Prostate-Specific Membrane Antigen in Anaplastic and Poorly Differentiated Thyroid Cancer—A New Diagnostic and Therapeutic Target?

Simple Summary Prostate-specific membrane antigen (PSMA)-targeted therapy represents a promising therapeutic option for the treatment of advanced carcinoma, but data on the relevance of PSMA-targeted diagnostics and therapy for anaplastic (ATC) and poorly differentiated (PDTC) thyroid carcinoma are still lacking. Due to the limited therapeutic options for these entities, the aim of this study was to evaluate to what extent Gallium-68 (68Ga)-PSMA-positron emission tomography/computed tomography (PET/CT) is superior to F-18-Fluordeoxyglucose (18F-FDG)-PET/CT in the diagnostic of ATC/PDTC and could represent a new therapeutic option in terms of theranostic. The findings of this study confirm the high diagnostic sensitivity and superiority of 18F-FDG-PET/CT in comparison to 68Ga-PSMA-PET/CT in the diagnosis of ATC and PDTC. However, it can be suggested that 68Ga-PMSA-PET/CT can be considered as a beneficial adjunct to the well-established 18F-FDG-PET/CT for a few individual selected patients with ATC and PDTC to detect lesions not discovered by 18F-FDG-PET/CT and to determine patients’ eligibility for a radioligand therapy. With regard to our results, radiolabelled PSMA-ligands with Lutetium-177 (177Lu)-PSMA may, in the future, represent a theranostic approach with only minor side effects for a few individual selected patients with ATC and PDTC who need alternative treatment options in case of progression when established therapies are no longer effective. Abstract Several studies have demonstrated an expression of the prostate-specific membrane antigen (PSMA) in the cancer-related neovasculature of thyroid malignancies. Due to the poor prognosis and limited therapeutic options for patients with anaplastic (ATC) and poorly differentiated (PDTC) thyroid carcinoma, the aim of our study was to investigate the theranostic approach of PSMA expression in these patients. The PSMA uptake on Gallium-68 (68Ga)-PSMA-positron emission tomography/computed tomography (PET/CT) and glucose uptake on F-18-Fluordeoxyglucose (18F-FDG)-PET/CTs were analysed in two ATC and six PDTC patients. The PSMA expression in corresponding patients’ tissue samples was detected by immunohistochemistry. In addition, various tissue sections from 22 ATC and six PDTC patients were examined concerning PSMA expression. 68Ga-PSMA-PET/CT showed heterogeneous PSMA expression among patients and lesions. Six of the eight analyzed patients (two ATC, four PDTC) showed increased glucose metabolism without increased PSMA uptake after PET/CT. In one patient (PDTC), 18F-FDG-PET/CT tracer uptake was positive and 68Ga-PSMA-PET/CT showed heterogeneous results. Another patient (PDTC) evidenced only PSMA-positive lesions and received two cycles of Lutetium-177 (177Lu)-PSMA therapy, which kept his disease stable for seven months. There was a correlation between immunohistochemical PSMA expression and uptake on 68Ga-PMSA-PET/CT in three of the examined patients. Twenty-seven of the analyzed 39 ATC and 13 of the analyzed 22 PDTC tissue sections showed a strong PSMA expression. Considering the rarity of PDTC and ATC, which is the reason for the small patient population we studied, the findings of this study confirm the high diagnostic sensitivity and superiority of 18F-FDG-PET/CT in comparison to 68Ga-PSMA-PET/CT in the diagnosis of ATC and PDTC. However, it can be suggested that 68Ga-PMSA-PET/CT can be considered as a beneficial adjunct to the well-established 18F-FDG-PET/CT for a few individual selected patients with ATC and PDTC to detect lesions not discovered by 18F-FDG-PET/CT and to determine patients’ eligibility for a radioligand therapy. Radiolabelled PSMA-ligands may, in the future, represent a theranostic approach with only minor side effects for a few individual selected patients with ATC and PDTC who need alternative treatment options in case of progression when established therapies are no longer effective. However, due to the small sample size of our collective, larger studies are needed to allow for a final evaluation on the significance of PSMA-targeted diagnostic and therapy for ATC and PDTC.


