This study investigated the association between normalized iodine concentration (NIC derived from photon-counting CT (PCCT) and key histopathological features in head and neck squamous cell carcinoma (HNSCC). Our results demonstrate that NIC is a highly reproducible imaging parameter, with excellent interreader agreement. Importantly, NIC was moderately associated with HPV status. Also NIC showed relationships with nodal stage and tumoral cellularity. This finding indicates that NIC may be used as an additional parameter to characterize HSNCC non-invasively.

To date, most of the existing evidence stems from MRI-based studies. Previous studies have shown that MRI-based parameters such as apparent diffusion coefficient (ADC) and dynamic contrast enhanced (DCE) perfusion metrics correlate well with tumor cellularity, proliferation index (KI-67) (van der Hulst et al. 2023; Swartz et al. 2018; Surov et al. 2016, 2017; Driessen et al. 2016; Nakahira et al. 2014; Jansen et al. 2012). Furthermore, some reports indicated that ADC may discriminate HPV positive and negative tumors (van der Hulst et al. 2023; Driessen et al. 2016). As reported previously, HPV-positive tumors consistently exhibited lower ADC values than HPV-negative tumors (van der Hulst et al. 2023; Driessen et al. 2016; Nakahira et al. 2014). So far, several DCE MRI parameters correlated well with microvessel density and the proliferation marker Ki 67 Surov et al. 2017; Jansen et al. 2012). Finally, positron emission tomography (PET)- CT metrics such as standardized uptake value (SUV) and/or metabolic tumor volume have also demonstrated correlations with microvessel density and tumor in HNSCC (Surov et al. 2019).

However, while CT remains the primary imaging modality for head and neck tumor staging (Gage et al. 2017), its ability to reflect underlying histopathology has traditionally been limited. Previously, associations between CT imaging and histopathology in HNSCC were only demonstrated using complex post-processing imaging analyses such as radiomics feature extraction that are currently neither implemented in the clinical routine nor easy to standardize (Zheng et al. 2023). Only recently, with the advent of DECT and PCCT, has it become possible to extract functional and quantitative imaging data such as iodine concentration, that may reflect tumor vascularity of perfusion and are available in every radiological reading (Wu et al. 2023a, b; Fan et al. 2017). Ever since, there is a growing interest to identify CT derived imaging biomarkers and analyzed as a possible imaging biomarker in different tumors.

Our findings are consistent with recent PCCT and DECT studies in other tumor entities. So far, IC was shown to be associated with lymphovascular invasion in rectal cancer (Surov et al. 2024). Furthermore, IC correlated with expression of Ki 67 in lung and rectal cancers (Wu et al. 2023a, b). In rectal cancer, IC correlated well also with hypoxia-inducible factor 1α (HIF-1α) (Fan et al. 2017). In gastric cancer, IC was associated with microvessel density (Chen et al. 2017).

In HNSCC specifically, only two studies have explored similar questions (Wang et al. 2021; Geng et al. 2023). Wang et al. reported a moderate correlation between IC and Ki 67 in laryngeal cancer. Geng et al. demonstrated that arterial NIC values differed by tumor differentiation grade. Our study builds upon these results and is, to our knowledge, the first to link NIC to HPV status and tumor cellularity in a clinically applicable PCCT setting. Furthermore, our study revealed that the associations between NIC and histopathology were different in HPV positive and HPV negative tumors. This finding may be related to the different tumoral architecture in HPV positive and HPV negative lesions.

We observed that HPV-negative tumors had higher NIC-values that HPV-positive tumors. Of dedicated clinical interest is, that a NIC ≥ 0.5 was, despite it‘s limited sensitivity, highly specific for HPV negative tumors. Furthermore, the study results show that more aggressive tumors had lower NIC values. Prior studies have shown that HPV negative tumors express higher levels of pro-angiogenic markers, such as vascular endothelial growth factor (VEGF), and exhibit increased microvascular density (Baruah et al. 2015; Dok et al. 2017). This could lead to greater iodine accumulation and thus higher NIC values in HPV-negative lesions.

Interistingly, NIC correlated inversely with tumor cell count, supporting the hypothesis that hypercellular tumors may exhibit relatively reduced perfusion due to poor vascular supply or central necrosis (Zhu et al. 2014; Miles 1999; Dou et al. 2020). The observed association between lower NIC values and lymph node metastasis further supports the link between perfusion and tumor aggressiveness.

No relevant association was found between NIC and expression of Ki 67 in our study. This may be due to the complex, non-linear relationship between proliferation and perfusion, or to sampling variability in Ki-67. Furthermore, in the present study, NIC did not reflect tumor grade in HNSCC. This finding is in agreement with the literature. Similarly, Geng et al. did not find any relevant association between venous NIC and tumor stage/grade in HNSCC (Geng et al. 2023).

The ability to non-invasively estimate tumor cellularity or predict HPV status using PCCT could have meaningful clinical implications. HPV negative HNSCC is known to be associated with poorer treatment response, especially to radiotherapy, as well as reduced survival (Park et al. 2025). While biopsy and immunohistochemistry remain the gold standard, NIC could serve as a supplementary marker, particularly in cases where histologiccal confirmation is delayed or challenging. The high specificity of NIC (94.4%) for HPV-negative tumors may make it a valuable tool for treatment planning and stratification in clinical trials. Compared to MRI or PET, PCCT offers the advantage of being rapidly available in routine staging workflows, without the need for time-consuming post-processing or contrast protocols beyond standard practice.

This study has several limitations. First, its retrospective nature intrudes inherent selection bias. Second, our results are based on a relatively small sample size. Third, our histopathological analysis focused on a limited number of parameters; more comprehensive immune or vascular profiling could offer deeper mechanistic insights. We also used p16 staining and it is a surrogate marker for the HPV status. Selection bias may also be present in the histopathological analysis. Finally single-slice tumor sampling may not fully capture intratumoral heterogeneity.

Despite these limitations, NIC is a promising imaging biomarker in HNSCC. Future prospective studies should evaluate its role in larger multicenter cohorts and explore its integration with multiparametric imaging approaches (e.g. PCCT + PET + MRI). Moreover, correlation with outcome data could establish NIC as a prognostic marker.

In conclusion, NIC derived from PCCT is a reproducible parameter. It is associated moderately with tumor cellularity, nodal stage as well as HPV status in HNSCC. It may offer added value in non-invasive tumor characterization and risk stratification, especially in cases where histopathological information is limited or unavailable.