NEW YORK, July 12 — A new AI tool from New York University may offer a cheaper alternative to genomic tests for assessing breast cancer recurrence risk. The bottom line: NYU researchers have built an artificial intelligence system that predicts breast cancer recurrence risk from digital pathology images. The tool matches or outperforms the accuracy of commercial genomic tests like Oncotype DX and MammaPrint, which analyze gene expression in tumor tissue and are expensive.
Krzysztof J. Geras led the study.
The team combined tumor tissue images with basic clinical data. No extra biopsies. No specialized lab equipment.
By the numbers, the comparison matters. Genomic tests are the current clinical standard for assessing recurrence risk and guiding chemotherapy decisions.
They provide a recurrence score that helps oncologists determine who needs aggressive treatment. But those tests are expensive and may not be available in all healthcare settings. The figure that matters: the AI tool uses digital pathology images that are more widely accessible.
If NYU’s approach holds up in larger trials, it could offer a potential alternative that could democratize risk assessment.
How it works
The AI analyzes patterns in tumor tissue. It learns from images, correlating microscopic features with actual patient outcomes. The result is a risk score that, in NYU’s testing, performed as well as or better than the expensive genomic alternatives.
Investors will note the implications. Genomic testing is a multi-billion-dollar market dominated by a handful of companies.
A widely available AI alternative would disrupt that pricing structure entirely.
Access gap
The study addresses a real-world problem. Genomic tests are not universally available. Rural hospitals, community clinics, and healthcare systems in lower-resource countries often lack access to the specialized labs and trained personnel required to run them.
Digital pathology, by contrast, is spreading rapidly. The NYU tool offers a faster option as well.
Genomic tests can take time for results. An AI analysis of a digital image could be completed more quickly. The study results were based on combining tumor tissue images with basic clinical data.
Geras and his team have not yet published the full dataset or methodology in a peer-reviewed journal, so the medical community will be watching for the formal publication and independent validation. What to watch next: whether larger, multi-center trials confirm these findings.
If they do, the standard of care for breast cancer recurrence risk assessment could shift from molecular biology to machine learning. Consult your doctor for medical advice.
Sources
{“@context”: “https://schema.org”, “@type”: “NewsArticle”, “headline”: “AI Matches or Outperforms Standard Genomic Recurrence Tests”, “description”: “NEW YORK, July 12 — A new AI tool from New York University may offer a cheaper alternative to genomic tests for assessing breast cancer…”, “datePublished”: “2026-07-12T09:30:00+00:00”, “author”: {“@type”: “Person”, “name”: “Ava Anderson”}, “publisher”: {“@type”: “Organization”, “name”: “InfoPulse Today”, “url”: “https://newsfeed.infopulsetoday.com/”}, “citation”: [{“@type”: “CreativeWork”, “url”: “https://en.wikipedia.org/wiki/Genomic_testing”}]}