Introduction
Anaplastic (ATC) and poorly differentiated thyroid carcinoma (PDTC) represent less than 5% of thyroid cancers [1,2]. However, they are clinically relevant entities because of their extremely poor prognosis depending on their aggressive behaviour and high metastatic potential. ATC accounts only for approximately 1-2% of thyroid cancer. Patients affected by ATC have a mean survival time of 3-6 months and a mortality rate exceeding 90% [3][4][5][6][7]. The prognosis of PDTC, whose incidence varies between 1-15% is slightly better with a median 5-year-survival of 62-85%; nevertheless, PDTC is responsible for the majority of all fatalities from non-anaplastic follicular cell derived thyroid cancer [8][9][10][11]. Despite multimodal treatment strategies based on surgical resection in combination with radiochemotherapy and novel targeted therapies, including multikinase inhibitors (mKI), the prognosis of ATC and PDTC could not be significantly improved over the last 20 years [1,7,11,12]. New options to develop innovative diagnostic and therapeutic approaches for the treatment of these aggressive carcinomas have arisen in the field of theranostics and, in particular, by detecting prostate-specific membrane antigen (PSMA) [1,5,[13][14][15][16]. Theranostics using radiolabelled PSMA-ligands have gained significant influence in the management of prostate cancer. Meanwhile, targeted radionuclide therapy-for example, with Lutetium-177 ( 177 Lu) labelled PSMA peptides-represents a promising treatment option for castration-resistant prostate cancer, as currently confirmed in the VISION study (ClinicalTrials.gov Identifier: NCT03511664) [17,18]. Increased PSMA expression has also been observed in other carcinomas [19,20]. Interestingly, the overexpression of PSMA was histopathologically detected in the neovasculature of PDTC and ATC. Some case reports have already shown a high uptake on Gallium-68 ( 68 Ga)-PSMApositron emission tomography/computed tomography (PET/CT) in the primary tumor, cervical and mediastinal lymph nodes in ATC. However, the value of theranostic concepts using radiolabelled PSMA ligands for the diagnostic and treatment of ATC and PDTC remains still unclear [21].
Therefore, the aim of this study was to investigate the expression of PSMA in tissues from patients suffering from ATC and PDTC and to correlate the extent of expression with the feasibility and utility of the imaging with 68 Ga-PSMA-PET/CT. In particular, we evaluated to what extent 68 Ga-PSMA-PET/CT is superior to F-18-Fluordeoxyglucose ( 18 F-FDG)-PET/CT in the diagnostic of ATC/PDTC and could represent a new therapeutic option in terms of theranostics for these carcinomas.

Patient Selection
This is a single-center retrospective pilot study approved by Marburg University Hospital ethics committee (Az. 27/21), which included two patients with histologically confirmed ATC and six patients with PDTC treated between 2007 and 2021 at the Department of Visceral, Thoracic and Vascular Surgery and the Department of Nuclear Medicine at the Philipps University Hospital of Marburg. All of the patients enrolled in this study received a 68 Ga-PSMA-PET/CT in addition to 18 F-FDG-PET/CT. Data on patients' characteristics, including gender, age, treatment history, histopathological findings and information on diagnostic procedures involving 68 Ga-PSMA-PET/CT, in addition to 18 F-FDG-PET/CT, were retrieved from the medical files. Corresponding to the clinical data and PET/CT results, the PSMA expression in immunohistochemistry (IHC) was detected in paraffin-embedded formalin-fixed tissue sections from primary tumors, lymph nodes, pulmonary, soft tissue and osseous metastases, as well as from local tumor recurrences from these patients. We included one patient with PDTC for PET/CT evaluation in whom no tissue could be examined.

Imaging Protocol and Analyses of 18 F-FDG-and 68 Ga-PSMA-PET/CT
Both 18 F-FDG-PET/CT and 68 Ga-PMSA-PET/CT were performed according to the recommendations of the German Society of Nuclear Medicine [22].
When 18 F-FDG-PET/CT was performed, a median activity of 265.5 MBq (range 248-297 MBq) Fluorine-18-FDG was applied. The median blood glucose value at the time of application was 99.5 mg/dL (range 73-122 mg/dL). After a median time of 60 min after application, image acquisition was performed, which was contrast media-enhanced in five patients. 68 Ga-PMSA-PET/CT was performed after application of a median activity of 168 MBq (range 140-212 MBq) Gallium-68-PSMA after a median time of 60 min. This examination was performed with contrast agent in three patients.
An experienced nuclear medicine physician evaluated the PET/CTs. The experienced physician considered a lesion positive, after visually sufficient increased tracer accumulation.

Tissue Selection
In addition to the tissue samples from the seven patients who received a 68 Ga-PSMA-PET/CT, PSMA expression was further detected in tissue sections from 22 ATC patients and six PDTC patients that have previously undergone surgery or diagnostic punctures at the Department of Visceral, Thoracic and Vascular Surgery or the Department of Nuclear Medicine at the Philipps University Hospital of Marburg.
Overall, the histopathological analyses included 22 ATC primary tumors, five ATC lymph node metastases, four pulmonary/pleural metastases of ATC, seven metastases of the soft tissue surrounding the thyroid gland and one ATC presternal metastasis. Furthermore, six PDTC primary tumors, six PDTC lymph node metastases, three pulmonary metastases, one renal metastasis, two local recurrences and four osseous metastases of PDTC were evaluated with regard to their PSMA expression.

