AI Gives Pathologists 'Spatial Super Vision' to Detect Hidden Cancer Markers

A new AI screening tool can detect hidden genetic markers of cancer in standard patient tissue samples — the kind that pathologists examine every day in hospitals around the world. The tool, called STimage, was developed by scientists at QIMR Berghofer Medical Research Institute and published this week in Nature Communications.

The breakthrough is in what STimage can see. Standard pathology uses H&E staining — a 150-year-old technique that highlights cell structure but reveals nothing about the underlying molecular activity. STimage applies spatial transcriptomics analysis over a standard H&E slide, generating predictions about disease based on molecular patterns that are invisible to the human eye.

Superman Vision for Pathologists

"It's like giving pathologists the super-resolution vision of Superman or Superwoman to scan millions of invisible biomarkers in a tiny tissue sample to find the two or three that are showing signs of cancer," said Associate Professor Quan Nguyen, who led the development. "This capability is critical for earlier detection, more precise diagnosis, and better-informed treatment decisions."

In testing, STimage accurately predicted breast, skin, and kidney cancers, as well as primary sclerosing cholangitis, a liver immune disease. It outperformed the handful of comparable tools that exist in this nascent field. Critically, it also generated accurate predictions about prognosis and treatment response — correctly classifying patients as high or low risk of survival and likely to respond to existing drugs.

Transparency as a Feature

One of STimage's distinguishing characteristics is interpretability. Most AI diagnostic tools produce a prediction without explaining why. STimage shows the specific tissue and cellular features that led to its prediction, along with a mathematical confidence score. That transparency is essential for clinical adoption — pathologists need to understand and evaluate AI findings, not simply accept them.

"The STimage tool does not replace the experience and expertise of pathologists. Rather, it assists them by providing extra information about cell types and genetic activity that they can't see with their own eyes."

— Associate Professor Quan Nguyen, QIMR Berghofer

Access and Equity

The potential equity implications are significant. Spatial transcriptomics analysis currently requires specialist equipment and expertise concentrated in major research centres. STimage runs on standard H&E slides — the most common and affordable pathology preparation in the world. That means the technology could, in principle, extend advanced diagnostic capabilities to regional and rural hospitals that currently lack access to specialist oncology services.

The research team is continuing to broaden the cancer types STimage can detect and improve its accuracy on rarer cell types. The next stage is clinical trials in pathology labs. If those succeed, the team hopes the tool could enter clinical practice within two years.

What to Watch

The clinical trial results will be the critical test. AI diagnostic tools have a mixed track record when moving from controlled research settings to real-world clinical environments. The interpretability features and the focus on pathologist augmentation rather than replacement are both positive signals — but the proof will be in the pathology lab, not the research paper.