PSMA radioligand therapy for solid tumors other than prostate cancer: background, opportunities, challenges, and first clinical reports

In the past decade, a growing body of literature has reported promising results for prostate-specific membrane antigen (PSMA)-targeted radionuclide imaging and therapy in prostate cancer. First clinical studies evaluating the efficacy of [177Lu]Lu-PSMA radioligand therapy (PSMA-RLT) demonstrated favorable results in prostate cancer patients. [177Lu]Lu-PSMA is generally well tolerated due to its limited side effects. While PSMA is highly overexpressed in prostate cancer cells, varying degrees of PSMA expression have been reported in other malignancies as well, particularly in the tumor-associated neovasculature. Hence, it is anticipated that PSMA-RLT could be explored for other solid cancers. Here, we describe the current knowledge of PSMA expression in other solid cancers and define a perspective towards broader clinical implementation of PSMA-RLT. This review focuses specifically on salivary gland cancer, glioblastoma, thyroid cancer, renal cell carcinoma, hepatocellular carcinoma, lung cancer, and breast cancer. An overview of the (pre)clinical data on PSMA immunohistochemistry and PSMA PET/CT imaging is provided and summarized. Furthermore, the first clinical reports of non-prostate cancer patients treated with PSMA-RLT are described. Supplementary Information The online version contains supplementary material available at 10.1007/s00259-021-05433-w.


Introduction
Prostate-specific membrane antigen (PSMA) is a transmembrane protein that is encoded by the FOLH1 (folate hydrolase 1) gene and was first discovered in prostate cancer cells [1]. Contrary to what the name suggests, PSMA is not only selective to prostate cancer cells [2,3] but also expressed by neovascular endothelial cells of various cancers, including glioblastoma, kidney cancer, lung cancer, and breast cancer [4][5][6]. Preclinical data suggests that PSMA might be involved in cancer-related angiogenesis by degrading the extracellular matrix and participating in integrin signal transduction [7,8].
To date, most clinical research on PSMA focuses on prostate cancer due to its exceptional high level of PSMA expression by tumor cells. Clinical studies evaluated the potential of PSMA imaging using radiolabeled PSMA antibodies (ProstaScint®, J591) and ligands (namely [ 68 Ga] Ga-PSMA-11 and [ 18 F]F-PSMA-1007), mainly by positron emission tomography (PET), revealing higher tumor detection rates and higher tumor-to-background ratios compared to conventional imaging modalities [9][10][11][12]. Subsequently, PSMA targeting antibodies (J591) or ligands (PSMA-617 or PSMA-I&T) were labeled with therapeutic radionuclides such as lutetium-177 ( 177 Lu) or actinium-225 ( 225 Ac) respectively [13][14][15][16][17][18][19]. Driven by the favorable binding features and pharmacokinetics of ligands compared toavailable antibodies (low bone marrow toxicity due to faster clearance), PSMA ligands are currently the main focus for PSMA therapy in prostate cancer patients [13,19]. Yet, comparative studies are still lacking. [ 177 Lu]Lu-PSMA-617 has demonstrated promising results in prostate cancer patients in one prospective study, as well as several retrospective studies and compassionate use programs worldwide [14][15][16]. A phase II trial on [ 177 Lu]Lu-PSMA in heavily pretreated progressive prostate cancer showed efficacy, i.e., a PSA decline > 50%, in 57% of patients and a progression-free survival of 7.6 months [13]. Moreover, a phase III registration trial (VISION) in advanced prostate cancer patients completed recruiting and final results are awaited at the end of 2021 (NCT03511664).
Since PSMA radioligand therapy (PSMA-RLT) demonstrated remarkable therapeutic efficacy in prostate cancer patients, the question arises whether PSMA-RLT could also achieve beneficial effects in other cancers expressing PSMA on the tumors cells themselves, or in the tumor-associated neovasculature.
The aim of this review is to assess which other solid cancers could potentially benefit from PSMA-RLT, based on PSMA expression levels and PSMA imaging data. Potential challenges and differences compared to prostate cancer are discussed. Additionally, the results of the first clinical reports of PSMA-RLT in solid tumors other than prostate cancer are presented.