Immunohistochemical Staining for PSMA Antigens
Immunohistochemistry (IHC) was performed on 1,2 µm-thick paraffin-embedded formalin-fixed tissue sections. Sections were deparaffinised in xylene and subsequently rehydrated through ethanol. Heat-induced epitope retrieval was performed with trilogy. Staining was performed on a DAKO auto-stainer plus. After blocking of endogene peroxidase, sections were incubated for 45 min with mouse monoclonal anti-PSMA antibody (1:4000; ABIN 1302364, clone GCP-04; antibodies-online GmbH, Aachen, Germany). Sections were washed and incubated with Dako REAL EnVision HRP Rabbit/Mouse poly-mer (Agilent Technologies Deutschland GmbH, Waldbronn, Germany), which reacts with DAB-Chromogen, according to the manufacturer s protocol.

Assessment of PSMA Expression
Assessment of PSMA expression was calculated with the method described by Heitkötter et al. [23,24] and evaluated by an experienced pathologist (CK) on immune-stained whole ATC and PDTC slides. PSMA expression in either tumor cells or tumor-associated vessels was considered positive. Staining intensity was scored semiquantitatively as negative (0), weak (1 = barely perceptible staining at higher (400×) magnification), moderate (2 = readily apparent at lower power (40×) magnification) or strong (3). The fraction of PSMA positive cells was assessed as <5% or >5%. In the case of heterogeneous staining, the predominant pattern was recorded. A weak (1) or moderate (2) staining intensity in <5% of the neovasculature and a weak (1) staining intensity in >5% of the neovasculature was allocated to the "low expression" group (PSMA-labelling index = 1), whereas a moderate (2) staining intensity in >5% of the neovasculature and a strong (3) staining intensity in < or > 5% of the neovasculature was assigned to the "strong expression" group (PSMA-labelling index = 2). In contrast to Heitkötter et al., we also implemented the PSMA-labelling index = 3, which classified a strong (3) staining intensity not only in the neovasculature, but also in the tumor cells themselves.

Patient Characteristics
Two male patients with ATC and six patients with PDTC (four male, two female) underwent a 68 Ga-PSMA-PET/CT in addition to a 18 F-FDG-PET/CT. The median age of the patients suffering from PDTC was 60 years (range 44-70 years) at initial diagnosis. All patients developed distant metastatic disease: (a) pulmonary metastases in three PDTC patients over a median time span of 13 months (range 0-56 months); (b) osseous metastases in four patients with a median occurrence after initial diagnosis of 7.5 months (range 0-41 months); (c) cerebral metastasis in one patient 44 months after the first diagnosis, respectively. This patient also had evidence of inguinal lymph node metastasis and soft tissue metastasis of the thoracic wall 41 months after the first diagnosis (patient 7).
Furthermore, we evaluated a 59-and 47-year-old patient with ATC (patient 1 and 2). One of these patients already had an advanced Stage IVC (Union internationale contre le cancer (UICC) (Brierley J, Gospodarowicz MK (Mary K., Wittekind C (Christian). TNM classification of malignant tumours. 253 p.) carcinoma with evidence of lymph node and pulmonary metastases at the time of initial diagnosis (patient 2). In the other ATC patient, the presence of pulmonary metastases was detected eight months after initial diagnosis (patient 1). Patients characteristics and previous treatments are shown in Table 1.  1 According to the 8th Edition of UICC/TNM Classification 2 COSMIC study: Cabozantinib versus placebo in second line therapy of radioiodine refractory thyroid cancer. Abbreviations: M = male, F = female, mKI = multikinaseinhibitor.

68 Ga-PSMA-PET/CT
The overall median time span between initial diagnosis and the prior PET of the particular patient examined in this study was 20.5 months (range 5-112 months)-for ATC seven and eight months and for PDTC 33.5 months (range 5-112 months).
In each case, we evaluated the 18 F-FDG-PET/CT with the shortest time interval to the performed 68 Ga-PSMA-PET/CT. In seven patients, a 18 F-FDG-PET/CT was performed previously and the interval time between this prior 18 F-FDG-PET/CT and 68 Ga-PSMA-PET/CT in temporal connection was median 76 days (range 1-232 days). Only one of the evaluated PDTC patients (patient 6) received a 68 Ga-PSMA-PET/CT first and subsequently underwent further imaging via 18 F-FDG-PET/CT 171 days later. 68 Ga-PSMA-PET/CT was performed in this patient in the setting of theranostics when therapeutic options were exhausted after two different mKI (sorafenib and lenvatinib) and re-challenge of one mKI (sorafenib) (patient 6). Therefore, the overall evaluation time interval between 18 F-FDG-PET/CT and 68 Ga-PSMA-PET/CT amounted to a median 87.5 days (range 1-232 days). These data are included in the Table 2. Table 2. Lesion SUVmax in 18 FDG-PET/CT and 68 Ga-PSMA-PET/CT of (metastatic) disease locations and extent of PSMA expression in tissue samples for each patient.   Patients, in whom 18 F-FDG-PET/CT uptake was positive and 68 Ga-PSMA-PET/CT negative. This was the case in six of the eight patients evaluated (patient 1-6).

