Histopathologic Features of Mucosal Head and Neck Cancer Cachexia

Objective Determine the histopathologic features that correlate with head and neck cancer (HNC) cachexia. Methods A single-institution, retrospective study was performed on adults with HPV-negative, mucosal squamous cell carcinoma of the aerodigestive tract undergoing resection and free flap reconstruction from 2014 to 2019. Patients with distant metastases were excluded. Demographics, comorbidities, preoperative nutrition, and surgical pathology reports were collected. Comparisons of histopathologic features and cachexia severity were made. Results The study included 222 predominantly male (64.9%) patients aged 61.3 ± 11.8 years. Cachexia was identified in 57.2% patients, and 18.5% were severe (≥15% weight loss). No differences in demographics were identified between the groups. Compared to control, patients with severe cachexia had lower serum hemoglobin (p=0.048) and albumin (p < 0.001), larger tumor diameter (p < 0.001), greater depth of invasion (p < 0.001), and elevated proportions of pT4 disease (p < 0.001), pN2-N3 disease (p=0.001), lymphovascular invasion (p=0.009), and extranodal extension (p=0.014). Multivariate logistic regression identified tumor size (OR [95% CI] = 1.36 [1.08–1.73]), oral cavity tumor (OR [95% CI] = 0.30 [0.11–0.84]), and nodal burden (OR [95% CI] = 1.16 [0.98–1.38]) as significant histopathologic contributors of cancer cachexia. Conclusions Larger, more invasive tumors with nodal metastases and aggressive histologic features are associated with greater cachexia severity in mucosal HNC.


Introduction
Head and neck cancer (HNC) encompasses a diverse group of malignancies and accounted for 54,010 new cases and 10,850 cancer deaths in 2021 [1].Squamous cell carcinoma (SCC) is the predominant histology usually resulting from signifcant exposure to tobacco and/or alcohol.Terefore, most tumors occupy the mucosal surface of the upper aerodigestive tract (oral cavity, oropharynx, hypopharynx, and larynx) and cause signifcant morbidity due to impaired mastication, deglutition, and/or respiration [2].
Advanced HNC is associated with high proportions of unintentional weight loss often despite nutritional supplementation resulting in a catabolic state known as cancer cachexia [2][3][4][5].Tis is a detrimental condition caused by the interaction between the immune system and malignancy [6].Tis interplay releases infammatory and catabolic mediators which promote skeletal muscle and adipose wasting, fatigue, anorexia, and metabolic derangements [6].Ultimately, there is an imbalance of energy and unintentional wasting of body mass.As such, it is diagnosed primarily by unintentional weight loss with or without being underweight and sarcopenic [7,8].Cancer cachexia unfortunately thus far has limited, if any, improvement with conventional nutrition methods [9,10].Tis comorbidity results in poorer tolerance and outcomes with surgery, chemotherapy, radiation therapy, and survival [11][12][13][14][15][16][17].
HNC induces the third highest incidence of cancer cachexia behind pancreatic and gastroesophageal malignancies [6].It occurs most commonly in advanced tumors of the pharynx and supraglottis, which also have the poorest overall fve-year survival within the head and the neck [4].As an estimated 30% of cancer deaths are related to the consequences of cachexia (e.g., cardiopulmonary decline), a substantial portion of advanced HNC patients will also succumb to cachexia's efect [18].Despite the preponderance of cachexia in these patients, there is a paucity of investigations into HNC cachexia and limited in vivo or in vitro models [19,20].Terefore, we sought to identify histopathologic features of mucosal HNC which promote cancer cachexia that may provide insight into its pathoetiology.Given the increased incidence of cachexia in more advanced cancers and greater tumor burden [4,[21][22][23], we hypothesized that larger tumors with more aggressive features would be correlated with worse cachexia severity.

Patients and Methods
After exemption for review from our Institutional Review Board, a retrospective review was performed of consecutive, adult patients with head and neck mucosal SCC undergoing oncologic ablation and reconstruction at our academic, tertiary referral center from 2014 to 2019.Patients were excluded for any of the following: (1) no pathological report; (2) no primary tumor (Tx/T0 disease); (3) HPV/p16-positive disease; (4) concurrent primary malignancy or distant metastases; (5) lack of or inadequate 30-day preoperative abdominal CT images; (6) active immunosuppression; (7) obvious source of malnutrition (e.g., prior small bowel resection).
Statistical analyses were performed using SPSS v28.0 (IBM, Armonk, NY).Continuous and ordinal data reported as mean ± standard deviation or median [range], respectively.Categorical data were presented as number (%).Normality was determined using the Shapiro-Wilk test.Univariate analyses were performed using Welch's t-test, Mann-Whitney U-test, or Fisher's exact test.Comparisons of multiple groups after stratifying cachexia severity were performed with 2-sided Welch's one-way ANOVA with Tukey post-hoc analysis, Kruskal-Wallis one-way ANOVA with post-hoc analysis, or Pearson's χ 2 test.Signifcance was determined at p < 0.05.With two group comparisons, efect sizes and 95% confdence intervals (95%CI) were measured using Cohen's d for normally distributed continuous data, Goodman and Kruskal's c for ordinal data, and Cramer's v for categorical data.For three or more group comparisons, efect sizes with 95% CI were measured with ω 2 for continuous variables (fxed-efect model) or Cramer's v for categorical variables.All comparisons and efect sizes were determined with bootstrapping using 1000 samples.Binary logistic regression was performed on histopathologic features on the presence of cancer cachexia.Variables with p < 0.10 on univariate regression were included in the initial multivariate regression and then subsequently removed using the backward Wald method.Tose variables with p < 0.10 were included in the fnal model.

