Head and neck squamous cell carcinoma (HNSCC) is recognized as the eighth most prevalent cancer worldwide, accounting for approximately 771,307 new cases and 384,631 deaths annually.1 The incidence of HNSCC is higher among males, tobacco users, and individuals with a history of alcohol abuse,2,3 as well as among individuals infected with human papillomavirus (HPV).3–5

In clinical practice, therapeutic management of HNSCC is determined by the primary tumor site and staging.6–8 Staging and follow-up of HNSCC are traditionally performed through computed tomography (CT).918F-fluorodeoxyglucose (18F-FDG), a marker of glycolytic activity, has been used in positron emission tomography/computed tomography (18F-FDG PET/CT), which enables comprehensive evaluation of locoregional disease extent and distant metastases.8,10 Surgical resection (SR) and radiotherapy (RT) represent the main approaches for patients with localized disease,11 while patients with metastatic disease require platinum-based chemotherapy,12–14 and multimodal treatment is indicated for patients with locally advanced tumors.7,15

Despite the above-mentioned procedures, the overall survival (OS) rate for patients with HNSCC remains around 40%–50% at five-years,16,15 highlighting the need for improved strategies for tumor characterization, staging, treatment, and treatment monitoring.

Prostate-specific membrane antigen (PSMA), a marker of neoangiogenesis in tissues, was initially recognized for its overexpression in prostate cancer cells.17,18 Gallium-68 PSMA PET/CT scans have been used for detecting metastases of prostate cancer,19 and lutetium-177 PSMA has been indicated for treatment of castration-resistant metastatic prostate cancer.20 Previous studies have demonstrated PSMA expression in non-prostatic malignancies. Incidental detection of lung, lymphoma, thyroid, and base-of-tongue primary malignancies was identified by 68Ga-PSMA PET/CT in patients with prostate cancer.21 Renal cell carcinoma, breast cancer, hepatocellular carcinoma, and glioblastoma were also identified by PSMA PET/CT.22–28 PSMA uptake was observed in adenoid cystic carcinoma, salivary duct carcinoma,29 and concomitant prostate cancer and SCC of the oropharynx30 and larynx31 cases, but the clinical role of PSMA PET/CT in HNSCC remains undefined. In this context, the current study aimed to investigate the role of PSMA PET/CT in the characterization of HNSCC, considering 18F-FDG PET/CT as the reference method among imaging modalities.

We hypothesized that, in HNSCC, 18F-PSMA-1007 PET/CT may provide biologically distinct information from 18F-FDG PET/CT by preferentially reflecting tumor-associated neoangiogenesis, whereas FDG predominantly reflects glycolytic metabolism within the tumor microenvironment. Therefore, both tracers may have complementary clinical value for lesion characterization and for depicting intratumoral and interlesional heterogeneity.

The demographic and clinicopathological characteristics of the 14 patients included in the study are presented in Table 1. The median age was 61-years, with a predominance of male, smokers, and alcoholic patients. Most patients had good functional performance, one or more comorbidities, moderately differentiated tumors, and predominantly AJCC 8th edition stage III–IVB disease (M0) at presentation; one patient had a prior stage II primary but was imaged during locoregionally advanced recurrence. The tumors were located predominately in the oropharynx, larynx or oral cavity. Nine patients were analyzed at diagnosis and five patients at disease recurrence.

The numbers of primary and metastatic lesions detected by FDG PET/CT and PSMA PET/CT are presented in Table 2 and Supplementary Table S1. Excluding patient 11, who had undergone total laryngectomy and therefore lacked a primary anatomical site for evaluation, FDG PET/CT detected primary lesions in all 13 evaluable patients (100.0%), whereas PSMA PET/CT demonstrated uptake in 10 of the 13 patients (76.9%). All primary lesions detected by PSMA PET/CT were also detected by FDG PET/CT (Fig. 1). Excluding patient 6, who presented only locally advanced disease, FDG PET/CT identified metastases in cervical lymph nodes, pulmonary lymph nodes, pulmonary parenchyma, and bone in 13 patients (100.0%), whereas 6 out of 13 patients (46.2%) presented metastases in cervical lymph nodes, pulmonary lymph nodes, and pulmonary parenchyma detected by PSMA PET/CT (Fig. 2). All metastatic lesions detected by PSMA PET/CT were also detected by FDG PET/CT, except an unbiopsied sclerotic lesion in the sacral wing in patient 12, which was seen only by PSMA PET/CT.

