Fusobacterium nucleatum and IL-32 co-predict nodal metastasis preoperatively in head and neck cancer

Abstract

Background

Head and neck squamous cell carcinoma (HNSCC) is a common malignancy with high rate of lymph node metastasis (LNM), which significantly impacts prognosis. This study aims to evaluate the potential utility of Fusobacterium nucleatum (F. nucleatum) and interleukin-32 (IL-32) as co-predictors for cervical LNM in HNSCC, enhancing preoperative assessment of metastatic status in patients.

Methods

Clinical information, tissue, and plasma samples were collected from a cohort of 254 HNSCC patients. IL-32 expression was assessed using immunohistochemistry (IHC), while fluorescence in situ hybridization (FISH) was used to evaluate F. nucleatum enrichment in both tumor and paracancer tissue. Univariate and multivariate analyses identified significant risk factors.

Results

IL-32 levels were higher in tumor tissue compared to paracancer tissue (P<0.001). Patients with high F. nucleatum enrichment had higher IL-32 expression (P=0.04). Through univariate and multivariate analyses, alcohol, Ki-67 (antigen Kiel-67) expression, F. nucleatum enrichment and IL-32 expression were identified as significant risk factors of LNM. The comparison of receiver operating characteristic (ROC) curves for these factors showed that the combined use of all chosen indicators resulted in the highest diagnostic accuracy.

Conclusions

The enrichment of F. nucleatum in cancer tissue is positively correlated with high expression of IL-32, suggesting that these two novel indicators, F. nucleatum and IL-32, along with Ki-67 and alcohol, may serve as preoperative predictors for LNM in HNSCC. This discovery holds great potential to facilitate the development of more precise and personalized surgical strategies for patients.

Highlight box.

Key findings

• Interleukin-32 (IL-32) expression level elevates in Fusobacterium nucleatum (F. nucleatum) highly enriched tumor tissue.

• F. nucleatum and IL-32 have positive relation with head and neck squamous cell carcinoma (HNSCC) lymph node metastasis (LNM).

• F. nucleatum, IL-32, antigen Kiel-67 (Ki-67) and alcohol can co-predict HNSCC LNM.

What is known and what is new?

• F. nucleatum and IL-32 have been implicated in tumor progression and metastasis in various cancers, but their roles in HNSCC lymphatic metastasis remain underexplored. Common preoperative clinical factors, such as alcohol consumption and Ki-67 positivity, have been associated with metastasis risk in cancer.

• This study identifies a positive correlation between F. nucleatum enrichment and IL-32 expression in HNSCC, suggesting their synergistic role in promoting LNM. It demonstrates that F. nucleatum enrichment, IL-32 expression, alcohol consumption, and Ki-67 positivity are significant predictive markers for LNM in HNSCC. The combined assessment of these factors offers superior diagnostic accuracy for LNM compared to individual markers, contributing to more precise preoperative evaluation and personalized treatment strategies.

What is the implication, and what should change now?

• Incorporating F. nucleatum, IL-32, alcohol consumption, and Ki-67 into preoperative diagnostics can enhance LNM risk prediction in HNSCC, guiding personalized treatment and improving outcomes.

Introduction

Head and neck squamous cell carcinoma (HNSCC) is the six most common malignancies worldwide, characterized by its aggressive nature and high potential for metastasis. Lymph node metastasis (LNM) can affect the treatment and prognosis of patients (1). Clinical diagnosis of lymph nodes is mainly based on the imaging combination of computed tomography (CT) and ultrasound, but its sensitivity and specificity still need to be improved (2). Therefore, identifying additional preoperative biomarkers is crucial for accurately determining the presence of LNM before initiating treatments such as cervical lymph node dissection or neoadjuvant chemoradiotherapy. This approach can help avoid unnecessary treatments or missed LNM diagnoses, both of which may negatively affect patient survival and quality of life.