Discussion
Tis investigation reports the frst review of histopathologic features of nonmetastatic mucosal head and neck SCC which are correlated with cancer cachexia and its severity.We demonstrate the efect of tumor size, invasion, nodal disease, grade, and adverse features on cachexia severity and their associations with the clinical criteria for cachexia diagnosis.Tis lays important groundwork into undermining the pathoetiology of cancer cachexia for this population of patients in whom cachexia is highly prevalent and detrimental to prognosis.Te underlying mechanism of cancer cachexia is complex but originates in the tumor-host interaction.Tere is frst the tumor "secretome" -procatabolic molecules (e.g., TGF-β, heat shock proteins) produced from the tumor itself.Second to arise are the infammatory cytokines and mediators (e.g., IL-1, IFN-c, and TNF-α) which are induced by the tumor and host immune system interaction.Together, a widespread infammatory state is generated and incites skeletal muscle degradation, lipolysis and energy wasting, and central nervous system changes which promote anorexia and fatigue.Te target organs of these mediators demonstrate reciprocal interaction (i.e., "crosstalk") to perpetuate the cachectic response even further [6].Emerging evidence suggests that there are secondary and tertiary efects of these interactions including bone degradation and altered gut biome [29,30].Furthermore, the chemotherapeutic agents used to eradicate the cachexia-inducing tumor can result in similar alterations in homeostasis and propagate its progression [31].Investigating the tumor-host and multiorgan interactions therefore is imperative for development of targeted cachexia therapeutics.
Tumor size and DOI were strongly associated with cachexia severity in our investigation.Although it is known that larger tumors have greater incidence of cancer cachexia within head and neck cancer [4] as well as several other cancer pathologies [21,22], the relationship between DOI and cachexia has not previously been examined.DOI is an important histopathological feature that determines tumor classifcation and overall prognostic staging for carcinoma of the oral cavity [26,32,33].Tis is similar to cutaneous 4 International Journal of Surgical Oncology International Journal of Surgical Oncology carcinomas and melanomas where depth of invasion portends worse prognosis than tumor size [26].In the hypopharynx and larynx, although not included in TNM staging, DOI also predicts nodal disease and disease-free survival [34][35][36].
Tough there is concordance between tumor size and DOI, it is not perfect as tumors can be large and exophytic or small and endophytic and therefore represent separate entities.In a similar manner, higher tumor burden within local cervical node metastases was identifed at the greatest degree of cachexia severity, both number of involved regional nodes, and identifcation of ENE.Together these correlated to greater pathologic nodal classifcation in severely cachectic patients.As expected, nodal disease is more frequently identifed in larger, greater T-classifcation primary tumors.With both tumor and nodal classifcations combined, the incidence of cancer cachexia increased as staging increased, similar to prior analyses [23].It is unclear if nodal metastasis in head and neck SCC has a unique role in the cachectic process compared to the primary tumor or is simply an increase in total disease burden.However, evidence suggests that the development of cachexia is linked to the molecular events during progression of low-stage to high-stage disease, advancing from localized tumor to nodal and fnally distant metastases [37,38].Te genetic changes and resulting molecular expression which allow for the malignant propagation also have corresponding local and distant tissue efects.
Given the importance of tumor-host interplay in cancer cachexia, we speculate increased tumor size, tumor DOI, and greater nodal burden all increase the total surface area for such interactions.Tis may, in turn, enable more opportunities for the cachectic infammatory response to occur and subsequent systemic cachectic efects.Greater tumor burden in both primary and nodal disease also increases the total tumor cachexia "secretome" and production of tumorderived cachexia mediators.Te signaling and molecular changes that occur with each stage of tumor progression may also coincide with these changes in histopathology, but further investigations are required to confrm or deny these suspicions.
What cannot be explained precisely are the diferences between patients with large, invasive and/or regionallymetastatic SCC without any evidence of cachexia versus those with small, localized tumors experiencing signifcant weight loss and atrophy.Tis may be due to polymorphisms in cachexia-modulating cytokines and receptors produced by both the tumor and the host.Te unique tumor molecular alterations combined with unique patient resistance or susceptibilities can increase or decrease the rate at which cachexia develops, respectively [21].For instance, hundreds of single nucleotide polymorphisms have been identifed in the expression of infammatory cytokines, proteins, and hormones in several diferent cancers inducing cachexia  6 International Journal of Surgical Oncology [21,39].Similar variations in host cytokine receptors can also account for the diferences seen in tumor response.In congruence to individual genetic expression diferences is the variation in autophagy of the tumor microenvironment which modulates the tumor-immune infammatory response.Furthermore, the "tumor purity," proportion of tumor to immune cell infltrate, fuctuates by cancer histology, and the corresponding host infltrating immune phenotype has contributed to the heterogeneous cachexia response [40].All these variations may be akin to the variable infammatory response and symptoms produced by an illness or efectiveness of a therapeutic drug from one individual to the next.Further analyses are required to determine those factors which promote cachexia progression, potentially through single-cell DNA testing of tumors, host immune cells, skeletal muscle, adipose, etc. LVI and PNI, although adverse features which have demonstrated worse survival and locoregional control, do not have a role in TNM staging in mucosal HNSCC [26,41,42].Despite this, the presence of either on histopathology generally necessitates adjuvant radio-and/or chemotherapy after surgical resection [41,42].PNI occurs when tumor cells are identifed within any of the three peripheral nerve sheath layers and allows microscopic regional spread.LVI is identifed when tumor aggregates within the walls or endothelium of lymphatic vessels and, similar to PNI, provides a highway for metastasis [42,43].Both features had similar prevalence between cachectic patients and controls, but there was a signifcant increase in LVI prevalence when stratifying by cachexia severity (>70%).Tis may represent only a correlative incidence with the larger, more invasive tumor and be part of the bigger picture of advanced disease, particularly nodal disease [44].However, similar mechanisms which contribute to tumor invasion in the epithelial-mesenchymal transition, as seen with DOI and tumor size, are also identifed in PNI and LVI.Tese processes require reciprocal interaction of Schwann cells and lymphatic endothelial cells, which can induce infammatory cytokine release [41,44].Although PNI can induce signifcant neuropathic pain, limit food intake, and promote malnutrition particularly in oral cavity SCC [41], it was not correlated with greater weight loss in this investigation.A lack of correlation in cancer cachexia and PNI in pancreatic adenocarcinoma has similarly been demonstrated [45].
Tumor grade, a classifcation of cell atypia, was not associated with cancer cachexia within this study, and it, too, is not included in the AJCC staging of head and neck SCC [26].Yet, tumor grade has signifcant prognosis in locoregional and overall survival in these patients [46].Furthermore, higher levels of atypia in mucosal SCC are correlated with other adverse features including nodal disease, ENE, LVI, and need for postoperative chemotherapy and radiation therapy [46].Tumor grade remains a critical component of staging in several other head and neck malignancies, such as soft tissue sarcomas and cutaneous carcinomas, and may contribute to the development of cachexia in certain cancers [26].However, the molecular alterations which occur at varying levels of tumor grade and the resultant tumor-immune interactions and release infammatory mediators of cachexia are still unknown.
Tis investigation is limited in its retrospective nature which only included aerodigestive, non-HPV SCC of the head and neck.Although these are the most common head and neck malignancies, it excludes other pathologies such as salivary gland, cutaneous SCC, and sarcomas.Alongside this, there were no serologic biomarkers of infammation (IL-2, CRP, etc.) with which further correlations could be made.With this and the high rates of dysphagia and prior chemoradiation therapy, we are unable to fully parse out the malnutrition component of the weight loss identifed in our HNC patients, though the majority patients with dysphagia and aspiration were initiated on tube-feeding regimens many weeks prior to surgery.Finally, the cohort size is limited due to the requirement of abdominal CT imaging for the measurement of SMI, which are generally obtained only for metastatic workup and/or vascular imaging for fbular free fap reconstruction.