Search strategy
The selection of cancer types for this review was based on a combination of PSMA expression analysis and electronic library searches. First, the FOLH1 gene expression levels (this gene encodes PSMA) of all cancers included in the TCG Pan-Cancer Atlas were obtained from cBioPortal ( Fig. 1) [20,21]. Second, literature was searched by universal PubMed searches (see supplementary 1) for the fifteen cancers with the highest PSMA expression level on the PanCancer Atlas. Solely cancers with a substantial (> 20) amount of PubMed results were included in this review.
This resulted in the inclusion of glioblastoma, thyroid cancer, renal cell carcinoma, hepatocellular carcinoma, lung cancer, and breast cancer. Additionally, we included salivary gland cancer. Although this rare tumor entity is not included in the TCG PanCancer Atlas, several relevant PSMA-related studies were conducted for this cancer type.
For each type of cancer, the PubMed results were screened for papers or case reports which investigated PSMA expression levels through immunohistochemistry, PSMA imaging (e.g., PET/CT scans), or reports on PSMA-RLT. Both preclinical and clinical studies were included. The last search was performed on the 23rd of October 2020.

PSMA immunohistochemistry
All articles reporting on PSMA immunohistochemistry (IHC) of the above-mentioned seven solid cancer types were included. No selection was made based on the type of antibody used for IHC staining, since there is no golden standard for PSMA IHC. Antibodies targeting the intracellular and extracellular domains of PSMA were included. A distinction was made between IHC staining on the tumor cells and the neovasculature. For each tumor type, the percentage of tumors which are PSMA positive on the IHC staining was described.

PSMA PET/CT imaging
Although mRNA and PSMA IHC data provide relevant information on PSMA expression levels, in clinical practice, eligibility for PSMA-RLT in prostate cancer is based on in vivo tracer uptake revealed by PSMA PET/CT, semi-quantitatively expressed as standardized uptake values (SUV). According to the European Association of Nuclear Medicine (EANM) guideline based on the phase II trial on [ 177 Lu]Lu-PSMA-617, the required maximum SUV (SUVmax) at dominant sites of tumor involvement should be at least 1.5-fold higher than the baseline mean SUV (SUVmean) of the liver on PET/CT (using renally excreted ligands such as [ 68 Ga]Ga-PSMA-11) to qualify for therapy [13,22]. Therefore, we looked at the tumor/liver ratio; if this was not reported, we used a SUVmean of 4-8 for liver as reported in the literature [23,24]. This suggests that a tumor uptake (SUVmax) of at least 12 might be considered sufficient to explore PSMA-RLT.

Salivary gland cancer
Salivary gland cancer (SGC) is a rare and complex disease with an annual incidence of 2 per 100,000, consisting of 22 subtypes each with different clinical behaviors and prognoses [25,26]. PSMA-related research has solely been conducted for adenoid cystic carcinoma (ACC) and salivary duct carcinoma (SDC).
Healthy salivary glands show high physiological tracer uptake on PSMA PET scans [9]. Interestingly, unlike prostate cancer cells, the uptake of PSMA ligands by the salivary glands does not seem to be completely mediated by PSMA; at least part of the uptake is aspecific [27,28].
PSMA expression has been examined using IHC for both ACC and SDC in primary tumor material as well as metastases (details can be found in Table 1). The majority of ACC express PSMA on the tumor cells (91%-145/159 patients), while none of the tumors showed PSMA expression in the vasculature. In contrast, in SDC, the majority of the vessels express PSMA (90%-9/10), and some of the tumor cells express PSMA (40%-4/10) [29][30][31][32][33].
PSMA tracer uptake in ACC, as visualized with PSMA PET/CT, was first described in several case reports. Some patients demonstrated high PSMA uptake (SUVmax: 23.3) in metastatic lesions, compared to other patients who only showed low to modest tracer uptake in the tumor cells (SUVmax: 1.2) [31,32,34,35]. This variation in uptake was confirmed in larger studies (van Boxtel et al. also included SDC patients), also describing a large variation of PSMA tracer uptake between patients (Fig. 2) [29,30]. Even within a patient, a relatively large inter-metastatic variation in tracer uptake was detected [29]. van Boxtel et al. reported a tumor/ parotid ratio, which was < 1 in the vast majority of cases [29]. Therefore, PSMA PET/CT imaging might be of limited value for detecting primary tumors or local recurrences, but could be useful for detecting lymph node or distant metastases. Overall, SUVmax values ranged from 1.1 to 30.2 in ACC patients and from 0.3 to 25.9 in SDC patients. This suggests that PSMA-RLT might be of interest for a subset of salivary gland cancer patients, since some patients showed lesions with a SUVmax > 12.
Regarding clinical studies on PSMA-RLT (Table 2), one patient with stage IV ACC received a single dose of [ 177 Lu]Lu-PSMA (7.5 GBq) [34]. Treatment was well tolerated with no side effects and some pain relief was reported. Whole-body [ 177 Lu]Lu-PSMA SPECT/CT imaging after therapy showed intense uptake in the metastases. A planned second cycle of [ 177 Lu]Lu-PSMA was canceled