2.
Patients, in whom 18 F-FDG-PET/CT uptake was positive and 68 Ga-PSMA-PET/CT showed heterogeneous results. This was the case in one of the eight patients evaluated (patient 7).

3.
Patients, in whom 68 Ga-PSMA-PET/CT uptake was clearly positive and thus a therapeutic option in terms of theranostics-for example, with a 177 Lu-PSMA therapyexists. This was the case in one of the eight patients evaluated (patient 8).
In the following, we will describe a detailed case example for each of the three possibilities mentioned above. These data can be found along with further information on each patient in Tables 1 and 2. For a better overview we evaluated the five most conclusive metastases, if more than ten metastases of one site were present in a patient. As part of the follow-up after performing thyroid surgery, as well as radiochemotherapy and initiation of mKI therapy, 18 F-FDG-PET/CT and 68 Ga-PMSA-PET/CT were performed seven months after initial diagnosis. The time interval between these two examinations was one day.
Imaging analysis: 18 F-FDG-PET/CT showed in the area of the left thyroid bed near the esophagus a soft tissue lesion of 1.9 cm with increased glucose utilization (standardized uptake value (SUV)max 16.3); 68 Ga-PMSA-PET/CT did not show pathologically increased tracer accumulation at this location (SUVmax 2.5) ( Figure 1). Furthermore, 18 F-FDG-PET/CT revealed one lymph node of 2.2 cm with slightly increased glucose accumulation in the mediastinum with suspicion of lymph node metastasis, which did not show clearly increased uptake on 68 Ga-PMSA-PET/CT (SUVmax 3.3) and seven pleural lesions between 0.3-1.2 cm with increased glucose uptake (SUVmax 2.7 (range 1.5-7.9). None of these lesions showed increased PSMA uptake (SUVmax 1.3 (range 0.6-2.6).
performed seven months after initial diagnosis. The time interval between these two examinations was one day.
Imaging analysis: 18 F-FDG-PET/CT showed in the area of the left thyroid bed near the esophagus a soft tissue lesion of 1.9 cm with increased glucose utilization (standardized uptake value (SUV)max 16.3); 68 Ga-PMSA-PET/CT did not show pathologically increased tracer accumulation at this location (SUVmax 2.5) (Figure 1). Furthermore, 18 F-FDG-PET/CT revealed one lymph node of 2.2 cm with slightly increased glucose accumulation in the mediastinum with suspicion of lymph node metastasis, which did not show clearly increased uptake on 68 Ga-PMSA-PET/CT (SUVmax 3.3) and seven pleural lesions between 0.3-1.2 cm with increased glucose uptake (SUVmax 2.7 (range 1.5-7.9). None of these lesions showed increased PSMA uptake (SUVmax 1.3 (range 0.6-2.6).  IHC analysis: The primary tumor as well as tumor infiltration of the soft tissue surrounding the thyroid gland could be pathologically investigated concerning the PSMA expression-both samples showed a "strong expression" (PSMA-labelling index = 2) ( Figure 2).
In summary: This patient showed multiple metastases with increased glucose metabolism without increased PSMA uptake on PET/CT, despite a strong PSMA expression in the examined tissue samples.
IHC analysis: The primary tumor as well as tumor infiltration of the soft tissue surrounding the thyroid gland could be pathologically investigated concerning the PSMA expression-both samples showed a "strong expression" (PSMA-labelling index = 2) ( Figure 2).
In summary: This patient showed multiple metastases with increased glucose metabolism without increased PSMA uptake on PET/CT, despite a strong PSMA expression in the examined tissue samples. Due to the fact that patient 2, 3, 4, 5 and 6, all suffering from ATC/PDTC, showed similar to patient 1 metastases of ATC/PDTC with increased glucose metabolism on 18 F-FDG-PET/CT without an increased PSMA uptake on PET/CT, we provide further detailed information to these six patients in Tables 1 and 2  Due to the fact that patient 2, 3, 4, 5 and 6, all suffering from ATC/PDTC, showed similar to patient 1 metastases of ATC/PDTC with increased glucose metabolism on 18 F-FDG-PET/CT without an increased PSMA uptake on PET/CT, we provide further detailed information to these six patients in Tables 1 and 2     IHC analysis: The primary tumor, as well as a cervical lymph node metastasis, were examined with regard to their PSMA expression and revealed a "strong expression" (PSMA-labelling index = 2).
In summary: There was a heterogeneous PSMA expression of the metastases on PET/CT and a strong PSMA expression in the examined tissue samples. Imaging analysis: While the first performed 18 F-FDG-PET/CT showed no lesions with increased glucose metabolism, four osseous lesions with clearly (9th thoracic vertebra 1.6 cm, SUVmax 7.1; Os ilium right 1.2 cm, SUVmax 9.0) or discretely (7th rib without morphological correlate, SUVmax 2.9; acetabulum left 1.3 cm, SUVmax 4.0) pathologically increased PSMA expression (overall median SUVmax 5.55 (2.9-9.0)) were detected after the 68 Ga-PSMA-PET/CT, which was performed 81 days later (Figures 5 and 6). IHC analysis: The primary tumor, as well as a cervical lymph node metastasis, were examined with regard to their PSMA expression and revealed a "strong expression" (PSMAlabelling index = 2).
In summary: There was a heterogeneous PSMA expression of the metastases on PET/CT and a strong PSMA expression in the examined tissue samples. Imaging analysis: While the first performed 18 F-FDG-PET/CT showed no lesions with increased glucose metabolism, four osseous lesions with clearly (9th thoracic vertebra 1.6 cm, SUVmax 7.1; Os ilium right 1.2 cm, SUVmax 9.0) or discretely (7th rib without morphological correlate, SUVmax 2.9; acetabulum left 1.3 cm, SUVmax 4.0) pathologically increased PSMA expression (overall median SUVmax 5.55 (2.9-9.0)) were detected after the 68 Ga-PSMA-PET/CT, which was performed 81 days later (Figures 5 and 6).  A second 68 Ga-PSMA-PET/CT which was not included in our comparative analysis and which was conducted 187 days later, revealed a disease progression with multiple lesions with clearly pathologically increased PSMA expression (median SUVmax 10.0 (6.0-15.0) (Figure 7). Therefore, the patient received two cycles of 177 Lu-PSMA therapy in A second 68 Ga-PSMA-PET/CT which was not included in our comparative analysis and which was conducted 187 days later, revealed a disease progression with multiple lesions with clearly pathologically increased PSMA expression (median SUVmax 10.0 (6.0-15.0) (Figure 7). Therefore, the patient received two cycles of 177 Lu-PSMA therapy in the sense of theranostics within the framework on an individual therapeutic approach (Figures 8 and 9). Based on the experience gained in the treatment of prostate and neuroendocrine carcinomas, 6.3 GBq were applied intravenously in the first cycle and 7.4 GBq were administered in the second cycle eight weeks later. Despite the fact that no side effects were evidenced after the first cycle, the patient experienced temporary nausea after the second one. During follow-up, stable disease was described until seven months after the first cycle, after which there was an increase in thyroglobulin levels; another three months later, FDG-positive metastases appeared for the first time in an 18 F-FDG-PET/CT performed in the context of the tumor marker elevation.
IHC analysis: Unfortunately, no tissue sample could be examined for this patient.

