The effect of androgen blockade on [⁶⁸Ga]Ga-PSMA-11 and [¹⁸F]FDG PET uptake in patients with recurrent or metastatic salivary duct carcinoma: a prospective imaging study

Abstract

Background

The expression of prostate-specific membrane antigen (PSMA), a target for oncological imaging and treatment, is upregulated by androgen blockade in prostate cancer. Salivary duct carcinoma (SDC), an aggressive histological subtype of salivary gland cancer, resembles prostate cancer in terms of PSMA and androgen receptor (AR) expression. A similar upregulation of PSMA in SDC would have implications for future studies with PSMA-targeted imaging and therapy. Additionally, FDG PET/CT scans are frequently used for SDC imaging, but the effect of androgen blockade on FDG uptake is unknown. This study investigated the effect of combined androgen blockade (CAB) on tumour PSMA and FDG uptake in patients with SDC.

Results

Eight patients with recurrent and/or metastatic AR-positive SDC who started CAB (goserelin plus bicalutamide) as standard of care were prospectively enrolled. [⁶⁸ Ga]Ga-PSMA-11 and [¹⁸F]FDG PET/CT scans were performed within 21 days before and 21 ± 7 days after CAB initiation. PET parameters, including SUV_max, were obtained for PSMA and FDG positive lesions. A total of 80 metastatic lesions were analysed on a per-lesion basis. SUV_max changes after CAB initiation were categorised as increased (≥ + 20%), stable (between -20% and + 20%), or decreased (≤ -20%). The PSMA SUV_max increased in 20 lesions (25%), remained stable in 46 lesions (58%), and decreased in 14 lesions (18%), with no significant overall change (Wilcoxon signed rank test, p = 0.74). The FDG SUV_max increased in 35 lesions (44%), remained stable in 39 lesions (49%), and decreased in 6 lesions (8%), with a significant overall median increase of + 0.97 (Wilcoxon signed rank test, p < 0.001). The median PFS was 2.2 months (95% confidence interval 1.7–2.7 months).

Conclusions

Androgen blockade in patients with recurrent and/or metastatic SDC did not induce a significant increase in tumour PSMA uptake after three weeks. In contrast, tumour FDG uptake increased significantly after three weeks of CAB, which may reflect the poor tumour response in this cohort and/or a transient treatment-related effect.

Trial registration: ClinicalTrials.gov, NCT04214353. Registered 13 December 2019.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13550-025-01275-x.

Background

Salivary duct carcinoma (SDC) is an aggressive histological subtype of salivary gland cancer [1, 2]. Approximately 50% of patients develop recurrent and/or metastatic (R/M) disease [3]. In the R/M setting, the median overall survival (OS) is five months with best supportive care only [4]. As 78–96% of SDC tumours express the androgen receptor (AR), combined androgen blockade (CAB) consisting of a gonadotropin-releasing hormone analogue plus bicalutamide is commonly used [3–6]. In a phase II trial, CAB treatment resulted in an objective response rate of 42%, a median progression-free survival (PFS) of 8.8 months, and a median OS of 30.5 months [7]. Additionally, human epidermal growth factor receptor 2 (HER2) can be targeted in cases with positive receptor status (29–46%) [8]. Other systemic treatment options are limited, making exploration of novel targets imperative [9–11].

The theranostic target prostate-specific membrane antigen (PSMA) has revolutionized both imaging and therapy in prostate cancer. PSMA positron emission tomography (PET) results in superior detection rates compared to conventional imaging [12], and substantial efficacy of PSMA-targeted radioligand therapy has been achieved [13–17]. Interestingly, in prostate cancer, androgen blockade increases PSMA uptake after 1–4 weeks. This effect is attributed to the abolishment of androgen-mediated suppression of the PSMA-encoding gene folate hydrolase 1 (FOLH1) [18–29].