Fig. 1.

PET/CT images with 18F-PSMA-1007 (top row, A–E) and 18F-FDG (bottom row, F–J) in a patient with squamous cell carcinoma involving the right oral cavity and oropharynx, associated with left-sided cervical lymphadenopathy (Patient 9). Focal radiotracer uptake is seen in the primary lesion site (solid arrows) and in a left cervical nodal conglomerate with central necrosis, centered in the parotid chain (dashed arrows). Axial CT (B and G), PET (C and H), fused PET/CT (D and I), and sagittal reconstructions (E and J) confirm the anatomical and functional findings. 18F-FDG PET/CT uptake was markedly higher in both the primary tumor (SUVmax1 h: 59.8 vs. 6.2; SUVmax2 h: 63.7 vs. 6.9) and the metastatic lymph node (SUVmax1 h: 15.8 vs. 3.9; SUVmax2 h: 22.0 vs. 4.2). Small mediastinal lymph nodes with FDG uptake (F), but no PSMA uptake (A), were observed, likely reflecting inflammatory changes rather than metastasis (a potential false-positive FDG finding).

Fig. 2.

Maximum Intensity Projection (MIP) of 18F-PSMA-1007 (A) and 18F-FDG (B) PET/CT images of a patient with right-sided oropharyngeal squamous cell carcinoma and mediastinal and pulmonary metastases (Patient 13). Axial slices of fused 18F-PSMA PET/CT images of the head and chest (soft tissue and pulmonary windows) are shown in (C), and the corresponding slices of 18F-FDG PET/CT are shown in (D). Lesion uptake of both tracers is seen in the primary site, located in the right pharyngeal tonsil (solid arrows), with more intense and extensive for 18F-FDG (SUVmax1 h: 10.7, SUVmax2 h: 12.9) than for 18F-PSMA-1007 (SUVmax90 min: 3.0, SUVmax2 h: 4.4). Additionally, multiple metastatic mediastinal lymph nodes, including upper and lower paratracheal, subcarinal, and right hilar chains are seen with both tracers (dashed arrows), with markedly higher FDG uptake (SUVmax2 h up to 34.4) than PSMA uptake (SUVmax2 h up to 6.9). A spiculated pulmonary mass is also identified in the posterior basal segment of the right lower lobe, with intense FDG uptake (SUVmax2 h: 25.9) and moderate PSMA uptake (SUVmax2 h: 15.7), consistent with pulmonary metastasis (arrowheads).

Ten out of 13 patients with primary tumors (76.9%), 6 out of 13 patients with metastatic disease (46.2%) were identified by both FDG PET/CT and PSMA PET/CT (Fig. 1–2). In patient 1 and patient 3, both radiotracers demonstrated concordant uptake in the same indeterminate lesions; in patient 6, FDG showed uptake in two lesions, while PSMA identified only one of them. Interestingly, patients with largest tumors presented intralesional patterns with marked differences (Fig. 3).

Fig. 3.

Coronal (A) and axial (B) slices showing PET (right) and fused PET/CT images (left) obtained with 18F-PSMA-1007 from a patient with squamous cell carcinoma of the right retromolar trigone, with bone invasion (Patient 6). The corresponding 18F-FDG images are shown in (C) and (D). Lesion uptake of both tracers is clearly seen in the ill-defined primary lesion (arrows). However, their intralesional patterns differ markedly: 18F-PSMA-1007 displays peripheral uptake (A and B), likely reflecting PSMA expression in the tumor's neovascular periphery, while 18F-FDG shows predominant central uptake (C and D), consistent with increased glycolytic activity in a hypoxic or partially necrotic tumor core.

In the analysis of FDG-positive/PSMA-negative findings, PSMA PET/CT showed no visually relevant uptake at the primary site in three cases, whereas FDG PET/CT identified the primary lesion in these patients. These PSMA-negative primaries involved the oropharynx (patient 2), oral cavity (patient 10), and larynx (patient 14). At the metastatic level, several FDG-avid deposits, predominantly cervical nodal lesions across levels II–V/VI, showed no corresponding PSMA uptake, underscoring lesion-dependent heterogeneity of PSMA targeting in this cohort. This description is based on lesion-level mapping (Supplementary Table S1) and was not subjected to formal statistical testing.