Imaging Findings
Disease locations, median lesions SUVmax and the following analysis are listed in Table 3. Out of 11 evaluated ATC lesions, 10 were clearly visually FDG-positive (detection rate 91%) and none were clearly PSMA-positive (detection rate 0%). The detection rate for PDTC lesions was 89% for 18 F-FDG-PET/CT (out of 37 evaluated PDTC lesions 33

Imaging Findings
Disease locations, median lesions SUVmax and the following analysis are listed in Table 3. Out of 11 evaluated ATC lesions, 10 were clearly visually FDG-positive (detection rate 91%) and none were clearly PSMA-positive (detection rate 0%). The detection rate for PDTC lesions was 89% for 18 F-FDG-PET/CT (out of 37 evaluated PDTC lesions 33

Imaging Findings
Disease locations, median lesions SUVmax and the following analysis are listed in Table 3. Out of 11 evaluated ATC lesions, 10 were clearly visually FDG-positive (detection rate 91%) and none were clearly PSMA-positive (detection rate 0%). The detection rate for PDTC lesions was 89% for 18 F-FDG-PET/CT (out of 37 evaluated PDTC lesions 33 were visually FDG-positive) and 11% for 68 Ga-PMSA-PET/CT (out of 36 evaluated PDTC lesions four were visually PSMA-positive). * If only one lesion of a disease location is present, only the SUVmax for the respective lesion is given. 1 The patient with surgery of known FDG-positive vertebral metastases after 18 FDG-PET/CT was excluded for absence of comparability. 2 Only the clearly discordant lesions are described here. Lesions that had a discrete and thus not clearly positive uptake were not evaluated. 3 Of ten lesions, three are discretely PSMA-and FDG-positive, which is why they are not evaluated as clearly discordant. 4 Of two lesions, one is discretely PSMAand FDG-positive, which is why it is not evaluated as clearly discordant. 5 Of four lesions, one is PSMA-and FDG-positive (concordant); one is discretely PSMA-and FDG-positive, which is why it is not evaluated as clearly discordant. + = positive; − = negative.
In ATC patients, discordant lesions with clearly increased FDG uptake and no clear PSMA expression were observed in pulmonary metastases (9/9) and local tumor recurrence/thyroid bed lesions (1/1). We observed no discordant lesion with explicit/increased PSMA expression and no or decreased FDG uptake in ATC.

