Spatial neuroimmune crosstalk driving perineural invasion in head and neck squamous cell carcinoma

Perineural invasion (PNI)-the infiltration of tumor cells within, around, or through nerve fibers-is a hallmark of aggressive tumor biology and an independent predictor of poor prognosis in head and neck squamous cell carcinoma (HNSCC). Despite its high prevalence, reported in up to 90% of cases, the molecular and spatial mechanisms enabling PNI remain poorly characterized. Here, we apply spatial transcriptomics to delineate the cellular, molecular, and microenvironmental features that drive PNI and its associated immune contexture. Spatial transcriptomic datasets (GSE300147, GSE252265, and GSE281978) were processed using standardized pipelines for quality control, normalization, and unsupervised clustering (UMAP). Cell populations were annotated based on canonical markers, identifying distinct epithelial, neural, Schwann, immune, and axon guidance-related clusters. Composite module scores were calculated to quantify PNI activity, neuronal and Schwann cell identity, axon guidance signaling, and immune infiltration. Correlation analyses and Gene Ontology (GO) enrichment were used to define key biological pathways associated with spatially resolved PNI phenotypes. UMAP visualization revealed spatially distinct neural-enriched tumor clusters co-localizing with Schwann and axon guidance signatures, indicating a neural-like invasive niche. Correlation analysis demonstrated strong associations between PNI and neuronal (r = 0.89) and axon guidance (r = 0.65) programs, with moderate correlation to Schwann cell activity (r = 0.49). In contrast, PNI exhibited weak negative associations with T-cell (r = -0.26) and macrophage (r = -0.12) scores, consistent with an immune-excluded perineural microenvironment. GO enrichment further supported these trends. PNI-high regions showed activation of pathways involved in chemokine-mediated signaling, cellular adhesion, and fluid and lipid transport-processes that may facilitate neural infiltration and axon-tumor communication. PNI-low regions, by contrast, were enriched for immune-related functions including T-cell and B-cell receptor signaling, interleukin-2 production, and tumor necrosis factor regulation, consistent with active antitumor immunity. Collectively, these data reveal that PNI progression coincides with the emergence of a neural-dominant, immune-silent niche enriched for chemotactic and axon guidance signals. Together, these findings define the spatial neuroimmune landscape of PNI in HNSCC, implicating reciprocal nerve-tumor signaling and localized immune evasion as synergistic drivers of invasion. By integrating spatial resolution with transcriptomic profiling, this study provides mechanistic insight into how cancer cells remodel the perineural microenvironment to support invasion and immune escape.