The metastasis of various tumors, including HNSCC, is increasingly believed to be influenced by specific microbiota within the tumors’ microbial environment (3). Among them, Fusobacterium nucleatum (F. nucleatum) is a gram-negative anaerobe that has been implicated in the progression of various cancers, including head and neck cancers (4). Studies have demonstrated that F. nucleatum can adhere to and invade epithelial cells, disrupt cell-cell junctions, and promote the secretion of inflammatory cytokines, thereby facilitating LNM (5,6). Meanwhile, our previous research confirmed that the abundance of F. nucleatum in HNSCC tissue is higher compared to adjacent non-cancerous tissue, indicating the extent of F. nucleatum infection is closely related to the occurrence and progression of HNSCC (7). Our study also found that F. nucleatum stimulates the growth and metastasis of HNSCC tumors by preventing TGHB2 expression, playing a role in controlling the epigenetics of HNSCC to affect its metastasis (8,9).

Additionally, investigations into various tumor types have revealed the intricate link between cytokines and the processes of carcinogenesis and metastasis. For instance, in malignant cells of oral squamous cell carcinoma (OSCC) lymph nodes compared with primary tumors, there is a significant decrease in CXCL14 expression, which plays a crucial role in suppressing tumor invasion and metastasis (10). Also, through promoting lymphangiogenesis and transendothelial migration of lymphatic circulation OSCC cells, CX3CL1 facilitates cancer dissemination through the lymphatic system (11). In this study, we noticed that interleukin-32 (IL-32) is a pro-inflammatory cytokine involved in various inflammatory and immune responses, promoting the expression of other pro-inflammatory cytokines and chemokines, enhancing the inflammatory milieu that supports tumor growth and dissemination (12). Elevated levels of IL-32 have been observed in several cancer types, including breast cancer, colorectal cancer (CRC), and melanoma, and have been shown to correlate with poor prognosis and increased lymph node metastatic potential (13-15). This suggests that IL-32 may serve as a critical biomarker for predicting disease progression and metastasis in various malignancies (16). Furthermore, F. nucleatum was found to upregulate tumor-derived CCL20 expression, leading to increased infiltration of M2 macrophages and subsequent metastasis in CRC patients (17). These results highlight the critical role of microbe-cytokine interactions within the tumor microenvironment in influencing LNM. Thus, we sought to investigate the impact and predictive value of F. nucleatum infection and IL-32 expression on LNM in HNSCC.

Given the pivotal role of accurately assessing LNM in HNSCC patients, as it significantly influences treatment plans and surgical strategies, this study evaluates the combined predictive value of F. nucleatum and IL-32 for LNM, with the purpose of providing a robust clinical indicator prior to surgery. We present this article in accordance with the TRIPOD reporting checklist (available at https://tcr.amegroups.com/article/view/10.21037/tcr-2024-2539/rc).

Methods

Data collection and ethics

A total of 254 patients treated at the Department of Otorhinolaryngology of Eye & Ear, Nose, Throat (ENT) Hospital, Fudan University, between 2017 and 2022 were enrolled in this study, including 180 cases of laryngeal squamous cell carcinoma (LSCC), 71 cases of hypopharyngeal squamous cell carcinoma (HSCC) and 3 benign masses. None of the patients included in this study had received preoperative radiotherapy nor immunotherapy. All patients signed informed consents prior to sample collection. We completed the collection of patient medical history including LNM as an outcome. The preoperative complete blood count (CBC) data were systematically documented. The tissue was embedded in paraffin to create tissue microarrays (TMA). Plasma samples were stored at −80 ℃ refrigerator which were collected as a copy of preoperative blood test. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Ethics Committee of Eye & ENT Hospital of Fudan University (No. 2022092-2) and individual consent for this retrospective analysis was waived.

TMA construction

Formalin-fixed paraffin-embedded (FFPE) HNSCC donor blocks were sectioned for hematoxylin and eosin (H&E) staining. Tumor regions (≥70% tumor cellularity) and non-cancerous regions (histologically normal tissue within 1 cm of the tumor border, without dysplasia or inflammation) were demarcated on H&E slides by three pathologists blinded to clinical data. One core per predefined region (tumor or non-cancerous) was extracted from each donor block using a TMArrayer (Asiagene, Shanghai, China) and assembled into recipient TMA blocks.

