Accurately identifying the extent of prostate cancer is critical to treatment planning, but remains a significant challenge in oncology. Traditional diagnostic methods often result in underestimation of cancer margins, leading to suboptimal interventions. Physicians face difficulty in precisely outlining cancer margins while minimizing damage to non-cancerous tissue. This limitation affects the efficacy of treatments such as focal therapy and may compromise patient outcomes. Now, a new study has demonstrated that using artificial intelligence (AI) significantly improves physicians’ ability to define prostate cancer margins, enhancing diagnostic precision and minimizing misidentification of cancer extent. By analyzing a patient’s MRI data, fusion biopsy, pathology, and biomarkers, AI identifies subtle patterns and indicators of cancerous tissues that are often undetectable in traditional imaging methods.

The study found that Avenda Health’s (Culver City, CA, USA) AI-powered cancer mapping technology called Unfold AI improved the abilities of urologists and radiologists in identifying cancer extent by 45x. In a multi-reader, multi-case study involving seven urologists and three radiologists from five institutions, researchers tested the performance of Unfold AI in comparison to traditional cognitive and hemi-gland cancer margin assessments. Each physician evaluated 50 prostate cancer cases by first defining cancer margins manually, then repeating the task after a four-week interval using the AI-assisted software. The study aimed to evaluate how well Unfold AI could support clinicians in delineating cancer boundaries with accuracy while reducing the inclusion of healthy tissue.

The findings, published in The Journal of Urology, showed that AI-assisted cancer margins achieved a balanced accuracy of 84.7 percent, outperforming cognitively-defined (67.2 percent) and hemi-gland (75.9 percent) approaches. Unfold AI also significantly reduced the underestimation of cancer extent, with a negative margin rate of 72.8 percent compared to just 1.6 percent for cognitively-defined margins. These results suggest that integrating AI tools like Unfold AI into clinical workflows can lead to more precise and effective treatment strategies, ultimately improving patient care. The technology’s recognition by the American Medical Association as a Category III CPT code further signals its potential for widespread adoption. Future plans include expanding the use of Unfold AI to support physicians in delivering more personalized and data-driven prostate cancer interventions.

"This study is important because it shows the ability of AI to not only replicate expert physicians, but to go beyond human ability. By increasing the accuracy of cancer identification in the prostate, more precise and effective treatment methods can be prescribed for patients," said Dr. Wayne G. Brisbane, MD, co-author and physician researcher involved in the study.

http://www.allmycbs.com/en/index.asp