AI uncovers hidden bladder cancer warning signs years before diagnosis

Science Daily
Read full story· 7 min
Researchers at the University of Plymouth developed an AI model, PRECISE-AGZ, that analyzes electronic health records to detect bladder cancer risk years before diagnosis, identifying 85% of cases by recognizing subtle warning signs.

What happened

Researchers at the University of Plymouth created an AI model named PRECISE-AGZ to analyze electronic health records for early signs of bladder cancer. The team studied records from nearly 70,000 patients collected over 25 years. The AI system was able to identify subtle warning signs of bladder cancer as early as five years before diagnosis. The model detected 85% of bladder cancer cases by recognizing patterns in medical data that are often missed by current screening methods.

Key details

  • The AI model, PRECISE-AGZ, examined 48,261 possible health indicators and narrowed them to 38 key features for predicting risk.
  • The study used electronic health records from 1995 to 2020, covering almost 70,000 patients.
  • PRECISE-AGZ correctly identified 85% of bladder cancer cases and 91% of non-cases.
  • The AI system outperformed existing NHS referral guidelines for bladder cancer detection.
  • Some warning signs were detectable up to five years before official diagnosis.

Why it matters

Doctors and healthcare providers could use this AI system to identify patients at risk for bladder cancer much earlier than current methods allow. This could lead to earlier treatment, fewer unnecessary invasive procedures, and better allocation of medical resources. Patients may benefit from more accurate screening and a reduced likelihood of missed or delayed diagnoses.

#medical-diagnostics#precise-agz#university-of-plymouth

More in Healthcare