Insilico Medicine Uses AI to Discover Targets for Rare Sinonasal Cancer

CAMBRIDGE, Mass. — Aug. 4, 2026 — A new study is showing how artificial intelligence and multi-omic biology can expose therapeutic opportunities in one of the world’s rarest and least understood cancers. Published in npj Precision Oncology, the research provides the first comprehensive molecular characterization of inverted papilloma-associated sinonasal squamous cell carcinoma, or IP-SNSCC, an aggressive cancer that develops when a previously benign growth in the nasal cavity undergoes malignant transformation.

The study, conducted by researchers from Insilico Medicine, the University of Chicago, and Johns Hopkins University, addresses a central challenge in rare-cancer research: there are often too few patients, tissue samples, and molecular datasets to support the conventional process of identifying drug targets. Instead of relying on a single genetic alteration or a large population study, the investigators combined multiple layers of biological information to reconstruct how IP-SNSCC evolves and to identify molecular vulnerabilities that could guide future treatment development.

IP-SNSCC arises from inverted papillomas, benign but locally aggressive tumors that form in the sinonasal tract. In a subset of patients, these lesions progress into squamous cell carcinoma, a malignant disease capable of invading surrounding tissue and spreading. Surgery and radiation remain central to treatment, but therapeutic options are limited once the cancer becomes advanced or recurrent. Because the disease is so uncommon, it has received far less molecular research than more prevalent cancers, leaving clinicians with few targeted strategies.

To follow the transition from benign growth to invasive cancer, the researchers examined matched samples representing normal tissue, inverted papilloma, and carcinoma. They analyzed whole-exome sequencing to identify changes in protein-coding regions of the genome, RNA sequencing to measure gene activity, and mitochondrial DNA sequencing to investigate alterations in the cell’s energy-producing organelles. Studying these sample stages together allowed the team to distinguish changes associated with malignant progression from molecular features that may simply reflect a patient’s normal genetic background.

The results did not point to one dominant mutation responsible for the disease. Instead, they revealed a coordinated biological shift involving several interconnected systems. Abnormalities in cell-cycle regulation suggested that cancer cells were acquiring greater capacity for uncontrolled growth. Changes in extracellular matrix remodeling indicated that the tissue environment was being reorganized in ways that could help tumor cells invade nearby structures. Disrupted immune signaling pointed to altered communication between malignant cells and the immune system, while metabolic reprogramming suggested that tumor cells were changing how they generate and use energy.

This pattern is important because cancer biology is often driven not by a single defective gene but by networks of interacting pathways. A tumor may compensate when one pathway is blocked, making isolated genetic findings difficult to translate into treatment. By examining gene expression, genetic variation, mitochondrial alterations, signaling pathways, and protein networks together, the study offers a more detailed view of the biological machinery that supports IP-SNSCC. The resulting molecular profile could serve as a reference point for researchers investigating how the disease begins, progresses, and responds to therapy.

The team then used PandaOmics, Insilico Medicine’s artificial-intelligence platform for biological target discovery, to prioritize potential treatment targets. Its TargetID algorithms integrated the study’s transcriptomic data with pathway-level biology, protein–protein interaction networks, genetic evidence, and assessments of whether candidate proteins could realistically be targeted by drugs. This approach is designed to convert a complex molecular dataset into a ranked set of biological hypotheses, helping researchers focus experimental resources on the most promising opportunities.

Among the prioritized candidates were proteins already targeted by approved medicines, raising the possibility that some existing drugs could eventually be evaluated for repurposing in IP-SNSCC. The analysis also highlighted previously unexplored targets that could support new drug-discovery programs designed specifically for this cancer. These findings do not establish that any candidate will benefit patients, and the researchers emphasized that laboratory studies, independent validation, and clinical trials will be required before therapeutic conclusions can be drawn. Nevertheless, the work demonstrates how AI can help overcome the “small data” problem that has historically slowed research into rare diseases.

Alex Zhavoronkov, PhD, founder and chief executive officer of Insilico Medicine and co-corresponding author of the study, said that combining comprehensive multi-omic profiling with PandaOmics enabled the team to generate actionable therapeutic hypotheses in a disease where conventional approaches have struggled. The study’s broader significance extends beyond sinonasal cancer: it presents a potential framework for investigating other rare malignancies in which patient numbers are small but the need for effective treatment is substantial. By linking molecular evolution to druggability, the researchers hope to move rare-cancer research more rapidly from biological description toward translational testing.

The paper, titled “Comprehensive multi-omic dissection and AI-prioritized target identification in inverted papilloma–associated sinonasal squamous cell carcinoma,” was published online in npj Precision Oncology on July 10, 2026. Insilico Medicine, a clinical-stage biotechnology company, uses artificial intelligence and automated technologies in drug discovery programs spanning oncology, fibrosis, immunology, pain, obesity, metabolic disorders, and other fields. The company is listed on the Main Board of the Hong Kong Stock Exchange under the ticker 3696.

News Publication Date:
Aug. 4, 2026

References:
npj Precision Oncology; article publication date: July 10, 2026.