Detecting Prostate Cancer Subtypes with Empire AI

Prof. Ekta Khurana | Weill Cornell Medicine
Using artificial intelligence to analyze pathology images and identify aggressive, treatment-resistant prostate cancer subtypes earlier and more accurately.

Challenge

Each year, approximately 15,550 patients are diagnosed with prostate cancer, making it one of the most common cancers among men. While many cases are treatable, a subset of patients develop aggressive forms such as castration-resistant prostate cancer (CRPC), which no longer responds to standard hormone therapies.

These aggressive subtypes are particularly difficult to detect early because their defining characteristics can be subtle and easily missed using traditional diagnostic approaches. Pathologists rely on visual inspection of tissue samples, which can vary between observers and may not capture complex patterns embedded in large datasets.

As a result, patients may not receive the most effective treatment at the earliest stage, leading to poorer outcomes and significantly higher healthcare costs. The challenge lies in identifying these high-risk cases earlier and more reliably

Researcher’s Approach Using Empire AI

The team uses AI to analyze thousands of pathology images to identify patterns linked to aggressive cancer subtypes.

Empire AI enables large-scale training on high-resolution medical imaging datasets.

Empire AI Enables

• Detection of patterns invisible to clinicians
• Integration of large datasets
• Faster and more accurate model development

Without Empire AI, Limited computing would restrict dataset size and reduce diagnostic accuracy.

Potential Impacts

• Enables earlier and more accurate detection of aggressive prostate cancer
• Improves treatment selection and patient outcomes
• Reduces reliance on costly late-stage treatments
• Advances the use of AI as a standard tool in clinical diagnostics