunotherapeutic approaches to mCRPC, such as bispecific T-cell engagers (BiTEs). BiTEs are a class of novel antibodies that form a link between T cells and tumor cells, permitting T-cell cytotoxic activity and initiating apoptosis of malignant cells. The PSMA/CD3-bispecific BiTE antibody BAY2010112 (AMG212, MT112) has been found to potently suppress tumor growth in preclinical research and was found to have an acceptable safety profile and dose-dependent clinical activity in a Phase I study of 16 patients [79,80].

Safety of PSMA RLT
Studies have reported encouraging results on the safety of PSMA RLT. In the aforementioned multicenter study of 145 patients with mCRPC treated with 177 Lu-PSMA-617, Grade 3 to 4 toxicities including anemia, leukopenia, and thrombocytopenia were reported in 10%, 3%, and 4% of patients, respectively [69]. Salivary gland toxicity, including mild or transient xerostomia, occurred in 8% of patients. A similar safety profile was reported in a study of 49 patients treated with three cycles of 177 Lu-PSMA-617, but no Grade 4 hematotoxicity was observed and there were no significant differences between the PSMA RLT group and the control group in incidence of leukopenia or thrombocytopenia [71]. Mild nausea, loss of appetite, and fatigue are the most common nonhematologic adverse effects reported in studies of 177 Lu PSMA-617. Despite the renal binding of PSMA ligands, research suggests 177 Lu-PSMA-617 is relatively non-nephrotoxic. Low-grade nephrotoxicity has been reported, but there have been no reports of Grade 3 or 4 toxicity [81,82]. Risk factors for nephrotoxicity with 177 Lu-PSMA-617 have also been identified. They include age (p = 0.049), hypertension (p = 0.001), and pre-existing kidney disease (p = 0.001) [81].

Conclusions
Extensive research has demonstrated the excellent diagnostic accuracy of PSMA-targeted imaging for the detection of BRPC, but the impact on overall survival of earlier initiation of therapy based on PSMA-targeted imaging has not yet been elucidated. Research is ongoing to define PSMA-targeted imaging's exact role in other stages of the disease. Preliminary efficacy and safety data on PSMA RLT are very encouraging, and confirmatory data from larger studies will read out soon. While no PSMA RLT agent has yet obtained regulatory approval, federal approval is expected in the near future as ongoing studies read out ( Table 2). The future of PSMA therapeutics may include novel radioisotopes, immunotherapeutic ligands, and combined approaches.  Author Contributions: Writing-original draft preparation, W.J., K.G.; writing-review and editing, K.G., W.J., P.C.B., C.J.P.; supervision, P.C.B., C.J.P. All authors have read and agreed to the published version of the manuscript.
Funding: This research received no external funding.

Conflicts of Interest:
The authors declare no conflict of interest.