AI outperforms conventional methods of predicting prostate cancer spread

September 23, 2024
by Gus Iversen, Editor in Chief
New research suggests an algorithm may be better at predicting the spread — or extracapsular extension (ECE) — of prostate cancer than conventional methods like MR, PSMA PET/CT, and nomograms.

A study published in BJUI Compass documents the superior accuracy of Unfold AI, an FDA-cleared algorithm developed by Avenda Health, when utilized by researchers at UCLA.

ECE, the spread of cancer beyond the prostate, is a critical consideration when weighing treatment approaches, as it can increase the risk of positive surgical margins and recurrence. Accurate detection is essential for guiding treatment decisions and reducing cancer recurrence risks.

The study analyzed data from 147 patients who underwent MR-targeted biopsies followed by radical prostatectomy. Results showed that Unfold AI significantly outperformed conventional methods in predicting ECE, including a 40%–60% reduction in false negatives compared to MR.

"By providing a more precise assessment of extracapsular extension, our AI tool empowers physicians to make better-informed decisions, ultimately improving patient outcomes," said Shyam Natarajan, CEO of Avenda Health.

The study builds on earlier findings published in The Journal of Urology, where Unfold AI demonstrated 84% accuracy in identifying the extent of prostate cancer. Recently, Unfold AI was added to the 2024 Current Procedural Terminology (CPT) code set as a Category III code, making it billable by insurance and improving access to personalized prostate cancer care.

Avenda Health, based in Culver City, California, continues to advance the use of AI in prostate cancer management, aiming to set new standards for care and improve patient outcomes.