PSMA is also an interesting theranostic target for SDC, where it is primarily expressed in the tumour-associated neovasculature [30]. Prior research has demonstrated clinically relevant PSMA uptake in SDC, with a tumour-to-liver ratio > 1 in 20–40% of patients [30, 31]. However, initial PSMA-targeted radioligand therapy outcomes in SDC have been unfavourable [31]. Since androgen blockade is standard of care for patients with AR-positive SDC and it has shown to increase PSMA uptake in prostate cancer, we hypothesize that CAB also enhances PSMA uptake in SDC. A CAB-induced increase of PSMA uptake in SDC would have implications for the timing and sequencing of PSMA-targeted therapy and for the interpretation of PSMA PET/CT scans in future studies. Furthermore, FDG PET/CT scans are frequently used for SDC imaging, but the effect of CAB on FDG uptake is unknown [32–41]. Therefore, we aim to evaluate the effect of CAB on tumour PSMA and FDG uptake in R/M SDC.

Methods

Study design and patient selection

This single-centre, single-arm, prospective pilot study was conducted at Radboud university medical center, a salivary gland cancer expertise centre. Eligible patients were aged ≥ 18 years, had AR-positive R/M SDC, and intended to start CAB (goserelin 10.8 mg subcutaneously every 12 weeks and bicalutamide 50 mg once daily) as standard of care. Patients required ≥ 1 lesion ≥ 1.5 cm in diameter. Exclusion criteria were contraindications for PET/CT imaging or impaired renal (MDRD < 30 ml/min/1.73 m²) or liver function (AST/ALT ≥ 2.5 × the upper limit of normal).

Imaging

[⁶⁸ Ga]Ga-PSMA-11 PET/CT and [¹⁸F]FDG PET/CT scans were performed at baseline (pre-CAB) and repeated after CAB initiation (post-CAB) (Fig. 1). Pre-CAB scans were performed within 21 days before CAB initiation, with a maximum interval of 21 days. Based on prior studies in prostate cancer [22, 25], post-CAB scans were performed after 21 (± 7) days of CAB, with a maximum interval of three days. [⁶⁸ Ga]Ga-PSMA-11 was produced as described by Eder et al. [42]. Patients were injected with 2.0 MBq/kg of [⁶⁸ Ga]Ga-PSMA-11 and 2.1 MBq/kg of [¹⁸F]FDG. PET/CT imaging was performed 60 min post-injection on a Biograph mCT system (Siemens Healthineers, Erlangen, Germany). Acquisition and reconstruction parameters are provided in the Additional file 1.

Fig. 1.

Study design. [¹⁸F]FDG = 2-deoxy-2-[¹⁸F]fluoro-D-glucose; ⁶⁸ Ga = gallium-68; CAB = combined androgen blockade; CT = computed tomography; OPD = outpatient department; PET = positron emission tomography; PSMA = prostate-specific membrane antigen; R/M = recurrent and/or metastatic; SDC = salivary duct carcinoma

Image analysis

An experienced nuclear medicine physician (MvR) reviewed the PET/CT scans and identified SDC lesions characterised by pathological PSMA and/or FDG uptake. All lesions with well-defined margins were delineated by average-based iterative thresholding on PSMA and FDG PET/CT scans, using PMOD 4.4 software (PMOD Technologies, Zurich, Switzerland) [43]. The following parameters were obtained: SUV_max, SUV_mean, and tumour volume. Tumour volumes derived from PSMA and FDG PET/CT scans were termed PSMA-derived tumour volume and metabolic tumour volume, respectively. Total lesion PSMA was calculated as the product of PSMA-derived tumour volume and PSMA SUV_mean, while total lesion glycolysis was determined as the product of metabolic tumour volume and FDG SUV_mean [44]. Post-CAB SUV_max changes were categorised as increased (≥ + 20%), stable (between −20% and + 20%), or decreased (≤ −20%).

Follow-up and clinical outcomes

Patients visited the outpatient department at six weeks after CAB initiation, followed by CT scans and outpatient department visits from three months after CAB initiation at three-month intervals. Tumour response, PFS, and OS were determined. Tumour response was evaluated according to RECIST 1.1 [45]. PFS was defined as the duration from CAB initiation to clinical progression, radiographic progression per RECIST 1.1, or death. OS was defined as the duration from CAB initiation to death.