Glioblastoma
Glioblastoma is the most frequently occurring type of brain cancer, with an annual incidence of 5 per 100,000 and is highly aggressive [36]. Glioblastomas are known to be highly vascularized tumors [37]. The first reports on immunohistochemical staining in glioblastoma tumors observed PSMA expression only in the neovasculature and not in the tumor cells [38][39][40]. Therefore, subsequent IHC studies primarily focused on the PSMA expression of the neovasculature [41][42][43][44]. Overall, 72% (92/128) of the glioblastoma tumors express PSMA in the neovasculature (Table 1). Two reports also quantified the vasculature staining by scoring the percentage of PSMApositive vessels and staining intensity [41,44]. Wernicke et al. [41] reported that in 69% of the tumors > 50% of the vessels were PSMA positive, while this was only the case in 32% of the tumors in Mahzouni et al. [44].
PSMA ligand uptake by glioblastoma tumors has been observed with different diagnostic radiotracers [40,42,[45][46][47][48][49][50][51]. Bertagna et al. previously published a systematic review with a focus on the possible diagnostic role of PSMA PET/CT imaging, including most of these studies [52]. They concluded that glioblastomas are PSMA-avid tumors and that PSMA PET/CT imaging could be a useful diagnostic tool in glioblastoma. Articles published since then are in line

Thyroid cancer
Thyroid cancer is an endocrine malignancy with an annual incidence of 2-6 per 100,000 [56]. The most common subtype is differentiated thyroid carcinoma (DTC), which includes papillary thyroid carcinoma (PTC) and follicular thyroid carcinoma (FTC) [57]. Other rare subtypes of thyroid cancer are medullary thyroid carcinoma (MTC) and anaplastic thyroid carcinoma (ATC) that have a dismal prognosis.
Regarding PSMA-RLT, literature reports on three treated thyroid cancer patients (  [77]. Three patients were considered eligible for PSMA-RLT, of whom two were treated with 2 cycles of 6 GBq [ 177 Lu]Lu-PSMA-617. One of the patients did not respond to therapy and showed disease progression on [ 18 F]FDG PET/CT after 1 month. Interestingly, the second patient did have a partial response of lung and liver metastases on imaging, and a transient decrease of the tumor marker thyroglobulin from 17 to 9 μg/L. Seven months post treatment, disease progression was observed on imaging and the thyroglobulin level increased to 14 μg/L. Both of these DTC case reports did not report on side effects of PSMA-RLT [76,77].