Conclusions
Mucosal HNC cachexia is associated with larger, more invasive tumors with more adverse features.Te molecular role these features have in the development of cancer cachexia requires further investigation.

Table 1 :
Overall summary and comparison of cachectic versus control patients.

Table 2
< 0.001), but nodal burden had no signifcant correlation to Hgb or albumin.Expectedly, both tumor size and DOI were well correlated to nodal burden and size.3.5.Logistic Regression of Histopathologic Features on CancerCachexia.Univariate and multivariate regressions of the histopathologic features upon cancer cachexia status are displayed in Table4.Variables with p < 0.10 on univariate regression included tumor size, tumor DOI, tumor located other than oral cavity, nodal burden, largest node, and PNI.

Table 2 :
Comparison of patient and histopathologic features by cancer cachexia severity.

Table 4 :
Binary logistic regression of histopathologic variables on cancer cachexia.
Variables with p < 0.10 on univariate regression were included in the initial multivariate regression model (tumor size, tumor DOI, oral cavity, nodal burden, total nodes, PNI).Variables were removed sequentially in the backward Wald method with all fnal variables having p < 0.10.CI, confdence interval; CRT, chemoradiation therapy; DOI, depth of invasion; ENE, extranodal extension; LVI, lymphovascular invasion; OR, odds ratio; PNI, perineural invasion.