Statistical analysis

Data were analysed on a per-lesion basis. To compare pre- and post-CAB PET parameters, the paired Wilcoxon signed-rank test was applied. Associations between PSMA SUV_max and FDG SUV_max were evaluated using Spearman’s r. Time-to-event outcomes were analysed using Kaplan–Meier statistics. A p-value < 0.05 was considered statistically significant. Data were analysed using SPSS 29.0 (IBM, Armonk, NY, USA).

Results

Eight patients enrolled in the study and completed all PET/CT scans. Table 1 summarises patient characteristics. All eight patients had metastatic disease, and three patients had locoregional disease as well. All patients had lymph node metastasis, making it the most common metastatic site, followed by lung and bone metastasis each in five patients. Seven patients received CAB as first-line treatment, while the only patient with positive HER2 status received CAB as third-line treatment after HER2-targeted therapies. At CT evaluation three months after CAB initiation, one patient had a partial response, two patients had stable disease, and two patients had progressive disease. Three patients showed disease progression before the three-month evaluation (Fig. 2). The median PFS was 2.2 months (95% confidence interval 1.7–2.7 months) and the median OS was 12.9 months (95% confidence interval 8.3–17.5 months).

Table 1.

Patient characteristics

Characteristic
Age, median, years (range)	72 (54–79)
Sex, n (%)
Male	6 (75%)
Female	2 (25%)
Primary tumour site, n (%)
Parotid gland	6 (75%)
Submandibular gland	2 (25%)
Disease status
Metastatic disease	5 (62.5%)
Metastatic and locoregional disease	3 (37.5%)
Metastatic site, n (%)
Lymph nodes	8 (100%)
Bone	5 (62.5%)
Lung	5 (62.5%)
Brain	1 (12.5%)
Liver	1 (12.5%)
Cutaneous lymphangitis carcinomatosa	1 (12.5%)
Prior local treatment, n (%)
Resection of primary tumour	5 (62.5%)
Postoperative radiotherapy	5 (62.5%)
Palliative radiotherapy primary tumour	1 (12.5%)
Palliative radiotherapy metastasis	3 (37.5%)
Prior systemic treatments for R/M disease, n (%)
0	7 (87.5%)
1	0
2	1 (12.5%)
AR positivity (immunohistochemistry)
90–100%	6 (75%)
70–80%	1 (12.5%)
40–70%	1 (12.5%)
0–40%	0
AR staining intensity (immunohistochemistry)
Strong	6 (75%)
Moderate	2 (25%)
Weak	0
HER2 status,a n (%)
Positive	1 (12.5%)
Negative	7 (87.5%)

AR androgen receptor, HER2 human epidermal growth factor receptor 2, R/M recurrent and/or metastatic disease

^aHER2 positivity defined as a HER2 immunohistochemical staining score of 3 + or a HER2 immunohistochemical staining score of 2 + combined with a positive HER2 FISH

Fig. 2.

Swimmer plot starting at diagnosis of recurrent and/or metastatic (R/M) disease, illustrating the systemic therapies that study participants received until death or last follow-up, and the tumour response on combined androgen blockade (CAB) treatment. The colour scheme represents all systemic therapies, while the symbols indicate the tumour response during CAB treatment. The “ + ” and “-” signs indicate the tumour human epidermal growth factor receptor 2 (HER2) status. Patients are sorted by progression-free survival after CAB initiation. PD = progressive disease; PR = partial response; SD = stable disease

The median interval between CAB initiation and the post-CAB scans was 21 (range 14–31) days for the PSMA PET/CT scans and 22 (range 13–33) days for the FDG PET/CT scans. A total of 80 lesions were analysed, with a median of 10 lesions per patient (range 4–19). Analysis was limited to bone (n = 30), lymph node (n = 26), and lung lesions (n = 24), as liver and brain metastases could not be delineated due to high physiological tracer uptake in these organs.