Immunohistochemistry (IHC)

FFPE HNSCC TMA samples were sliced into 5 µm sections. For antigen retrieval, the sections were pretreated in a microwave oven with citrate buffer (pH 6.0) for 12 minutes, and then cooled in deionized water to 25 ℃. The sections were then incubated in methanol containing 3% hydrogen peroxide to inhibit endogenous peroxidase activity. After rinsing with PBS for 5 minutes, the sections were incubated with normal goat serum at 37 ℃ for 1 hour, followed by overnight incubation at 4 ℃ with anti-IL-32 antibodies (Proteintech, Wuhan, China). The sections were then rinsed again with PBS containing 0.1% BSA and incubated with a horseradish peroxidase (HRP)-conjugated secondary antibody. Finally, the sections were stained with 3,3’-diaminobenzidine (DAB) solution and hematoxylin.

Fluorescence in situ hybridization (FISH)

FISH was conducted to detect the enrichment of F. nucleatum in HNSCC tissue samples using a commercially available FISH kit (RiboBio, Guangzhou, China). The universal bacterial probe, targeting a conserved region across various bacterial species, had the probe EUB338 (Sequence 5' to 3': GCTGCCTCCCGTAGGAGT). For specific detection of F. nucleatum, the probe FUS664 (Sequence 5' to 3': CTTGTAGTTCCGCTACCTC) was employed.

Enzyme-linked immunosorbent assay (ELISA)

The level of IL-32 in patient plasma was measured using a commercial ELISA kit (Boster Biological Technology, Wuhan, China). All steps are carried out in accordance with the product manual.

Statistical analysis

IHC- and FISH-stained TMA sections were digitally scanned using a 3DHISTECH Panoramic scanner and analyzed with Caseviewer software. Three non-overlapping regions of interest (ROIs) per core were randomly selected for quantitative analysis via ImageJ (v1.54e): IL-32 expression was measured as average optical density (AOD), while F. nucleatum positivity in FISH was classified into negative (F. nucleatum positive area =0) and positive (F. nucleatum positive area >0) groups based on F. nucleatum probe signal coverage within each ROI. Data analysis was performed using IBM SPSS Statistics (version 29.0), GraphPad Prism (version 10.1) and R (version 4.4). Differences between groups were compared using the Student’s t-test, Student’s paired t-test, or non-parametric test. Categorical variables were analyzed using Fisher’s exact test or Chi-squared test. Multivariate analysis was conducted using logistic regression.

Results

High expression of IL-32 in HNSCC tumor tissue exhibits a positive correlation with the occurrence of tumor LNM

In this study, a total of 254 clinical samples and patient records were collected. Depending on the type of sample collected, these patients were divided into two cohorts after exclusion criteria: cohort 1 (n=159) and cohort 2 (n=83), see Figure 1 for details of the workflow. Since cohort1 data are more comprehensive than cohort2, containing F. nucleatum infections and other pathological information, the analyses and comparisons in this paper were carried out based on cohort1, with the exception of the ELISA experiments which was based on cohort2.

Figure 1.

The flowchart demonstrates the main process and rationale of this retrospective study. ELISA, enzyme-linked immunosorbent assay; F. nucleatum, Fusobacterium nucleatum; FISH, fluorescence in situ hybridization; IHC, immunohistochemistry; IL-32, interleukin-32.

In cohort 1, a comparative analysis of IL-32 expression levels between tumor and paracancer tissue revealed a significantly higher expression of IL-32 in both LSCC and HSCC tissue (P<0.001) (Figure 2A,2B). Furthermore, an elevated expression of IL-32 was consistently observed in the tumor tissue with LNM across both LSCC and HSCC (P=0.002) (Figure 2A,2C). This enhancement in IL-32 expression suggests its potential role in the progression and metastatic behavior of HNSCC.

Figure 2.