Correlation between the Extent of Immunohistochemical PSMA Expression in Tissue Samples and Imaging in 68 GA-PSMA-PET/CT with Regard to Different Metastatic Sides of ATC and PDTC Patients
In seven of the eight patients we examined, tissue samples of metastatic sides could be evaluated with regard to their PSMA expression and compared with the respective imaging results on 68 Ga-PSMA-PET/CT (Table 4). In this regard, we only evaluated the metastases for which there was also a specific reproducibility in 68 Ga-PSMA-PET/CT.
An increased PSMA expression in ATC or PDTC tissue samples and concordantly PSMA-positive lesions in 68 Ga-PSMA-PET/CT was only observed in a cervical lymph node metastasis of a patient with PDTC (strong PSMA expression and discreet PSMA-positive lesion in PET/CT). In all other tissue samples examined, there was either a discordance between PSMA expression and uptake on 68 Ga-PSMA-PET/CT or a low PSMA expression in IHC was associated with an absent uptake on PET/CT. In this regard, however, it must be noted that from patient 8 in particular, who evidenced clearly PSMA-positive lesions in PET/CT, no tissue samples were available and therefore could not be included in this investigation.  (Table 5) and was-with one exception-detected exclusively in the tumor microenvironment (tumor-associated neovasculature), but not in the tumor cells themselves. 27 of the analyzed ATC tissue sections showed a strong PSMA expression and nine displayed a weak expression. In 13 ATC patients, the PSMA expression was analyzed in the primary tumor as well as in existing metastases. Nine of these 13 patients displayed a concordantly high PSMA expression in the primary tumor as well as in the metastases; the PSMA expression was higher in the primary tumor than in the metastases in only four patients. In none of the analyzed ATC samples was the PSMA expression higher in the metastases than in the primary tumor.
13 of the analyzed PDTC tissue sections displayed a strong and six a weak PSMA expression in the tumor-associated neovasculature. From a total of six PDTC patients, the primary tumor and tissue sections of metastases were examined. In this context, PSMA expression in the primary tumor and associated metastases were consistently high in four of these patients. In one patient, a high PSMA expression was detected in the primary tumor and in the pulmonary metastasis, while the osseous metastasis showed no PSMA expression. In the other patient, PSMA expression was low in the primary tumor and high in the corresponding pulmonary metastasis.
One PDTC tumor sample showed an unusually high PSMA expression with a strong membraneous and cytoplasmatic immunoreactivity; not only of the neovasculature, but also of the tumor cells themselves, compared to all other analyzed tissue samples ( Figure 10). Thus, we assigned a PSMA-labelling index = 3 for the evaluated primary tumor and associated cervical lymph node metastasis of this patient.

Discussion
PSMA-targeted therapy represents a promising therapeutic option for the treatment of advanced carcinomas and there are currently a number of PSMA-targeted therapeutic approaches being approved for clinical trials e.g., clear cell renal carcinoma, glioblastoma and hepatocellular carcinoma (ClinicalTrials.gov Identifier: NCT02607553, NCT02067156, NCT01777594, NCT01856933) [25]. PSMA-radioligands are, meanwhile, an established theranostic approach with excellent results and less severe side effects for the management of prostate cancer as shown in the VISION study (ClinicalTrials.gov Identifier: NCT03511664) [17,18,26,27]. In contrast, data on the relevance of PSMA-targeted diagnostic and therapy for ATC and PDTC are still lacking. Hence, this study evaluates, to the best of our knowledge, the largest collective of PDTC and ATC patients with advanced metastatic disease investigated by 68 Ga-PMSA-PET/CT, in terms of theranostics regarding an individual treatment concept in purpose of a 177 Lu-PSMA therapy. In addition, this study provides the first comparison between IHC and 68 Ga-PSMA-PET/CT for thyroid cancer patients.
In our cohort, the overall detection rate for PDTC/ATC lesions was 89/91% for the 18 F-FDG-PET/CT and 11/0% for 68 Ga-PMSA-PET/CT and thus much lower for the 68 Ga-PMSA-PET/CT than in prior reports. Verma et al. evaluated ten patients with meta-