At the post-CAB scans, the PSMA SUV_max increased in 20 lesions (25%), remained stable in 46 lesions (58%), and decreased in 14 lesions (18%). Overall, the PSMA SUV_max did not change significantly (median change −0.17; p = 0.74) (Fig. 3; Additional file 1: Table S1). Notably, lung lesions showed a distinct pattern: the PSMA SUV_max increased in most lung lesions (54%), remained stable in 46% of lung lesions, and did not decrease in any of the lung lesions. This was different from the pattern observed in lymph node and bone lesions. The PSMA SUV_max of most lymph node lesions remained stable (58%), while fewer lymph node lesions showed a PSMA SUV_max increase (27%) or decrease (15%). This was similar to the bone lesions, where the PSMA SUV_max remained stable in 67% of lesions, while fewer bone lesions showed a PSMA SUV_max increase (10%) or decrease (23%). In addition to the PSMA SUV_max, there were no significant overall changes in PSMA SUV_mean (median change + 0.05; p = 0.31) and PSMA-derived tumour volume (median change + 0.23; p = 0.07), while total lesion PSMA showed a significant median increase of + 0.58 (p = 0.04) (Additional file 1: Figs. S1-S2).

Fig. 3.

Changes in tumour lesion prostate-specific membrane antigen (PSMA) maximum standardised uptake value (SUV_max) after 21 (± 7) days of combined androgen blockade (CAB). A Paired box plot demonstrating the PSMA SUV_max before and after CAB treatment. B Relative changes per individual patient, sorted by progression-free survival after CAB initiation, and number of lesions with an increased (≥ + 20%), stable (between −20% and + 20%), and decreased (≤ −20%) PSMA SUV_max

The FDG SUV_max increased in 35 lesions (44%), remained stable in 39 lesions (49%), and decreased in 6 lesions (8%), resulting in an overall significant median SUV_max increase of + 0.97 (p < 0.001) (Fig. 4; Additional file 1: Table S1). The increase in FDG SUV_max was most pronounced in lung lesions, with 75% showing an increase and 25% remaining stable, while no lung lesions showed a decrease in FDG SUV_max. For lymph node and bone lesions, the FDG SUV_max most frequently remained stable, while increases were more common than decreases. For lymph node lesions, the FDG SUV_max was stable in 50%, increased in 39%, and decreased in 12%. Among bone lesions, 67% showed stable FDG SUV_max, while 23% showed an increase and 10% a decrease. Consistent with the FDG SUV_max, overall significant increases were also observed for FDG SUV_mean, metabolic tumour volume, and total lesion glycolysis, with median changes of + 0.86, + 0.35, and + 2.70, respectively (all p < 0.001) (Additional file 1: Figs. S3-S4).

Fig. 4.

Changes in tumour lesion fluorodeoxyglucose (FDG) maximum standardised uptake value (SUV_max) after 21 (± 7) days of combined androgen blockade (CAB). A Paired box plot demonstrating the FDG SUV_max before and after CAB treatment. B Relative changes per individual patient, sorted by progression-free survival after CAB initiation, and number of lesions with an increased (≥ + 20%), stable (between −20% and + 20%), and decreased (≤ −20%) FDG SUV_max

The percentage change in PSMA SUV_max and FDG SUV_max showed a significant but weak positive correlation (r = 0.23; p = 0.04). Most lesions with increased PSMA SUV_max also demonstrated increased FDG SUV_max (15/20). Similarly, lesions with stable PSMA SUV_max tended to show stable FDG SUV_max (25/46) (Fig. 5). Figure 6 illustrates a typical case with a vertebral metastasis showing stable uptake of both PSMA and FDG, while a lung metastasis demonstrated increased PSMA and FDG uptake after CAB treatment.

Fig. 5.

Correlation between the relative change in tumour lesion prostate-specific membrane antigen (PSMA) and fluorodeoxyglucose (FDG) maximum standardised uptake value (SUV_max) after 21 (± 7) days of combined androgen blockade

Fig. 6.

