NEW YORK — A team at New York University has developed an artificial intelligence tool that predicts the recurrence of breast cancer using a combination of digital images of tumour tissue and basic clinical information. The research, led by NYU’s Krzysztof J. Geras, demonstrates that the AI system can match or outperform widely used genetic tests.
The work sits at the intersection of NYU’s strengths in computer science and clinical medicine. The university, a major research institution in New York City, has been active in developing AI applications for healthcare, including diagnostic tools for various diseases.
This latest effort targets a specific, high-stakes question: after a breast cancer diagnosis, will the disease return? The AI does not rely on genomic sequencing or expensive molecular assays. Instead, it learns patterns from digital images of tumour tissue and augments that with basic clinical information.
The detail that matters is that this approach, if validated further, could make recurrence prediction accessible in settings where genetic tests are unavailable or too costly. Geras and his team compared their model’s performance against established genetic tests.
The findings require careful interpretation: the AI matched or outperformed those benchmarks, but the study’s scope and the precise margins of improvement are not yet public in full. What is clear is that NYU is betting on a future where machine learning augments — and in some cases replaces — the expensive molecular biology that has dominated oncology for the past decade. The research is part of a wider push at NYU to embed AI into clinical workflows.
The university has already developed diagnostic tools for other diseases, leveraging its dual expertise in machine learning and patient care. For breast cancer specifically, the ability to predict recurrence without sending tissue samples to a central lab would represent a genuine shift. The medical AI field is littered with promising algorithms that never made it past the paper stage.
NYU’s institutional commitment to translating research into practice — and its track record with earlier diagnostic tools — suggests this may be different. The immediate question is how quickly this tool can move from the research bench to the clinic.
Geras and his team will need to demonstrate that the AI works consistently across different patient populations and hospital systems. NYU’s existing infrastructure for AI in healthcare gives it a running start, but the regulatory pathway for a software-as-medical-device remains demanding. For now, the detail that matters is this: a tool that matches genetic tests without the genetic testing.
That is the kind of substitution that could reshape how oncologists think about risk.
Sources
{“@context”: “https://schema.org”, “@type”: “NewsArticle”, “headline”: “New York University’s Role in Advancing AI for Cancer Diagnostics”, “description”: “NEW YORK — A team at New York University has developed an artificial intelligence tool that predicts the recurrence of breast cancer using a combination…”, “datePublished”: “2026-07-12T09:30:00+00:00”, “author”: {“@type”: “Person”, “name”: “Sebastian Smith”}, “publisher”: {“@type”: “Organization”, “name”: “InfoPulse Today”, “url”: “https://newsfeed.2bme.xyz/”}, “citation”: [{“@type”: “CreativeWork”, “url”: “https://en.wikipedia.org/wiki/New_York_University”}]}