Discussion
PSMA-targeted therapy represents a promising therapeutic option for the treatment of advanced carcinomas and there are currently a number of PSMA-targeted therapeutic approaches being approved for clinical trials e.g., clear cell renal carcinoma, glioblastoma and hepatocellular carcinoma (ClinicalTrials.gov Identifier: NCT02607553, NCT02067156, NCT01777594, NCT01856933) [25]. PSMA-radioligands are, meanwhile, an established theranostic approach with excellent results and less severe side effects for the management of prostate cancer as shown in the VISION study (ClinicalTrials.gov Identifier: NCT03511664) [17,18,26,27]. In contrast, data on the relevance of PSMA-targeted diagnostic and therapy for ATC and PDTC are still lacking. Hence, this study evaluates, to the best of our knowledge, the largest collective of PDTC and ATC patients with advanced metastatic disease investigated by 68 Ga-PMSA-PET/CT, in terms of theranostics regarding an individual treatment concept in purpose of a 177 Lu-PSMA therapy. In addition, this study provides the first comparison between IHC and 68 Ga-PSMA-PET/CT for thyroid cancer patients.
In our cohort, the overall detection rate for PDTC/ATC lesions was 89/91% for the 18 [28,29]. One possible explanation for the lower detection rates may be the small patient group collective analyzed in all other studies, which makes differences statistically distinctly divergent. Despite the fact that this study confirms the high diagnostic sensitivity of 18 F-FDG-PET/CT, it must be further investigated to what extent 68 Ga-PMSA-PET/CT may offer lower detection rates for dedifferentiated than for differentiated thyroid carcinomas.
According to our results, especially lymph node and bone metastases of dedifferentiated thyroid carcinoma can be visualized in 68 Ga-PSMA-PET/CT. The varying degree of PSMA expression in the metastases can be caused by different factors, of which neovascularization is of particular importance. The angiogenesis of thyroid carcinoma is controlled by multiple receptors, signaling pathways and genetic targets, which may vary between the primary tumor and different metastases [32]. Furthermore, it is known that the morphology of the tumor angiogenesis influences the tumor environment and can be inhomogeneous between different metastatic sides [33]. Considering these aspects of neovascularization and angioneogenesis, the spectrum of PSMA expression in the investigated metastases, which is mainly found in the tumor-associated neovasculature, can be partly explained.
This study confirmed the previous findings [28,31] highlighting that the lesion FDG uptake was higher than the PSMA uptake in 7/8 evaluated patients with an overall median lesion SUVmax of 5.0 for 18 F-FDG-PET/CT and 1.5 for 68 Ga-PMSA-PET/CT in ATC and 17.9 for 18 F-FDG-PET/CT and 4.0 for 68 Ga-PMSA-PET/CT in PDTC. Nevertheless, 68 Ga-PMSA-PET/CT was able to detect metastatic lesions in PDTC including two certain osseous metastases (and two suspicious osseous lesions), which could not be detected by 18 F-FDG-PET/CT imaging (patient 8), as well as one mediastinal lymph node metastasis and one cerebral metastasis (patient 7). Similar to our findings in patient 7 and 8, Lütje et al. evaluated six patients with iodine-negative metastatic differentiated thyroid carcinoma and found FDG-negative and PSMA-positive metastases in one patient, so we must emphasize at this point that we can expect to identify more of such patients in larger patient collectives, who may benefit from 68 Ga-PMSA-PET/CT in terms of staging [31].
For a reasonable implementation of individual treatment concepts, all lesions should have a high expression in order to be therapeutically targetable. In one PDTC patient (patient 8), all lesions showed a high PSMA expression; subsequently, as the disease progressed, he received two cycles of 177 Lu-PSMA therapy, from which he benefited for several months without experiencing serious side effects. Comparing this result with the results of de Vries et al. [30], who showed a slight, temporary response in one of two 177 Lu-PSMA treated patients and considering the lack of therapeutic options in these cases, PSMA expression offers a promising therapeutic target for a few individually selected PDTC and ATC patients who need alternative treatment options in case of progression when established therapies like tyrosinekinaseinhibitors are no longer effective. Nevertheless, it is important to consider that PSMA in thyroid carcinomas has been detected only in the tumor-associated neovasculature [23]. To achieve a therapeutic effect, the compounds need to the stored in/on/near the tumor cells, due to the short range of the emission (1.5 mm for 177 Lu). We were able to show visually intense nuclide accumulation in an early scan after 90 min and also in a late scan after at least 43 h of radioligand injection (patient 8), indicating that the neovasculature represents a sufficient target due to a long retention time.