The expression of IL-32 in HNSCC tissue and its correlation with LNM in tumors. (A) Representative images of IL-32 IHC staining of cancer and paracancer tissue in LSCC and HSCC with or without LNM. Total magnification for IHC is 200× and 630×. (B) IL-32 expression level between cancer and paracancer tissue of HNSCC. (C) IL-32 expression level between HNSCC with and without LNM. AOD, average optical density; HNSCC, head and neck squamous cell carcinoma; HSCC, hypopharyngeal squamous cell cancer; IHC, immunohistochemistry; IL-32, interleukin-32; LNM, lymph node metastasis; LSCC, laryngeal squamous cell cancer.

The expression level of IL-32 in HNSCC tumor tissue exhibits a positive correlation with the extent of F. nucleatum enrichment

Patients exhibiting high F. nucleatum enrichment in their tumor tissue demonstrated significantly increased IL-32 expression compared to those with lower levels of F. nucleatum enrichment in both LSCC and HSCC (P=0.04) (Figure 3A-3C). Additionally, tumors characterized by high F. nucleatum enrichment were more prone to LNM in LSCC and HSCC (P=0.001) (Figure 3A,3B,3D). The correlation highlights the positive relationship between F. nucleatum enrichment and IL-32 expression, as well as the possible contribution of the microbial composition of the tumor microenvironment to the facilitation of metastatic growth in these malignancies.

Figure 3.

The correlation between the expression of IL-32 and the enrichment of F. nucleatum in tumor tissue. (A) Representative images of FISH staining for F. nucleatum and corresponding IL-32 IHC staining in LSCC tumor tissue. EUB338, AF488-labeled; F. nucleatum, AF594-labeled. (B) Representative images of FISH staining for F. nucleatum and corresponding IL-32 IHC staining in HSCC tumor tissue. Total magnification for IHC and FISH is 630×. (C) IL-32 expression level between F. nucleatum low and high enriched tumor tissue. (D) LNM percentage between low and high F. nucleatum enriched tumor tissue. AOD, average optical density; IHC, immunohistochemistry; FISH, fluorescence in situ hybridization; IL-32, interleukin-32; F. nucleatum, Fusobacterium nucleatum; LSCC, laryngeal squamous cell cancer; HSCC, hypopharyngeal squamous cell cancer; LNM, lymph node metastasis.

The expression of IL-32 exhibits a positive correlation with the expression of Ki-67 in HNSCC tumors

Furthermore, IL-32 expression was higher in the group with high antigen Ki-67 expression compared to the low Ki-67 expression group in both LSCC and HSCC (P=0.04) (Figure 4A,4B). Patients with high Ki-67 expression in tumor tend to develop LNM (P=0.03) (Figure 4A,4C). This observation suggests that high cellular proliferation, indicated by Ki-67 levels, may be associated with enhanced inflammatory cytokine expression and an increased propensity for metastatic dissemination.

Figure 4.

The correlation between IL-32 expression and Ki-67 in HNSCC tumor tissue. (A) Representative images of IL-32 IHC staining of cancer and paracancer tissue in low or high Ki-67% group among LSCC and HSCC with or without LNM. Total magnification for IHC is 200× and 630×. (B) IL-32 expression level between low and high Ki-67%+ tumor in HNSCC. (C) LNM percentage between low and high Ki-67%+ tumor tissue. AOD, average optical density; HNSCC, head and neck squamous cell carcinoma; HSCC, hypopharyngeal squamous cell cancer; IHC, immunohistochemistry; Ki-67, antigen Kiel 67; LSCC, laryngeal squamous cell cancer; LNM, lymph node metastasis.

High plasma concentration of IL-32 is positively associated with the presence of LNM in HNSCC

In cohort 2, plasma levels of IL-32 were observed to be significantly elevated in patients presenting with LNM (P<0.001) (Figure 5A). This finding indicates that IL-32 possesses predictive value in blood, complementing its role discerned from pathology in preoperative biopsies. Consequently, this suggests the potential utility of incorporating IL-32 as a biomarker in the preoperative blood tests to predict LNM. Considering that IL-32 is predominantly secreted by immune cells, we stratified immune cell parameters (absolute counts and relative percentages) into high/low subgroups based on median values to investigate potential differences in IL-32 concentrations between these subgroups (Figure 5B). Furthermore, comparative analysis of immune cell profiles was conducted between patients with and without LNM (Figure 5C,5D). However, all statistical analyses demonstrated no significant differences.

