AI Model Predicts Radiation Dose Before Prostate Cancer Therapy

Metastatic castration-resistant prostate cancer (mCRPC) is an advanced form of disease that progresses despite androgen-deprivation therapy and frequently spreads to bone and visceral organs. In this setting, clinicians must balance radiopharmaceutical efficacy against dose to healthy tissue, yet accurate dosimetry typically relies on post-therapy imaging that is time- and resource-intensive. This limits opportunities to adjust treatment plans before harm occurs. To help address this challenge, researchers have now developed an AI approach that predicts absorbed radiation dose before therapy begins.

Investigators at University Hospital Southampton and the University of Southampton (Southampton, UK) presented a pre-therapy, positron emission tomography/computed tomography (PET/CT)–based machine-learning model tailored for prostate-specific membrane antigen (PSMA)–targeted radiopharmaceutical therapy. The approach was reported at the Society of Nuclear Medicine and Molecular Imaging 2026 Annual Meeting. It is designed to estimate dose to tumors and organs to inform selection and planning for lutetium-177–labeled PSMA therapy.

The model uses routinely acquired pre-therapy 18F‑PSMA PET/CT scans to derive uptake metrics and radiomic features, integrating these with clinical biomarkers. A mixed-effects framework accounts for patient-level variability when forecasting absorbed dose to disease sites and dose-limited organs such as the kidneys and salivary glands. By shifting dose estimation to the pre-treatment phase, the tool aims to personalize schedules and mitigate toxicity risk.

In a proof-of-concept study, nine patients with mCRPC referred for 177Lu‑PSMA therapy contributed data from 57 tumors, 36 salivary glands, and 18 kidneys. Predictive estimates from the pre-therapy model were compared with dosimetry calculated after the first treatment cycle. The model demonstrated a promising ability to predict tumor and organ absorbed dose based on pre-therapy information.

According to the team, the work forms part of a planned five-year program to expand data collection and refine the algorithm across larger, multicenter cohorts. Future efforts will include independent validation to support patient stratification and guide individualized 177Lu‑PSMA radiopharmaceutical therapy in clinical practice. If validated, the strategy could use widely available imaging to move beyond diagnosis and actively inform pre-treatment decision-making for advanced prostate cancer.

"If validated in larger studies, this approach may improve patient selection and support better decision-making during pre-treatment assessment, helping to optimize 77Lu-PSMA therapy for individual patients. More broadly, it highlights how imaging can move beyond diagnosis to actively guiding personalized treatment," said Amit Nautiyal, Ph.D., scientist and National Institute for Health and Care Research (NIHR) fellow at University Hospital Southampton and the University of Southampton in the United Kingdom.

Related Links
University of Southampton
University Hospital Southampton
Society of Nuclear Medicine and Molecular Imaging

Visit expo >

Digital Intelligent Ferromagnetic Detector

Digital Ferromagnetic Detector

Mammo DR Retrofit Solution

DR Retrofit Mammography

Post-Processing Imaging System

DynaCAD Prostate

Medical Radiographic X-Ray Machine

TR30N HF