Representative PET/CT images of patient no. 3, before and after treatment with combined androgen blockade (CAB). Lesion A is a vertebral metastasis showing stable uptake of both prostate-specific membrane antigen (PSMA) and fluorodeoxyglucose (FDG) after CAB treatment. Lesion B is a lung metastasis showing increased uptake of both PSMA and FDG after CAB treatment. The interval between CAB initiation and scanning was 31 days for the PSMA PET/CT and 33 days for the FDG PET/CT. A PSMA PET, B FDG PET

Discussion

This is the first study to evaluate the effect of CAB on tumour PSMA and FDG uptake in patients with R/M SDC. After a median of 21 days of CAB treatment, we did not observe a significant overall increase in tumour PSMA uptake, whereas tumour FDG uptake increased significantly. Because the PSMA uptake in prostate cancer does increase 1–4 weeks after the initiation of androgen blockade, the absence of a similar effect in SDC raises questions about the underlying mechanisms driving these differences.

We propose three potential explanations for the lack of a PSMA increase in SDC. First, unlike prostate cancer, where PSMA is predominantly expressed on the tumour cells, PSMA in SDC is mainly expressed in the tumour-associated neovasculature [30]. Although the AR is also expressed in cancer endothelial cells [46], it is uncertain whether androgen blockade upregulates FOLH1 transcription in these cells as it does in prostate cancer cells [18–20]. A second explanation could be that the effect of androgen blockade on PSMA uptake may be smaller in hormone-sensitive tumours, such as those in our study, than in castrate-resistant tumours. This difference between hormone-sensitive and castrate-resistant tumours was observed in prostate cancer, where the increased PSMA uptake was larger in the castrate-resistant tumours treated with enzalutamide or abiraterone compared to the hormone-sensitive tumours treated with gonadotropin-releasing hormone analogue ± bicalutamide [23]. Third, considering the aggressive disease course in this cohort, it is plausible that the lack of PSMA response may reflect a shift toward more dedifferentiated tumour phenotypes with reduced PSMA expression and heightened glycolytic metabolism, as seen in prostate cancer [47].

The increased tumour FDG uptake after the initiation of CAB might be caused by tumour progression, a transient treatment-related effect, or both. The cohort’s poor PFS (median 2.2 months) suggests that FDG accumulation due to tumour progression played a role. However, tumours with increased FDG uptake were not only present in patients who experienced rapid tumour progression, but also in patients with stable disease or a partial response at CT evaluation after three months CAB (Fig. 4). Therefore, it can be speculated that CAB may induce an initial rise of FDG uptake in SDC, followed by a decline beyond three weeks.

It was notable that the patterns of PSMA and FDG SUV_max changes in lung lesions were different from lymph node and bone lesions. Most lung lesions (54%) showed an increase in PSMA SUV_max after CAB initiation, a markedly higher proportion compared to lymph node (27%) and bone lesions (10%). Increases in FDG SUV_max were also more common in lung lesions (75%) than in lymph node (39%) and bone lesions (23%). Among 13 lung lesions with increased PSMA SUV_max, 11 lesions were present in patients with a PFS of less than three months (Fig. 3). Similarly, 17 of 18 lung lesions with increased FDG SUV_max were observed in patients with a PFS of less than three months (Fig. 4). This could suggest that tumour progression rather than treatment-related upregulation of PSMA and glucose metabolism contributed to the increased PSMA and FDG SUV_max among lung lesions. However, it cannot be ruled out that biological differences between metastatic sites – such as local microenvironmental factors and vascularisation—contributed to the distinct pattern observed in lung lesions.

