Key points are not available for this paper at this time.
Oral squamous cell carcinoma (OSCC) shows substantial immune and clinical heterogeneity that is not fully captured by conventional clinicopathologic risk factors. This systematic review synthesized primary studies evaluating immune biomarker signatures in OSCC identified through spatial transcriptomics, single-cell transcriptomics, and artificial intelligence (AI)-enabled pathology. PubMed/MEDLINE, Scopus, and Embase were searched without language or date restrictions. Eligible studies included original human OSCC investigations reporting immune-relevant biomarker outputs derived from spatial/single-cell transcriptomics or AI-enabled pathology. Nine studies met the inclusion criteria. Six used spatial and/or single-cell transcriptomic approaches, and three used AI-enabled pathology applied to histopathological whole-slide images. Functional meta-synthesis identified four interconnected domains: AI-derived tissue immune infiltration for prognostic stratification; T-cell states and tertiary lymphoid structure-associated antitumor immunity; spatial and metabolic immunosuppressive niches; and stromal–myeloid programs linked to T-cell exhaustion and resistance. Quantitative synthesis was considered but not performed because no group of studies was sufficiently comparable in biomarker construct, comparator, outcome, and effect measure. Clinical confidence remains limited by heterogeneity and prospective validation gaps. These findings suggest that emerging OSCC immune biomarkers may integrate tissue architecture, cellular states, and stromal–immune interactions; however, the current evidence remains exploratory and requires standardized, prospective validation before clinical translation can be considered.
Ardila et al. (Mon,) studied this question.