Renal cell carcinoma
Renal cell carcinoma (RCC) has an incidence of 4.4 per 100,000 [78]. Renal tumors are divergent and their clinical behavior is highly dependent on the histological subtype [79]. Clear cell RCC is the most common subtype and accounts for the majority of kidney cancer-related deaths [80]. Importantly, pro-angiogenic factors (VEGF, PDGF) are strongly upregulated in clear cell RCC, leading to high vascularized tumors. Other frequently occurring RCC subtypes include papillary RCC and chromophobe RCC.
Regarding the PSMA expression in RCC, most research has been conducted for clear cell RCC and papillary RCC (Table 1). PSMA expression of primary renal neoplasms demonstrated an exclusive PSMA expression in the tumorassociated neovasculature [5]. This holds true for clear cell RCC, papillary RCC, chromophobe RCC, and oncocytoma. Clear cell RCC was found to have the highest percentage of PSMA-positive tumors and also the highest PSMA staining intensity. In contrast, transitional cell and angiomyolipoma showed no PSMA expression [81,82]. Overall, seventy-nine percent (236/299 patients) of the primary clear cell RCC tumors showed positive PSMA staining, in contrast to 27% (16/59) in primary papillary RCC. In addition, in metastatic clear cell RCC, 75% (15/20) of the tumors showed PSMA expression in the neovasculature [81][82][83]. Spatz et al. presented the largest cohort, with 257 RCC patients (including papillary, clear cell, and chromophobe subtypes). Interestingly, this cohort related stronger PSMA expression patterns with high-grade and advanced tumors and increased staining intensity was associated with poorer overall survival [81].
The role of PSMA PET/CT imaging in RCC has yet to be defined, but its potential has been investigated in multiple clinical studies. Due to the highest PSMA expression in clear cell RCC, this subtype has gained the most interest for clinical application. This consensus is reinforced by a recent report that showed inconsistent detection of non-clear cell RCC lesions [84]. Several explorative studies showed heterogeneity of PSMA uptake in clear cell RCC lesions; in primary/recurrent tumors, SUVmax ranged from 1.7 to 39.4, and in metastatic lesions, a range between 0.9 and 48 was reported [85][86][87]89]. Since some of the SUVmax values described in literature are above 12, a part of the patients might be considered for PSMA-RLT. This is supported by Siva et al. who also mentioned that [ 177 Lu]Lu-PSMA treatment might be feasible in a part of the recurrent RCC patients based on the high PSMA tracer uptake [89].
To date, no RCC patients have been treated with PSMA-RLT according to literature.

Hepatocellular carcinoma
Hepatocellular carcinoma (HCC) is the most frequent primary liver cancer with an incidence of 10.1 cases per 100,000 person-years [90].
Overall, PSMA expression was mostly observed in the tumor-associated neovasculature (83%-235/282), and it was associated with poor prognosis in patients with HCC [91]. Only one of the studies also identified some PSMA expression by the tumor parenchyma (41% of samples), in a canalicular pattern [93].
Regarding PSMA-RLT (Table 2), two HCC patients received PSMA-RLT [10]. Both patients were treated with one cycle of [ 177 Lu]Lu-PSMA-617 (activity 5.9-6.2 GBq). Although the treatment was well tolerated, intra-therapeutic SPECT/CT-based dosimetry revealed disappointing radiation dosages. According to the authors, the PSMA-RLT dose was at least tenfold lower than typically achieved by one cycle of external beam radiation therapy for HCC. Therefore, PSMA-RLT was discontinued after one cycle for both patients.

Lung cancer
The incidence of lung cancers varies largely between countries and differs between sexes. It ranges from < 10 to > 50 per 100,000 person-years [103]. Lung cancer is generally divided into small cell lung cancer (SCLC) or non-small cell lung cancer (NSCLC). NSCLC can be further classified into adenocarcinoma (most common form), squamous cell carcinoma, and large cell carcinoma [103].
In lung cancer, presence of PSMA is mainly observed on the neovasculature (Table 1), with expression levels in primary tumors ranging from 45% (63/141 patients) in adenocarcinoma to 70% in large cell carcinoma (49/70) and SCLC (21/30) [104,105]. Positive staining of the tumor cells was shown by Wang et al. in all three subtypes of NSCLC, in approximately half of the cases [104]. However, this was not observed by Schmidt et al. where only a small fraction of the NSCLC cases were PSMA positive on the tumor cells (2-12%) [105]. PSMA expression on SCLC tumor cells has only been studied by Wang et al., who showed no PSMA expression [104].
All literature on PSMA tracer uptake by lung cancer derives from accidental findings of lung lesions on PSMA PET/CT scans in patients who received PSMA imaging for their prostate cancer [106][107][108][109][110][111]. The reported SUVmax of lung cancer ranged from 4.8 to 5.6 in lung adenocarcinoma (Table 1). In 9 patients with NSCLC (without further details on subtype), SUVmax ranged from 3.7 to 7.0. However, these lung cancers are identified due to their PSMA tracer uptake, and most likely PSMA-negative lung cancers have not been published. Therefore, it remains unclear which proportion of lung cancer patients show PSMA tracer uptake and to what extent (SUV values).
Case reports on PSMA-RLT for lung cancer patients were not identified.

