Prediction of Prognosis and Response to Androgen Deprivation Therapy in Intermediate to High-Risk Prostate Cancer Using 18 F-FDG PET/CT Radiomics

Jianyi Qiao,Bitian Liu,Jun Xin,Siang Shen,Han Ma,Shen Pan

DOI: https://doi.org/10.1016/j.acra.2024.06.034

IF: 5.482

2024-07-19

Academic Radiology

Abstract:Rationale and Objectives This study aims to predict intermediate to high-risk prostate cancer (PCa) prognosis based on 18-fluoro-2-deoxyglucose ( 18 F-FDG) positron emission tomography (PET)/computed tomography (CT) radiomics. Additionally, subgroup analysis will be performed on the androgen deprivation therapy (ADT) group and the metastatic PCa group. Materials and Methods In the retrospective analysis of 104 intermediate to high-risk PCa patients who underwent 18 F-FDG PET/CT prior to treatment. The data set was divided into a training set ( n = 72) and a testing set ( n = 32). Two different PET/CT models were constructed using multivariate logistic regression with cross-validation: radiomics model A and an alternative ensemble learning-based model B. The superior model was then selected to develop a radiomics nomogram. Separate models were also developed for the ADT and metastatic PCa subgroups. Results Model A, which integrates eight radiomics features showed excellent performance with an area under curve (AUC) of 0.844 in the training set and 0.804 in the testing set. The radiomics nomogram incorporating the radiomics score (radscore) from model A and the tumor-to-liver ratio (TLR) showed good prognostic accuracy in the testing set with an AUC of 0.827. In the subgroup analyses for endocrine therapy and metastatic cancer, the PET/CT radiomics model showed AUCs of 0.845 and 0.807 respectively, suggesting its potential effectiveness. Conclusion The study establishes the utility of the 18 F-FDG PET/CT radiomics nomogram in predicting the prognosis of intermediate to high-risk PCa patients, indicating its potential for clinical application.

radiology, nuclear medicine & medical imaging