Figure 5.

Analysis of plasma IL-32 levels stratified by lymph node metastasis status and leukocyte level. (A) Plasma IL-32 concentrations in HNSCC patients with and without LNM. (B) IL-32 concentrations in high and low leukocyte subgroups. (C,D) Comparative leukocyte profiles between patients with and without LNM. HNSCC, head and neck squamous cell carcinoma; IL-32, interleukin-32; LNM, lymph node metastasis; ns, not significant.

Univariate and multivariate analyses of LNM in HNSCC

We conducted univariate analysis and multivariate analysis on cohort 1, whose clinicopathological characteristics was shown in Table 1. The univariate analysis identified four significant risk factors for LNM in HNSCC: alcohol consumption (P=0.002), Ki-67 expression (P=0.004), F. nucleatum enrichment (P=0.006), and IL-32 expression (P=0.002) (Table 2).

Table 1. Clinicopathological characteristics of HNSCC patients in cohort 1.

Clinicopathological characteristics	Values
Age (years)	65.09±8.57
<60	39 (24.5)
≥60	120 (75.5)
Sex
Male	153 (96.2)
Female	6 (3.8)
Primary subsites
Larynx	121 (76.1)
Hypopharynx	38 (23.6)
Smoking
No	27 (17.0)
Yes	132 (83.0)
Alcohol
No	74 (46.5)
Yes	85 (53.5)
Diabetes
No	149 (93.7)
Yes	10 (6.3)
Hypertension
No	101 (63.5)
Yes	58 (36.5)
cT staging
1	4 (2.5)
2	42 (26.4)
3	65 (40.9)
4	48 (30.2)
Pathological nodal metastasis
Absent	95 (59.7)
Present	64 (40.3)
Ki-67 (%)
<60	96 (60.4)
≥60	43 (27.0)
Not stained	20 (12.6)
p53
No	55 (34.6)
Yes	66 (41.5)
Not stained	38 (23.9)
p16
No	74 (46.5)
Yes	33 (20.8)
Not stained	52 (32.7)
IL-32 expression	0.06±0.02
F. nucleatum enrichment
Low	40 (25.2)
High	116 (73.0)
Not stained	3 (1.9)

Data are presented as n (%) or mean ± SD. cT staging, clinical T staging; p53, tumor protein p53; F. nucleatum, Fusobacterium nucleatum; HNSCC, head and neck squamous cell carcinoma; Ki-67, antigen Kiel 67; p16 (P16INK4; CDKN2A), cyclin dependent kinase inhibitor 2A; SD, standard deviation.

Table 2. Univariate and multivariate analysis of LNM in HNSCC based on clinicopathological characteristics.

Clinicopathological characteristics	Univariate analysis		Multivariate analysis
	P value (χ2/t)†		OR (95% CI)	P value
Age groups	0.63
Sex	0.23
Primary subsite	0.19
Smoking	0.42
Alcohol, yes vs. no (ref)	0.002*		2.915 (1.349–6.298)	0.006*
Diabetes	0.18
Hypertension	0.65
cT staging	0.49
Ki-67 (%)	0.004*		1.031 (1.006–1.056)	0.01*
p53	0.12
p16	0.11
F. nucleatum enrichment, high vs. low (ref)	0.006*		2.759 (1.077–7.069)	0.03*
IL-32 expression	0.002*		1.331 (1.030–1.720)	0.03*

^†, comparisons performed using Chi-square test for categorical variables and Student’s t-test for continuous variables. All Student’s t-tests and Chi-square tests were two-sided. *, indicates significant P values <0.05. cT staging, clinical T staging, according to AJCC 8th; HNSCC, head and neck squamous cell carcinoma; F. nucleatum, Fusobacterium nucleatum; Ki-67, antigen Kiel 67; LNM, lymph node metastasis; p53, tumor protein p53; p16 (P16INK4; CDKN2A), cyclin dependent kinase inhibitor 2A; CI, confidence interval; OR, odds ratio.