Our study has some limitations. First, the small cohort size possibly resulted in overrepresenting patients with a more aggressive disease course. The median PFS and OS of 2.2 and 12.9 months, respectively, are notably shorter than in the phase II study investigating CAB (median PFS 8.8 months; median OS 30.5 months) [7] and a real-world-study on androgen blockade (median PFS 4 months; median OS 17 months) [4]. Potentially, the effect of CAB on PSMA and FDG uptake might differ in patients with more treatment-responsive tumours. However, another possible explanation for the shorter PFS observed in our study could be the PET/CT scans after three weeks of CAB, which possibly led to earlier detection of disease progression. Second, PET/CT scans were performed at only one post-CAB time point after approximately three weeks of treatment. A more comprehensive temporal analysis could give insight into the dynamics of PSMA and FDG uptake after CAB initiation, such as a possible more short-lived PSMA flare [23] or a reduction of PSMA uptake after long-term androgen blockade, as observed in prostate cancer [48–52]. Third, PSMA expression was not assessed on tissue level. Analysing FOLH1 RNA and/or PSMA protein levels in pre-CAB and post-CAB tissue biopsies would provide the most reliable evidence on the effect of androgen blockade on PSMA expression and its underlying mechanisms. This is particularly relevant in salivary gland cancers, where the correlation between PSMA immunohistochemistry and PSMA PET uptake has been shown to be weak [30]. Also, with regard to the effect of androgen blockade on FDG uptake in SDC, serial biopsies would be valuable to distinguish treatment-related effects and tumour progression. Relevant molecular biomarkers include the expression of glucose transporters and markers of metabolic stress and tumour proliferation.

Future prospective studies are warranted to validate and build on the findings of this study. First, these studies should, aim to recruit a more diverse and larger cohort. Second, a more longitudinal design is needed with multiple early and late post-CAB time points, to better capture the kinetics of PSMA and FDG PET/CT scans over time. Finally, future studies should integrate serial tumour biopsies for the analysis of molecular endpoints, such as the expression of FOLH1, PSMA, glucose transporters, and markers of metabolic stress and tumour proliferation.

Conclusions

Treatment with CAB in patients with R/M SDC did not result in a significant increase in tumour PSMA uptake after three weeks. In contrast, tumour FDG uptake increased significantly after three weeks of CAB, which may reflect the poor tumour response in this cohort and/or a transient treatment-related effect.

Abbreviations

AR

Androgen receptor

ALT

Alanine transaminase

AST

Aspartate aminotransferase

CAB

Combined androgen blockade

CT

Computed tomography

FDG

Fluorodeoxyglucose

FOLH1

Folate hydrolase 1

HER2

Human epidermal growth factor receptor 2

MDRD

Modification of diet in renal disease

OS

Overall survival

PET

Positron emission tomography

PFS

Progression-free survival

PSMA

Prostate-specific membrane antigen

RECIST

Response evaluation criteria in solid tumors

R/M

Recurrent and/or metastatic

SDC

Salivary duct carcinoma

SUV_max

Maximum standardized uptake value

SUV_mean

Mean standardized uptake value

Author contributions

All authors were involved in the writing and reviewing of the manuscript. NvR was involved in the data collection, data analysis, and manuscript design. MU was involved in the study concept and design, and data collection. CD and MvR were involved in the data collection. SP, BP, and MG were involved in the study concept and design. GV and AvEvG were involved in the data analysis. JN and CvH were involved in the study concept and design, data collection, and data analysis.

Funding

This study was funded by the Dutch Cancer Society (KWF Kankerbestrijding; grant number 11807). The effect of androgen blockade on [68Ga]Ga-PSMA-11 and [18F]FDGPET uptake in patients with recurrent or metastatic salivary duct carcinoma: a prospective imaging study.

Availability of data and materials

The datasets generated during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

This study was performed in adherence with the 1964 Declaration of Helsinki and its later amendments or comparable ethical standards. This study was approved by the Medical Review Ethics Committee Arnhem-Nijmegen, the Netherlands (NL71139.091.19). Written informed consent was obtained from all the participants before study enrolment.

Consent for publication

Not applicable.

Competing interests

James Nagarajah has received research grants from ABX and Novartis, and has received consulting and teaching fees from ITM, POINT biopharma, CURIUM, Novartis, and Bayer. Carla van Herpen has been member of advisory boards for Bayer, Bristol-Myers Squibb, Elevar, Ipsen, MSD, and Regeneron, and has received research grants from Astra Zeneca, Bristol-Myers Squibb, MSD, Merck, Ipsen, Novartis, and Sanofi. The other authors declare no competing interests.