Comprehensive multi-omic dissection and AI-prioritized target identification in inverted papilloma–associated sinonasal squamous cell carcinoma

Sinonasal squamous cell carcinoma (SNSCC) is a rare malignancy arising de novo or from inverted papilloma (IP), a benign neoplasm with malignant potential. The molecular drivers of IP-associated SNSCC (IP-SNSCC) remain poorly defined, and no effective therapies are available. Progress is hindered by limited preclinical studies and prospective clinical investigations. We performed multi-omic profiling, including whole-exome, RNA, and mitochondrial DNA sequencing (mtDNA-Seq), of matched normal sinonasal epithelium, IP, and SNSCC samples from 11 patients. Analyses revealed a stepwise transcriptional continuum across histological stages, marked by progressive activation of the cell cycle, extracellular matrix remodeling, and metabolic pathways, with suppression of immune and apoptotic signaling. Shared genomic aberrations were detected in only a subset of paired IP and SNSCC specimens, whereas mtDNA-Seq revealed no overlapping mutations, indicating divergent mitochondrial evolution even in clonally related lesions. Given the need for targeted therapies, we applied PandaOmics, an AI-driven target discovery platform, to genes progressively upregulated during IP-SNSCC development. We first prioritized targets with FDA-approved inhibitors, identifying CDK6, EGFR, HDAC, and SRC/YES1 as repurposing candidates, and nominated AURKA, PLK4, TTK, and CDK1/7 as druggable preclinical targets. Together, this study defines the molecular basis of IP-SNSCC and provides a foundation for future translational investigation.