Two students built a map for a difficult cancer decision

Immunotherapy can help the immune system attack cancer, but it does not work for every patient.

Anshul Raghav and Krishay Singh, two students from Washington state, focused on that uncertainty. Their project, OncoMap, uses patterns in tumour data to predict whether a person with head-and-neck cancer may respond to a type of immunotherapy called a PD-1 inhibitor.

In April, their team, The Cartographers, was named a 2026 Pete Conrad Scholar in Health and Nutrition.

The students describe OncoMap as a two-stage platform. One part uses RNA-sequencing data to predict response. A second compares features with similar cases to give more context around the result.

The aim is not to discover a new drug. It is to improve a difficult treatment decision using information already present in a tumour.

That is a serious claim. A model can perform well on historical data and still fail when it meets patients from a different hospital or population. Medical tools also need clinicians to understand when a result is uncertain and when it should be ignored.

The Conrad Challenge recognised the project’s scientific and entrepreneurial promise. Its award page says more than 1,750 projects from over 70 countries entered the 2025–2026 competition, with finalists presenting to expert judges at Space Center Houston.

Winning the category gives Raghav and Singh a strong outcome. It does not make OncoMap a clinical product.

Before a hospital could use the model, researchers would need to inspect its training data, repeat the results on independent patients and compare its recommendations with current medical practice. Regulators and hospitals would also need evidence that using it improves decisions without creating new harm.

For now, OncoMap is best understood as an ambitious student research platform built around a real medical uncertainty, with a major competition award behind it and the hardest validation still ahead.

■