The maximum FDG and PSMA uptake values at 1h and 2h post-injection are presented in Table 3. The mean SUVmax values in primary lesions detected by FDG PET/CT at 1 and 2h were higher than those detected by PSMA PET/CT. For metastatic lesions, the mean SUVmax values detected by FDG PET/CT at 1 and 2h tended to be higher than those detected by PSMA PET/CT.

The current study provides a head-to-head comparison of 18F-FDG and 18F-PSMA-1007 PET/CT in locoregionally advanced HNSCC, demonstrating that while FDG remains superior for tumor detection, PSMA offers complementary insights into tumor biology, particularly in characterizing angiogenesis and intratumoral heterogeneity.

It was observed that HNSCC lesions were more consistently identified by 18F-FDG PET/CT than by 18F-PSMA-1007 PET/CT, particularly for metastatic lesions, and 18F-FDG PET/CT outperformed 18F-PSMA-1007 PET/CT in sensitivity (100% vs. 76.9% for primary lesions; 100% vs. 46.2% for metastases).

Although PSMA uptake is biologically linked to tumor-associated neovasculature, the PSMA PET/CT detection rate observed in our cohort was lower than might be expected from the reported frequency of vascular PSMA expression in HNSCC histopathological series. In our study, 18F-PSMA-1007 demonstrated visually relevant uptake in the primary tumor in 10/13 evaluable patients (76.9%) and detected metastatic disease in 6/13 patients (46.2%), whereas FDG detected all evaluable primary and metastatic lesions. Tissue-based data have shown frequent, yet heterogeneous, endothelial PSMA expression in HNSCC, with a graded distribution of microvessel involvement and higher vascular PSMA expression in metastatic lymph nodes than in reactive nodes.34,35 Therefore, histologic vascular PSMA positivity does not necessarily translate into lesion detectability on PET, consistent with lesion-dependent heterogeneity of endothelial expression.

In fact, FDG PET/CT has assumed an important role in the initial assessment36 and early detection of recurrences in HNSCC.37,38 More recently, the diagnostic potential of PET/CT using radiotracers targeting PSMA, already established in prostate cancer, has been investigated in other solid tumors.22–29

In HNSCC, synchronous SCC (squamous cell carcinoma) of the oropharynx was identified in a retrospective study involving 764 patients with prostate cancer,21 an invasive SCC of the oropharynx was detected during prostate cancer staging,30 and, more recently, the use of a dual-tracer PET/CT approach with 18F-FDG and 18F-PSMA-1007 demonstrated FDG and PSMA uptake in both prostate and laryngeal carcinoma.31 Considering that PSMA uptake has been previously reported only in incidental cases of HNSCC, this is the first study that analyzed PET/CT using radiotracers targeting PSMA in a series of HNSCC patients. In addition, as 18F-FDG PET/CT enables the detection of glycolytic activity across the tumor microenvironment,8,10 and PSMA uptake reflects tumor neo vasculature,34,35 the results of the current study present, for the first time, preliminary evidence that metabolic reprogramming and neoangiogenesis are hallmarks of HNSCC.

It is worth commenting that one false positive finding was attributed to PSMA PET/CT. In patient 12, an increased PSMA uptake was observed in a unbiopsied sclerotic bone lesion located in the right sacral wing, without corresponding metabolic activity on FDG PET/CT or anatomical findings on CT. A high prevalence of nonspecific bone uptake on PSMA-1007 PET in patients with prostate cancer has been reported,39 and this finding was validated with follow-up PSMA PET/CT.40 Such nonspecific uptake may be attributed to PSMA expression in the neovasculature associated with inflammatory conditions and changes in vascular permeability.41

In the current study, it was observed that the mean SUVmax values in primary lesions detected by FDG PET/CT at 1 and 2h were higher than those detected by PSMA PET/CT, with an increase in both tracers’ uptake over time. In a review of PSMA-targeted radiopharmaceuticals in non-prostatic tumors, it was reported that SUVmax values observed on PSMA PET/CT were generally lower than those seen in prostate cancer, and that tracer distribution tended to be more heterogeneous.42 Although HNSCC was not included in that analysis, it was emphasized that in most of the evaluated tumors, PSMA expression was confined to the tumor neovasculature.43,44 In another report, when synchronous prostate and laryngeal carcinomas were evaluated by both tracers, the laryngeal carcinoma exhibited intense FDG uptake with only mild PSMA avidity, while the prostate carcinoma demonstrated the opposite pattern.31 The results of the current study support the concept that 18F-FDG PET/CT demonstrates more intense uptake and more diffuse distribution in HNSCC when compared to PSMA PET/CT.

