A new machine-learning approach for prostate-specific membrane antigen (PSMA) treatment of metastatic castration-resistant prostate cancer (mCRPC) could estimate radiation dose to tumors and healthy organs before therapy begins. Using data already available from pre-therapy PET/CT scans, this novel prediction tool could help personalize treatment plans, improve patient selection, and reduce toxicity risk.
This research was presented at the Society of Nuclear Medicine and Molecular Imaging 2026 Annual Meeting.
Dosimetry is critical for optimizing ??Lu-PSMA radiopharmaceutical therapy in mCRPC. Currently, post-therapy imaging is typically used to calculate dosimetry; however, it is time-consuming and resource-intensive. Pre-therapy PET/CT offers an opportunity to assess potential treatment effectiveness and risk before therapy.
“18F-PSMA PET/CT is already routinely performed and widely available in prostate cancer patients, but its potential to predict treatment radiation dose has not previously been explored,” 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.
“Our study sought to determine if information already available from these scans could guide treatment planning before therapy begins and support more personalized care.”
In this proof-of-concept study, nine patients with mCRPC referred for ??Lu-PSMA radiopharmaceutical therapy were included, contributing 57 tumors, 36 salivary glands, and 18 kidneys for analysis. Researchers developed a machine learning mixed effects model to predict absorbed doses in tumors and organs.
Predictors included uptake-based PET metrics, radiomic features, and clinical biomarkers. Predictive estimates were compared with dosimetry calculated after one cycle of ??Lu-PSMA therapy to assess accuracy.
The pre-therapy 18F-PSMA PET/CT-based machine learning model showed a promising ability to predict tumor and organ absorbed dose. By combining uptake features, radiomics, and clinical biomarkers while accounting for patient-level variability, the model shows potential for using pre-therapy information to predict post-therapy dosimetry.
“If validated in larger studies, this approach may improve patient selection and support better decision-making during pre-treatment assessment, helping to optimize ??Lu-PSMA therapy for individual patients. More broadly, it highlights how imaging can move beyond diagnosis to actively guiding personalized treatment,” said Nautiyal.
This proof-of-concept research is part of a planned five-year program aimed at collecting more data and developing a robust, validated model. Future work will focus on larger, multi-center cohorts to refine pre-therapy absorbed dose predictions and to perform independent validation to support patient stratification for personalized ??Lu-PSMA radiopharmaceutical therapy in clinical practice.
More information
Abstract 262138: Amit Nautiyal, et al “Machine Learning-Based Pretherapy Prediction of Tumour and Organ Absorbed Dose in 177Lu-PSMA Therapy Using 18F-PSMA PET/CT Radiomics and Biomarkers”
Key medical concepts
Clinical categories
Provided by
Society of Nuclear Medicine and Molecular Imaging
The content is provided for information purposes only.