Stromal inflammatory subtypes of oral squamous cell carcinoma correlate with patient clinical characteristics, demographics and gene expression

Abstract

<h3>Introduction</h3> Oral squamous cell carcinoma (OSCC) is the most common oral malignancy displaying significant mortality and morbidity. The current OSCC grading systems exclusively describe the tumor differentiation, and share limited morphologic characterization of the stroma. Herein, we present a comprehensive microscopic analysis of the OSCC stromal inflammation and its relation with clinical features, demographics and gene expression of these patients. <b>Materials and Methods:</b> 87 completely excised OSCC tissues were prospectively collected. OSCC microenvironment was characterized according to the immune phenotype, the degree of desmoplasia and the extent and localization of inflammation. Statistical correlations between the tumor stroma and the patients' characteristics were carried out using GraphPad Prism. KEGG pathway analysis of representative PanCancer IO 360^TMdata was also investigated using nSolver software. <h3>Results</h3> Inflamed tumors were more commonly encountered (49 cases) compared with non-inflamed OSCC (31 immune-excluded and 7 immune-deserted). Peritumoral fibrosis (36 tumors) correlated with male gender (p=0.0043), smoking (p=0.0455), alcohol consumption (p=0.0044), increased tumor size (p=0.0054) and stage (p=0.002) while a peritumoral inflammation (60 cases) was mainly seen in females (p=0.0105), non-drinkers (p=0.0329), and was associated with decreased size (p=0.0044). No association between stromal characteristics and tumor grade was seen. Pathway analysis showed differential gene expression between triplicates of inflamed and immune-excluded tumors in cancer pathways and chemokine or cytokine signaling. <h3>Conclusions</h3> Our work describes the relation of carcinogens (alcohol and tobacco) with distinct stromal phenotypes in oral cancer. Additionally, the correlation with tumor size and stage complements previous work indicating a worst biologic behavior in "cold" tumors. In addition, the underlying molecular pathways that can be implicated in different stromal phenotypes are highlighted. Collectively, these findings in addition to the indications that the microenvironment shares microscopic features independent of the tumor grade highlight the necessity of their incorporation into future grading systems of OSCC.