Breast cancer
Breast cancer is one of the most prevailing types of cancer, with an incidence of 128.5 per 100,000 woman per year [112]. The most common histopathological subtype, accounting for 75% of all breast cancers, is invasive carcinoma of no special type (IC-NST), formerly known as invasive ductal carcinoma [113]. The second most common type is invasive lobular carcinoma (ILC) (5-10%).
Immunohistochemical PSMA expression has been examined for both primary IC-NST and ILC tumors (Table 1) [38,114]. In metastatic breast cancer, PSMA expression was described in two reports; 96% (22/23) of these samples were positive for PSMA in the neovasculature [114,115]. Remarkably, Wernicke et al. described that PSMA expression of all tumor metastases correlated with PSMA expression intensity of the primary tumor. They also found that both estrogen and progesterone receptor-negative tumors were more likely to have higher PSMA expression compared to hormone receptorpositive tumors [115].
Reports on PSMA PET/CT imaging mainly consisted of case reports (Fig. 2), which provided limited data on SUVmax or tumor/liver ratios. One article reported on PSMA imaging in 19 breast cancer patients (68% with IC-NST) [116]. The SUVmean was 2.5 ± 2.6 for primary or local recurrences (n = 13) and 3.2 ± 1.8 for involved lymph nodes (n = 15). Distant metastases (n = 53) showed a significantly higher SUVmean of 6.9 ± 5.7 compared to primary tumor/ local recurrence (p = 0.04) and lymph node metastases (p = 0.011). SUVmean did not show a significant correlation with hormone receptor status; however, PSMA uptake increased with tumor grading and was more often seen in IC-NST compared to other histological subtypes. Based on the limited literature, the SUVmax values that have been reported in breast cancer are generally low (mean SUVmax: 2.5-6.9) suggesting limited potential for PSMA-RLT in breast cancer.
Interestingly, we found a preclinical PSMA-related study in breast cancer. This study investigated the potential of PSMA-RLT in breast cancer, showing that [ 177 Lu] Lu-PSMA strongly impaired the vitality and angiogenic capacity of endothelial cells cultured in breast cancer conditioned medium [117]. Regarding the clinical application of PSMA-RLT (Table 2), a 38-year-old female with an aggressive triple-negative breast carcinoma, previously unresponsive to chemotherapy and bevacizumab, received [ 177 Lu] Lu-PSMA-RLT (2 cycles 7.5 GBq) based on intense tumor tracer uptake on PSMA imaging (SUV not reported). Posttherapy SPECT imaging showed uptake in the tumor lesions and the treatment was well tolerated. However, severe disease progression was seen after the second treatment cycle and treatment was terminated [118]. No other case reports on PSMA-RLT in breast cancer were found.

