Comparing texture analysis of pretreatment 18F-FDG PET and 68Ga-PSMA PET in patients with prostate cancer: investigation of diagnostic efficacy and prognostic biomarker.

Abstract

This study aims to evaluate the predictive accuracy of textural parameters and current parameters of ¹⁸F-fluorodeoxyglucose and ⁶⁸Ga-labeled prostate-specific antigen positron emission tomography (FDG and PSMA PET) images in prostate cancer (PCa) and compare the features retrieved from both scans. Based on symptoms, digital rectal examination (DRE), prostate-specific antigen (PSA) level in the blood, or histopathology from transrectal ultrasound-guided biopsy and 4Kscore Test, 120 patients have confirmed PCa. Sixty of them were scanned on a PET/CT machine using ¹⁸F-FDG, and the other 60 patients were scanned using ⁶⁸Ga-PSMA of radiopharmacy. Each tumour was delineated using PET. Edge texture parameters were used to define each tumour, and 73 features in all were taken from eight distinct texture matrices and computed using the open-source program Chang-Gung Image Texture Analysis (CGITA). Using Spearman correlation, feature correlation with conventional quantitative metrics (Maximum Standardized Uptake Value (SUVmax), Total Lesion Glycolysis (TLG), Metabolic Tumor Volume (MTV)) was investigated, and it was found that the High-Intensity Low-Energy Radiation (HILRE) correlation was strong. PCa was best discriminated by HILRE (64-bin) in receiver operating characteristic curves. It is concluded that ⁶⁸Ga-PSMA-based PET imaging is better than ¹⁸F-FDG-based PET and is strongly associated with PCa tumour allocation. According to extracted features, HILRE is the most significant measure and it is, thus, considered here an independent predictor of PCa prognosis. Although the study's findings are helpful, confirmation by further prospective research is required.