Subsequently, multivariate logistic regression analysis was performed (Table 2). These results indicate that alcohol consumption, high Ki-67 expression, F. nucleatum enrichment, and high IL-32 expression in tumors are significant risk factors associated with LNM in HNSCC.

The diagnostic value of F. nucleatum enrichment and IL-32 in cervical LNM of HNSCC patients

The receiver operating characteristic (ROC) curves for five predictors—alcohol, F. nucleatum enrichment, tissue IL-32 expression, Ki-67 expression, and a combined model—were analyzed to evaluate their diagnostic value for LNM in HNSCC. The area under the curve (AUC) values with their 95% confidence intervals (CIs) are indicated in Figure 6.

Figure 6.

ROC curves plotted according to different indicators, with AUC values shown at bottom-right corner. AUC, area under the curve; F. nucleatum, Fusobacterium nucleatum; FPR, false positive rate; IL-32, interleukin-32; Ki-67, antigen Kiel 67; ROC, receiver operating characteristic; TPR, true positive rate.

The combined model demonstrates the highest AUC of 0.76, which is closest to the top-left corner of the plot, indicating superior diagnostic performance for detecting LNM compared to individual predictors.

Discussion

In this study, we investigate for the first time the potential correlation between IL-32 expression and the occurrence of LNM utilizing cancer and plasma samples from HNSCC patients. Although increasing attention has been directed towards studying IL-32 expression within primary tumor sites and its prognostic implications in cancer (16), the specific relationship between IL-32 levels and LNM diagnosis in HNSCC has not been previously explored. Therefore, this research innovatively examines this under-researched nexus.

The incidence of LNM in HNSCC is notably high. Currently, clinical assessments of cervical lymphatic metastasis predominantly depend on enhanced CT and ultrasound. Nevertheless, the accuracy of these imaging data in predicting LNM is also not entirely guaranteed, highlighting a critical area for improvement in diagnostic accuracy. For instance, metastatic lymph nodes may have a normal form, while reactive or inflammatory lymph nodes may be enlarged (18,19). As such, a subset of patients is very susceptible to clinical lymph node staging errors. Thus, in order to establish the most excised preoperative prediction, future studies should consider integrating imaging indicators with microbial infection and cytokine expression findings from this study to enhance predictive capability.

Current guidelines delineate distinct treatment approaches based on the clinical staging of tumor and lymph nodes, including the necessity for lymph node dissection in patients clinically diagnosed with LNM and chemoradiotherapy. Although well-established, this procedure poses risks such as unnecessary expansion of the surgical scope, impairing local lymphatic drainage and adversely affecting postoperative recovery in patients without metastases (20). Moreover, lymph node staging plays a critical role in determining the necessity for radiotherapy, which is associated with side effects including skin reactions, mucositis, functional alterations, difficulty swallowing, fatigue, and weight loss (21,22). Consequently, accurate classification of LNM is vital prior to initiating specific treatments. For surgical candidates, precise staging helps define the necessary extent of surgery, thereby avoiding both insufficient and excessive interventions. For those requiring radiotherapy, accurate staging aids in determining the appropriate scope of radiotherapy and the regimen of chemotherapy, thus enabling targeted and personalized treatment strategies.

Optimizing the use of preoperatively acquired materials and integrating them with clinical data to elucidate lymphatic metastasis embodies the principles of precision therapy in cancer management. For a definitive diagnosis, patients with HNSCC often undergo biopsy to obtain a sample from the primary tumor foci. Notably, none of the patients included in this study had received preoperative chemoradiotherapy or immunotherapy, ensuring greater homogeneity between the tumor tissues obtained during surgery and those acquired through biopsy. This similarity significantly enhances the reference value of our findings, affirming the reliability of the surgical specimens as representative of the biopsied tissues. In a related effort, a study utilized core needle biopsy (CNB) specimens from 534 patients to develop a predictive model for axillary LNM in breast cancer, tailored to each intrinsic subtype. This model incorporated histological features from the CNB specimens and immunohistochemical markers such as bcl2, CK5/6, and HER2. The implementation of an immunohistochemistry assay for IL-32 in addition to the current panel is straightforward and augments diagnostic capabilities (23).

