Researchers in the United States have developed an artificial intelligence (AI) model that identifies and measures aggressive prostate cancer lesions, aiding in treatment decisions. The findings were published in the medical journal Radiology on October 29, 2024.
Prostate cancer is the second most common cancer among men in Brazil, making early diagnosis crucial for successful treatment. The AI model was trained using MRI scans from 732 patients, successfully identifying and marking 85% of the most aggressive lesions.
The model also linked larger tumor volumes to a higher risk of treatment failure and metastasis, regardless of other factors. David D. Yang, the study's lead author, stated, "The tumor volume determined by AI has the potential to promote precision medicine for prostate cancer patients, enhancing our ability to understand cancer aggressiveness and recommend appropriate treatments."
The AI's size estimates showed prognostic potential, indicating that larger tumors were associated with a greater likelihood of cancer recurrence. The model can also assist radiologists in pinpointing tumor locations, enabling more targeted treatment.
Looking ahead, researchers plan to validate their findings with larger, multi-institutional datasets to ensure the model's generalizability across diverse patient populations.