AI-Enabled Lesion Analysis in PSMA-PET

AI Enabled Lesion Analysis in PSMA-PET: Comprehensive Lesion Segmentation, Delineation, and Classification

This poster, presented at PSMA and Beyond, 2026 highlights how AI is transforming PSMA‑PET imaging for prostate cancer clinical trials by enabling automated, comprehensive lesion analysis. It introduces a machine learning–driven workflow that segments total tumor burden, delineates individual lesions, and classifies them by anatomy, replacing historically time‑intensive manual processes. Built using a deep learning model trained on expertly annotated datasets, the approach delivers strong performance while enabling per‑lesion post‑processing for more sensitive and customizable analysis.

Why download this poster?

• Learn how AI-driven lesion segmentation and classification can streamline PSMA‑PET analysis while improving consistency across reads.
• Understand how per‑lesion post‑processing enhances sensitivity and provides more flexible, clinically relevant quantitative endpoints.
• See how automation can reduce reader burden, cutting interpretation time by over 50% while minimizing inter‑reader variability.
• Explore emerging capabilities, including longitudinal lesion tracking and expansion to additional tracers and disease settings.
• Gain practical insight into how these innovations can support more scalable, data-driven decision-making in radiopharmaceutical and oncology trials.