Insilico Study Identifies Targets in Rare Sinonasal Cancer
Insilico Medicine on August 4 highlighted a collaborative npj Precision Oncology study that used multi-omic profiling and its PandaOmics platform to prioritize therapeutic targets for inverted papilloma-associated sinonasal squamous cell carcinoma. The July 10 paper analyzed matched tissue samples from 11 patients and identified repurposing candidates including CDK6, EGFR, HDAC, and SRC/YES1, alongside preclinical targets such as AURKA and PLK4.
Insilico Medicine on August 4 highlighted a collaborative study in npj Precision Oncology that applied multi-omic profiling and AI-driven target prioritization to inverted papilloma-associated sinonasal squamous cell carcinoma, or IP-SNSCC. The paper, published July 10, analyzed matched normal epithelium, inverted papilloma, and carcinoma samples from 11 patients.
According to the study, researchers performed whole-exome sequencing, RNA sequencing, and mitochondrial DNA sequencing, then used Insilico's PandaOmics platform to prioritize genes whose expression increased during disease progression. The work involved researchers affiliated with the University of Chicago, Johns Hopkins University, and Insilico Medicine.
Molecular findings and target ranking
The authors report a transcriptional continuum from benign inverted papilloma to carcinoma, with progressive activation of cell-cycle, extracellular-matrix remodeling, and metabolic pathways. They also report suppression of immune and apoptotic signaling across the histological stages.
The genomic results were less uniform. Shared genomic aberrations appeared in only a subset of paired benign and malignant samples, while mitochondrial sequencing found no overlapping mutations, which the paper describes as evidence of divergent mitochondrial evolution even in clonally related lesions.
PandaOmics first ranked targets with existing FDA-approved inhibitors. The study identified the following repurposing candidates:
- •CDK6
- •EGFR
- •HDAC
- •SRC/YES1
It also nominated AURKA, PLK4, TTK, and CDK1/7 as druggable preclinical targets. These are computationally prioritized hypotheses, not clinical efficacy findings. The paper states that the results provide a foundation for future translational investigation, while also noting that effective therapies for IP-SNSCC are currently unavailable and that limited preclinical and prospective clinical research has constrained the field.
What the study adds
Insilico founder and CEO Alex Zhavoronkov said in the August 4 announcement that combining comprehensive multi-omic profiling with PandaOmics generated therapeutic hypotheses in a disease where conventional approaches have been limited by scarce data. The company described the work as the first comprehensive molecular characterization of IP-SNSCC.
For ML and computational-biology practitioners, the study is a concrete example of AI target ranking being used after disease-stage-specific molecular signals have been established, rather than as a substitute for sequencing and pathology data. In rare-disease settings, comparable workflows can help narrow a large candidate space, but target priority remains distinct from validation in disease models and human trials. The small cohort, 11 patients, is an important constraint when interpreting the generalizability of the molecular patterns and ranked targets.
Key Points
- 1The study combines whole-exome, RNA, and mitochondrial sequencing with PandaOmics to rank targets across IP-SNSCC disease progression.
- 2Researchers identified approved-drug repurposing candidates and preclinical targets, but the findings remain hypotheses rather than demonstrated treatments.
- 3In rare cancers, multi-omic evidence can constrain AI target-ranking workflows, while small cohorts still limit biological generalizability.
Scoring Rationale
The study provides a specific, peer-reviewed example of AI-assisted target discovery integrated with multi-omic cancer profiling. Its direct practitioner relevance is meaningful for computational drug discovery, although the 11-patient cohort and absence of clinical validation limit near-term impact.
Sources
Primary source and supporting public references used for this report.
Practice interview problems based on real data
1,625 SQL & Python problems across 15 industry datasets — the exact type of data you work with.
Try 250 free problems