Discussion
Increasing evidence shows that PSMA-RLT is an effective treatment for prostate cancer patients with a favorable toxicity profile [13,119]. Currently, PSMA-RLT is investigated in a phase III trial (VISION; NCT03511664). These promising results in prostate cancer in combination with literature showing PSMA expression and PSMA tracer uptake in other malignancies encouraged us to assess the potential role of PSMA-RLT for other solid cancer types [120]. We focused on PSMA expression, PSMA PET/CT tracer uptake, and results of clinical attempts of PSMA-RLT in seven different solid cancers.
Regarding PSMA immunohistochemistry, in the majority of the solid cancers included in this review, > 70% of the primary tumors showed PSMA expression on the tumorassociated neovasculature. Of all included primary tumors, medullary thyroid carcinomas and hepatocellular carcinomas most often expressed PSMA in the neovasculature. In contrast, in adenoid cystic carcinoma (subtype of salivary gland cancer) and papillary renal cell carcinoma, only few of the tumors showed PSMA-positive staining on the neovasculature. Interestingly, although most of the solid cancers did not express PSMA on the tumor cells, it was still observed in salivary gland tumors (especially in adenoid cystic carcinoma), and to minimal extent in hepatocellular, lung, and breast cancer tissue.
On PSMA PET/CT imaging, PSMA tracer uptake differed considerably between the solid cancers. In patients with salivary gland cancer, glioblastoma, thyroid cancer, hepatocellular cancer, and clear cell renal cell cancer, several patients showed relevant tumor tracer accumulation on PET imaging (SUVmax values ≥ 12). Extremely high SUVmax values up to 101.8 were seen in metastatic medullary thyroid carcinoma. On the other hand, in several types of lung cancer and breast cancer, tracer uptake was low to moderate at best (SUVmax < 10). Noteworthy, in solid cancers, intrapatient tumor heterogeneity was observed.
PSMA-RLT eligibility in prostate cancer is assessed through PSMA PET/CT imaging, with an eligibility cutoff tumor/liver ratio > 1.5 in [ 68 Ga]Ga-PSMA PET/CT according to the EANM guideline [22]. As elaborated on in the "Methods" section, for the aim of this review, we considered a tumor SUVmax of > 12 sufficient to potentially investigate PSMA-RLT. Bearing this in mind, we conclude that salivary gland cancer, glioblastoma, thyroid cancer (differentiated and medullary), hepatocellular carcinoma, and renal cell cancer (clear cell) are the most relevant tumors to further explore the potential of PSMA-RLT. In line with this, the first case reports on PSMA-RLT in patients other than prostate cancer included salivary gland cancer, glioblastoma, thyroid cancer, and hepatocellular carcinoma [30,34,53,55,76,77,102]. These nine heavily pretreated end-stage patients received 1-2 cycles of 5.9-8.4 GBq [ 177 Lu]Lu-PSMA per cycle in compassionate use programs (Table 2). In some of these case reports, positive treatment outcomes were reported. In one salivary gland cancer patient, pain reduction was observed [34]. One thyroid cancer patient showed a partial response that lasted 7 months [77]. In a glioblastoma patient, tumor volume decreased upon PSMA-RLT [55]. Importantly, the treatment was generally well tolerated, with no or low-grade adverse events.
Despite the use of the same PSMA PET/CT-based eligibility criteria as in prostate cancer patients to assess possible PSMA-RLT application, there are essential differences between prostate cancer and the solid cancers included in this review.
First, in prostate cancer, PSMA is expressed on the tumor cells compared with the mainly neovascular expression in most of other solid cancers. Still, even if PSMA is solely expressed on the neovasculature of well-perfused tumors, PSMA-RLT could hypothetically induce a tumoricidal local radiation dose to the tumor cells due to the tissue range (2 mm) of beta particles emitted by radionuclides such as 177 Lu [120]. Furthermore, radiation dosages to the neovasculature and tumor micro-environment might also lead to a harmful effect and induce secondary immune responses.
It has been speculated that PSMA expression solely on the neovasculature could result in a shortened tracer washout [53,94], meaning that PSMA-RLT is not retained in the tumor for a longer time, resulting in a lower radiation dose to the tumor and less effective treatment. Yet, a case report on a glioblastoma patient treated with [ 177 Lu]Lu-PSMA showed good tumor tracer retention on post-therapy imaging, resulting in a substantial tumor absorbed dose [53], and in another glioblastoma patient, [ 177 Lu]Lu-PSMA treatment resulted in a decrease in tumor volume [55]. This suggests that a sufficient radiation dose might still be reached while PSMA expression is limited to the tumor vasculature. Nonetheless, dedicated studies including dosimetry are required to prove this.
Second, SUVmax values in other solid cancers are generally lower than the SUVmax (> 15-40) values in prostate cancer [121,122]. This suggests that lower radiation doses in the tumor could be reached, likely leading to a lower fraction of patients responding to PSMA-RLT compared to prostate cancer patients.
Third, more intra-patient tumor heterogeneity in terms of PSMA expression is seen in other solid tumors compared to prostate cancer [123]. Supposedly, this is a result of the neovasculature versus tumor cell PSMA expression, as described above. To illustrate this, metastases with high neo-vascularization would have higher PSMA uptake compared to metastases with low neo-vascularization within a patient. However, even in patients with heterogenous PSMA tracer uptake, the bystander and abscopal effect known in radiation-oncology might lead to tumor responses in PSMAnegative tumors [124,125].
Fourth, it has been long known that cancers vary in radiosensitivity [126]. Prostate cancer is generally radiosensitive and external beam radiotherapy is effective in early-stage disease, which provided a good rationale for PSMA-RLT in metastatic prostate cancer [127]. In contrast, hepatocellular carcinoma, for example, is considered less radiosensitive [128,129]. Therefore, some of the other solid cancers might require higher PSMA-RLT radiation doses, compared to prostate cancer, to achieve a clinically relevant response.
Based on this review, PSMA-RLT could potentially be investigated for certain solid cancers (e.g., salivary gland cancer, glioblastoma, thyroid cancer, liver cancer, and clear cell renal cell cancer). This has also been proposed for these cancers by other authors [29,30,46,54,76,89,95]. Sufficient PSMA uptake on PSMA PET/CT is a crucial parameter to consider therapy, and since a significant fraction of prostate cancer patients with high PSMA PET/CT tracer uptake do not respond to PSMA-RLT, other parameters, which are not clearly identified yet, obviously play also a relevant role. Hence, a good pre-selection of patients is crucial to apply this therapy in these patients in the future.
The current literature on PSMA uptake in cancers other than prostate cancer is scarce and prospective studies are rare. Therefore, it is not possible to draw firm, generalized conclusions. Furthermore, many of the papers included in this review are case reports. It is likely that patients with high PSMA uptake are reported, while negative results are less likely to be published, known as publication bias. Therefore, it may appear as if more tumors are PSMA avid, or have higher SUV values than is actually the case. In addition, relevant data such as PSMA tracer uptake (SUVmax) was frequently not reported [40,49,50,64,67,69,107,130,131].

