Detection of ISUP≥2 Prostate Cancers Using Multiparametric MRI: Prospective Multicenter Comparison of the PI-RADS Score and an Artificial Intelligence System

When there's a concern about prostate cancer, doctors often recommend a special MRI scan called a multiparametric MRI (mpMRI). This scan helps them look for suspicious areas in the prostate before deciding where to take tissue samples (biopsies). The idea is to guide the biopsy needles precisely to areas that might contain cancer. However, reading these MRI scans can be tricky. Sometimes, the scans might show something suspicious that isn't actually a serious cancer, or different doctors might interpret the same scan slightly differently. This means it's not always easy to consistently find the more aggressive cancers that need treatment while avoiding unnecessary biopsies for less serious findings.

To improve this, researchers have been developing a new type of computer program, often called an Artificial Intelligence (AI) system. This AI system has been taught to recognise patterns in MRI images that could indicate important prostate cancers (specifically those graded ISUP≥2, which are generally more serious). Early tests of this AI have been positive, but these tests were done using MRI scans from a single hospital and its specific machines.

This study, called CHANGE, is designed to see if this new AI system works just as well in a wider range of real-world settings. It will involve many different hospitals and various MRI machines. By collecting MRI scans and biopsy results from men across different locations, the researchers can properly test if the AI system can consistently help doctors spot serious prostate cancers, regardless of where the MRI was done. The ultimate goal is to see if this AI can be a valuable tool to make prostate cancer detection more accurate and reliable, helping doctors make better decisions about care.