Related Experiment Video
Updated: Nov 10, 2025

Machine Learning Algorithms for Early Detection of Bone Metastases in an Experimental Rat Model
Published on: August 16, 2020
Predicting the Risk of Metastases by PSMA-PET/CT-Evaluation of 335 Men with Treatment-Naïve Prostate Carcinoma
Stefan A Koerber1,2,3, Johannes Boesch4, Clemens Kratochwil4,5
1Department of Radiation Oncology, Heidelberg University Hospital, 69120 Heidelberg, Germany.
Abstract:
Men diagnosed with aggressive prostate cancer are at high risk of local relapse or systemic progression after definitive treatment. Treatment intensification is highly needed for that patient cohort; however, no relevant stratification tool has been implemented into the clinical work routine so far. Therefore, the aim of the current study was to analyze the role of initial PSMA-PET/CT as a prediction tool for metastases. In total, 335 men with biopsy-proven prostate carcinoma and PSMA-PET/CT for primary staging were enrolled in the present, retrospective study. The number and site of metastases were analyzed and correlated with the maximum standardized uptake value (SUVmax) of the intraprostatic, malignant lesion. Receiver operating characteristic (ROC) curves were used to determine sensitivity and specificity and a model was created using multiple logistic regression. PSMA-PET/CT detected 171 metastases with PSMA-uptake in 82 patients. A statistically significant higher SUVmax was found for men with metastatic disease than for the cohort without distant metastases (median 16.1 vs. 11.2; p < 0.001). The area under the curve (AUC) in regard to predicting the presence of any metastases was 0.65. Choosing a cut-off value of 11.9 for SUVmax, a sensitivity and specificity (factor 1:1) of 76.0% and 58.4% was obtained. The current study confirms, that initial PSMA-PET/CT is able to detect a relatively high number of treatment-naïve men with metastatic prostate carcinoma. Intraprostatic SUVmax seems to be a promising parameter for the prediction of distant disease and could be used for treatment stratification-aspects which should be verified within prospective trials.