Apart from its well known strong expression in prostate cancer cells, immunohistochemical examination of different solid cancer types, such as clear cell renal carcinoma or lung cancer, demonstrated, in addition to a substantial immunohistochemical PSMA expression, a high tracer accumulation after 68 Ga-PMSA-PET/CT imaging [34][35][36][37][38][39][40]. Heitkötter et al. analysed 101 thyroid lesions and found a significantly higher PSMA expression in PDTC and ATC tissue sections than in differentiated thyroid tumors [24]. It is therefore not surprising that Sollini et al. demonstrated that PSMA expression in differentiated thyroid carcinomas is associated with more aggressive behavior; therefore, higher expression would have to be expected in dedifferentiated carcinomas [41]. This aspect also fits to the pathogenesis of PSMA expression. It is assumed that PSMA is expressed as part of the autoregulation in tumorigenesis and facilitates endothelial cell invasion of angiogenesis [23].
Consistent with the findings from Heitkötter et al. [23], in this study, a strong PSMA expression was identified in the tumor-associated neovasculature of 27 out of 39 ATC and in 13 out of 22 PDTC tissue samples. One PDTC tumor sample even showed a strong PSMA expression of the tumor cells themselves. Furthermore, there was a correlation between immunohistochemical PSMA expression and PSMA uptake on 68 Ga-PMSA-PET/CT in three of the examined patients (patient 3 and 6 showed a low PSMA expression correspondingly with no PSMA uptake on PET/CT, patient 7 presented a strong PSMA expression correspondingly with an increased PSMA uptake on 68 Ga-PMSA-PET/CT). On the other hand, four patients enrolled in this study had a high PSMA expression in their tissue samples but no PSMA uptake on the PET/CT examination (patient 1, 2, 4 and 5). A clear explanation for the discrepancy between immunohistochemical PSMA expression in the evaluated tissue samples and uptake on 68 Ga-PSMA-PET/CT cannot be concluded from the currently available literature and represents, therefore, a topic for further investigations. With regard to prostate carcinoma, it is well known that there is a close correlation between tracer uptake on 68 Ga-PSMA-PET/CT and PSMA expression in IHC [42]. In contrast, a case study of a glioblastoma was recently published in which a low tracer accumulation was seen in 68 Ga-PSMA-PET/CT despite an intermediate PSMA expression in IHC [43]. One possible reason for this discrepancy could be that PSMA expression in apical membranes-typical for prostate carcinoma cells-is connected with a higher tracer uptake on PET imaging than the level of PSMA expression in cytoplasm, which is often observed in nonprostatic disease [44].
A relevant limitation of this study is the retrospective design and the small patient population due to the rarity of the analyzed carcinoma entities. In addition, the long time interval between 18 F-FDG-and 68 Ga-PSMA-PET/CT in some patients might bias our results. Nevertheless we can confirm the high diagnostic sensitivity and superiority of 18 F-FDG-PET/CT in comparison to 68 Ga-PSMA-PET/CT in the diagnosis of ATC and PDTC.
It is known from preliminary work of other research groups that both anaplastic and poorly differentiated thyroid carcinoma express GLUT1 (Glucose transporter 1) and that the associated high 18 F-FDG uptake reflects dedifferentiation of thyroid tumors. Therefore, radioiodine-negative metastases display increased glucose uptake more frequently than radioiodine-positive neoplasms. Furthermore, the up regulation of GLUTs in thyroid cancer seems to be associated with tumor aggressiveness and loss of differentiation [45][46][47]. As a proposal for a consecutive study, the immunohistochemical expression of GLUT1 and associated uptake on 18 F-FDG-PET/CT could be evaluated and compared to the PSMA expression and imaging in 68 Ga-PSMA-PET/CT in ATC and PDTC.
Larger prospective studies are urgently needed to conclusively evaluate the impact of PSMA-based diagnosis and therapy for patients suffering from ATC and PDTC.

Conclusions
In summary, considering the rarity of ATC and PDTC, however, the findings of this study confirm the high diagnostic sensitivity and superiority of 18 F-FDG-PET/CT in comparison to 68 Ga-PSMA-PET/CT in the diagnosis of ATC and PDTC. However, it can be suggested that 68 Ga-PMSA-PET/CT can be considered as a beneficial adjunct to the well-established 18 F-FDG-PET/CT for a few individual selected patients with advanced and metastatic dedifferentiated and anaplastic thyroid carcinomas to detect lesions not discovered by 18 F-FDG-PET/CT. In addition, PSMA-targeted therapy can be used as an alternative option in selected patients if they show progression under established therapeutic concepts. Current therapeutic options for ATC and PDTC are limited, so there is a desperate need for novel innovative and targeted treatment modalities. Even though there is a need for larger collectives to allow a final evaluation, radiolabelled PSMA-ligands with 177 Lu-PSMA may represent in future a theranostic approach with only minor side effects for a rare proportion of these patients. 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 upon request from the corresponding author. The data are not publicly available due to ethical restrictions.