Avenda Health's Mapping Technology, Unfold AI Enhances Prostate Cancer Staging Accuracy
Revolutionizing Prostate Cancer Diagnosis with AI
Avenda Health recently revealed groundbreaking results from a pilot study demonstrating that Unfold AI, its AI-powered cancer mapping tool, significantly surpasses conventional MRI in accurately predicting prostate cancer spread. These promising findings were shared at the American Urological Association's 2025 Annual Meeting held on April 27.
Study Design and Results
The study, titled "Prediction of Seminal Vesicle Invasion Using Artificial Intelligence Prostate Cancer Risk Mapping", aimed to assess the ability of Unfold AI to predict seminal vesicle invasion (SVI), an essential factor in prostate cancer staging. The results indicate that Unfold AI achieved a 92% accuracy rate compared to the 52% accuracy of radiologist interpretations using traditional MRI. This substantial improvement highlights the potential of AI to enhance diagnostic precision in prostate cancer management.
The research was conducted by experts from Stanford University School of Medicine and UCLA's David Geffen School of Medicine. It involved two cohorts of men who had undergone MRI scans prior to prostate cancer surgery. The study compared predictions of SVI based on MRI alone with predictions made using Unfold AI, which integrates MRI data with clinical information to produce a 3D cancer map. After surgery, the researchers examined prostate specimens to verify the presence of SVI, thereby determining the accuracy of both diagnostic methods.
In the first cohort, comprising 147 patients, SVI was pathologically confirmed in 25 cases. Unfold AI accurately predicted 92% of these cases, while traditional MRI identified only 52%. In the second cohort, which consisted of 20 patients with 10 cases of SVI, Unfold AI missed 2 cases, while MRI missed 4.
Additionally, Unfold AI demonstrated a lower rate of false positives compared to MRI in both cohorts:
-
First cohort AUC: Unfold AI 0.95 versus MRI 0.80 (p = 0.003).
-
Second cohort AUC: Unfold AI 0.85 compared to MRI 0.74.
These findings underscore the superior performance of Unfold AI in accurately identifying SVI, which is crucial for precise cancer staging and prognosis.
Clinical Impact and Next Steps
Detecting whether prostate cancer has spread to nearby structures, such as the seminal vesicles, is vital for effective treatment planning, particularly when considering surgery or radiation. Current methods, primarily relying on MRI, often result in inaccurate predictions, leading to potential misdiagnosis. The integration of Unfold AI in clinical practice could reduce such errors, enabling physicians to make better-informed treatment decisions.
These results build on previous success for Unfold AI, which demonstrated accuracy in predicting extracapsular extension risk. That study earned Unfold AI the BJUI Compass Prize 2025 at the same annual meeting, further validating the tool’s potential in prostate cancer care.
Advancing Personalized Cancer Care
Commenting on the study, Shyam Natarajan, PhD, co-founder and CEO of Avenda Health, remarked:
These results demonstrate how Unfold AI continues to improve the diagnosis and staging of prostate cancer, enabling the physician to recommend and deliver the best therapy to the patient.
By significantly enhancing the accuracy of prostate cancer staging, Unfold AI supports more precise treatment planning, offering patients better prospects for successful outcomes.
Future Directions
The promising performance of Unfold AI in accurately predicting cancer spread highlights its potential as a transformative tool in personalized cancer care. These findings also lay the groundwork for further research into AI's role in diagnostic accuracy, with the goal of integrating advanced technologies into routine clinical practice for improved patient outcomes.
With Unfold AI proving to be more reliable than traditional MRI, Avenda Health is poised to continue pioneering the use of AI in prostate cancer care, fostering advancements that could significantly improve both diagnosis and treatment planning.