Interestingly, it was also observed in the current study that patients with large lesions presented markedly different intralesional patterns, as in patient 6. FDG showed central uptake in the primary lesion, aligning with its affinity for hypoxic, glycolytically active tumor cores.45 PSMA exhibited peripheral uptake, likely due to PSMA expression in tumor-associated neovasculature.34 The central necrosis or hypoxia may limit PSMA expression to viable peripheral regions. This spatial dichotomy underscores the complementary roles of FDG (metabolic activity) and PSMA (angiogenesis) in HNSCC characterization. Similar findings have been reported in prostate cancer43 and glioblastoma,26 where PSMA uptake correlates with tumor microvessel density rather than epithelial proliferation. The distinct uptake patterns observed may have therapeutic implications: FDG’s central dominance supports its role in staging and surveillance, particularly for identifying aggressive, hypoxic tumor subclones resistant to radiotherapy.46 PSMA’s peripheral localization suggests potential utility for guiding anti-angiogenic therapies (e.g., VEGF inhibitors) in PSMA-avid subgroups and for selecting patients for PSMA-targeted radioligand therapy (e.g., 177Lu-PSMA), as proposed for other solid tumors.47

The current study has some limitations. The small sample size, both at initial diagnosis and at recurrence, and the single-center design may limit the statistical power and the generalizability of the findings. Another point is that the proliferative index Ki-67 as well as molecular characterization were not available for these patients, precluding correlation with the metabolic and/or angiogenic activity observed on imaging studies.

In conclusion, our data present preliminary evidence that 18F-FDG PET/CT is indispensable for HNSCC staging while 18F-PSMA-1007 provides complementary angiogenic profiling, contributing to tumor characterization. Further larger studies with patients at diagnosis and relapse from various services and with well tumor characterization have to be conducted to define consistent conclusions about the roles of PET/CT with 18F-FDG and 18F-PSMA-1007 in staging of HNSCC and in the selection of candidates for targeted therapies.

Hádila da Silva Veras Sousa: 0009-0000-8400-9456

Victor Cabral Costa Ribeiro Heringer: 0000-0002-1649-5656

Matheus Yung Perin: 0000-0001-6981-9063

Nájua Abou Arabi Silveira: 0009-0003-5772-4288

Simone Kuba: 0009-0009-2414-7705

Ligia Traldi Macedo: 0000-0003-0575-586X

Gustavo Jacob Lourenço: 0000-0002-5944-0305

Carlos Takahiro Chone: 0000-0002-4217-4629

Carmino Antônio de Souza: 0000-0001-8656-8374

Celso Dario Ramos: 0000-0002-1312-1026

Carmen Silvia Passos Lima: 0000-0002-1314-2345

The number of authors exceeds the journal’s limit of seven; however, we request the inclusion of all eleven authors due to their substantial and distinct contributions to the study. This work was developed by a multidisciplinary team that included oncologists, radiotherapists, and surgeons, whose expertise was essential for patient selection, data acquisition, and interpretation.

HSVS contributed to patient selection, data analysis, manuscript drafting, and critical revision. MYP and LTM contributed to patient selection. VCCRH, and SC were responsible for performing imaging examinations and conducting image analysis. GJL performed the statistical analyses. CTC, CAS, CDR and CSPL contributed to research planning, study design, manuscript drafting, and critical revision. All authors have read and approved the final version of the manuscript and agree to be accountable for all aspects of the work.

We are grateful for financial support from Fundação de Amparo a Pesquisa do Estado de São Paulo, Brazil, [Grant FAPESP #2021/10265-8], Cancer Theranostics Innovation Center/Centro de Pesquisa, Inovação e Difusão, CancerThera/CEPID; and National Council for Scientific and Technological Development, CNPq, Research Productivity, [Grant #307944/2022-0(Lourenço GJ)], [Grant #312212/2025-9(Ramos CD)], and [Grant #302922/2025-3 (Lima CSP)].

The authors declare that all data are available in repository.