Future prospective
Currently, prospective PSMA PET/CT imaging studies in end-stage non-prostate cancer patients are lacking; consequently, reliable information on PSMA uptake is missing. In our opinion, prospective imaging studies are the key way towards exploring PSMA-RLT for non-prostate cancers, especially imaging studies in patients with advanced disease, as PSMA-RLT is likely to be explored in end-stage disease with limited other treatment options. This will provide essential information on PSMA uptake and enables estimating which proportion of patients could be eligible for PSMA-RLT.
Furthermore, when evaluating the potential of PSMA-RLT in other cancers, preclinical studies on the therapeutic effects of PSMA-RLT are currently lacking and would be advisable. These could provide relevant insight into the fundamental questions such as PSMA tracer retention in tumors where PSMA is limited to the neovasculature. Preclinical studies could also provide information on the sensitivity of non-prostate cancers to PSMA-RLT.
In a clinical setting, prospective therapeutic studies should be performed instead of single patient reports to prevent trial-and-error-based science. As a different PSMA localization (tumor cell surface in prostate cancer versus neovasculature in other solid cancers) might lead to different PSMA tracer kinetics, preferably these prospective studies should include multi-timepoint post-therapy imaging (dosimetry), to provide more information on PSMA tracer kinetics. Currently, a prospective therapeutic study in salivary gland cancer patients is recruiting (NCT04291300). Furthermore, pending the outcome of the pivotal trial for [ 177 Lu]Lu-PSMA (VISION trial; NCT03511664) in prostate cancer patients, it is anticipated that with positive results the translation to other solid cancers may be accelerated.

Conclusion
In summary, PSMA expression in solid cancers other than prostate cancer is primarily observed in the tumor neovasculature, with the exception of adenoid cystic carcinoma (subtype salivary gland cancer), where PSMA is expressed on the tumor cells. Although there is heterogeneity in PSMA expression and tracer uptake, a subset of patients with advanced salivary gland cancer, glioblastoma, thyroid cancer, hepatocellular carcinoma, and clear cell renal carcinoma show sufficient PSMA PET/CT tracer uptake in the tumor. These patients might potentially benefit from PSMA-RLT, so future research in this setting is encouraged. To date, ten patients with non-prostate solid cancers (salivary gland cancer, glioblastoma, thyroid cancer, hepatocellular carcinoma, and breast cancer) have been treated with PSMA-RLT and some beneficial effects were seen, making this an interesting topic for further exploration.

Declarations
Ethics approval This article does not contain any studies with human participants or animals performed by any of the authors.

Conflict of interest
The authors declare no competing interests.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.