A groundbreaking artificial intelligence system developed by the University of Plymouth can spot signs of bladder cancer up to five years before a clinical diagnosis by analyzing electronic health records. The model, named PRECISE-AGZ, achieved 85% accuracy in identifying affected patients and could transform early detection of a disease that causes 220,000 deaths annually worldwide.

A Promising Breakthrough in the Fight Against Bladder Cancer

Researchers at the University of Plymouth, in the United Kingdom, have developed an artificial intelligence (AI) system capable of identifying early indicators of bladder cancer up to five years before a clinical diagnosis is made. The findings, published in the journal IEEE Transactions on Biomedical Engineering, mark a significant step forward in combating a disease that claims around 220,000 lives each year globally.

The study, led by Professor Shang-Ming Zhou from the University of Plymouth's Centre for Health Technology, with data analysis by PhD student Xu Wang, was based on the examination of electronic health records from nearly 70,000 patients in Wales, collected between 1995 and 2020.

How Does the PRECISE-AGZ System Work?

The model, called PRECISE-AGZ, analyzed 48,261 health indicators from patients, from which it selected 38 as key factors associated with the development of bladder cancer. These factors include smoking, levels of physical exercise, medication use, and previous medical conditions.

The AI classified patients into three categories: low risk, uncertain zone, and high risk. Results showed that the system correctly identified cancer in 85% of affected patients and ruled out tumors in 91% of healthy individuals. Moreover, warning signals were consistently detected one year in advance, and in some cases, up to five years before diagnosis.

What is hematuria? Hematuria is the presence of blood in the urine, one of the most common signs of bladder cancer. However, it can also be linked to other benign conditions such as kidney stones or prostate disorders, making it difficult to use as a sole indicator. AI could help differentiate these cases.

Secondary Findings and Limitations

The study also revealed interesting associations: patients with Parkinson's disease or dementia showed a lower risk of developing bladder cancer, while prolonged use of tamoxifen (a drug used in breast cancer treatment) was associated with a higher risk. However, researchers clarify that these findings do not prove causal relationships.

One of the main limitations is that the data comes exclusively from Wales, so the model requires validation in other healthcare systems before widespread implementation. Additionally, hematuria (blood in urine), a key sign of the disease, overlaps with other conditions like kidney stones or benign prostate disorders, underscoring the need for more precise tools.

Reactions and Future Prospects

Helen Winter, clinical director of the SWAG Cancer Alliance, highlighted that there is an “unmet need for early detection” of bladder cancer and welcomed the advances of this research. The study authors hope the system can be integrated into healthcare systems to improve patient outcomes and reduce mortality.

This breakthrough adds to other recent developments in using AI for early disease detection, such as breast and lung cancer, reinforcing technology's role as an ally in preventive medicine.

Source: Infobae