Plasma, a readily accessible preoperative specimen, is commonly used for predictive biomarker analysis. A study on tongue squamous cell carcinoma (TSCC) patients demonstrated elevated plasma IL-17a levels, detected via ELISA, suggesting its potential as a predictor of cervical lymphatic metastasis (24). Combining our finding that there is a significant correlation between plasma IL-32 level and LNM in HNSCC, supporting the use of plasma cytokines for predicting HNSCC LNM.

Our investigation revealed a positive correlation between IL-32 expression and F. nucleatum enrichment at the primary tumor sites, suggesting a potential synergistic role for IL-32 and F. nucleatum in promoting LNM in HNSCC. As previously documented, interactions between microorganisms and cytokines within the tumor microenvironment significantly influence LNM (8,17). However, the specific mechanical interaction between F. nucleatum and IL-32 remains unexplored. The findings of this study, alongside existing evidence, support the hypothesis of a reciprocal synergistic effect between IL-32 and F. nucleatum on LNM in HNSCC. Additionally, the presence of F. nucleatum can be detected via FISH staining of pathological biopsy specimens, providing further diagnostic insights.

In addition, our study incorporated other common preoperative information and found that alongside F. nucleatum enrichment and IL-32 expression, alcohol consumption and Ki-67 positivity were also significant risk factors for LNM in HNSCC. Combined assessment of these factors demonstrated greater diagnostic value compared to individual markers alone. Although existing studies on HNSCC LNM risk factors have found that metastasis is more common in males, adults aged 31–50 years, higher clinical and pathologic tumor stage, lymphovascular invasion, larger and deeply invasive tumors, keratinizing squamous cell carcinoma (SCC) (25,26), our study did not find a significant association between LNM and sex, age, or T-staging. This may be attributed to the small proportion of female patients and the advanced age of most patients in our sample, which could have impacted the reliability of subgroup analyses. Additionally, several studies have consistently demonstrated that T-staging lacks reliable predictive value for LNM, thereby it remains ongoing debate regarding the efficacy of T-staging as an indicator for cervical LNM (25,27). Pathological data from complete tumor resections, such as tumor invasion depth and size, were not included in our study as our primary objective was to achieve reliable clinical judgment on LNM before surgical treatment. Therefore, all pathological data were derived from biopsy-obtainable information. The current investigation found a correlation between alcohol use and lymph node metastases. Our conclusion is supported by previous research showing F. nucleatum causes reprogramming of ethanol metabolism in HNSCC by inhibiting the expression of transforming growth factor β receptor 2 (TGFBR2) and alcohol dehydrogenase 1B (ADH1B) enabling F. nucleatum to accumulate and promote epithelial-mesenchymal transition, which further exacerbates the uncontrolled progression of LSCC and metastasis (8). Meanwhile, high Ki-67 expression is associated with increased tumor proliferation and aggressiveness, leading to a greater likelihood of metastatic spread to cervical lymph nodes (28).

Owing to constraints in the initial sample collection, it was not feasible to conduct matched plasma IL-32 testing for TMA patients. As a result, a separate cohort was employed, which restricted the comparison study between the expression of IL-32 in tumor tissue and the concentration of IL-32 in plasma. Therefore, it is necessary to increase the sample size further and it should be emphasized that both tissue and plasma samples from patients should be obtained to enable comparison and verification.

Conclusions

This study confirms that elevated IL-32 levels in primary tumor foci and plasma are positively correlated with LNM in HNSCC. We also found that IL-32 levels, F. nucleatum enrichment, Ki-67 positivity, and alcohol intake together serve as predictive markers for LNM in HNSCC patients, aiding in tailored preoperative planning. Future multicenter studies and advanced imaging integration are expected to improve LNM prediction accuracy, allowing for patient-specific surgical strategies in HNSCC.

Supplementary

The article’s supplementary files as

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Ethics Committee of Eye & ENT Hospital of Fudan University (No. 2022092-2) and individual consent for this retrospective analysis was waived.