Researchers analyzed records from 723 men admitted to a hospital for benign prostatic hyperplasia. They wanted to find a way to spot those who also had histological prostatitis, a condition found in the prostate tissue. The study found that about 53.5% of these patients had this inflammation.
The team built a machine learning tool to predict the risk. This artificial intelligence model performed better than five other methods tested. It correctly identified the condition in the validation set with an accuracy score of 0.852. The model also showed low error rates and provided the greatest benefit across different risk thresholds.
Several factors helped the model predict the risk. These included prostate volume, neutrophil-to-lymphocyte ratio, International Prostate Symptom Score, age, systemic immune-inflammation index, and acute urinary retention. Higher values for most of these factors, or the presence of acute urinary retention, were linked to a higher predicted risk.
This research comes from a single hospital and used internal data only. While the tool could help doctors identify high-risk individuals and guide treatment decisions, more research is needed to confirm these results in other settings. This study does not prove that these factors cause the inflammation, but they are associated with it.