Diagnostic impact of interpretation criteria for [¹⁸F]PSMA-1007 PET/CT: Prospective comparison with CT and bone scan.

Abstract

The PSMA Reporting and Data System (PSMA-RADS) has been proposed to standardize the interpretation of PSMA-PET. However, its diagnostic performance has not been fully validated. This study aimed to determine the optimal interpretation criteria for PSMA-PET and to compare its diagnostic performance with conventional imaging. This prospective study enrolled 284 patients across various disease stages who underwent [¹⁸F]PSMA-1007 PET/CT. Diagnostic performance was evaluated using different positivity thresholds based on PSMA-RADS categories, PSMA-expression scores, and SUVmax, and was compared with that of CT and bone scan (BS). True-positive rates were summarized across PSMA-RADS categories. A total of 771 lesions were analyzed. PSMA-RADS demonstrated superior diagnostic performance compared with other interpretation criteria. Using PSMA-RADS-4 as the positivity threshold, [¹⁸F]PSMA-1007 PET/CT achieved a sensitivity of 78.5% and a specificity of 95.3%. PSMA-RADS-3 increased sensitivity to 98.4% but reduced specificity to 58.8%. PSMA-expression score 2 yielded a sensitivity of 52.2% and a specificity of 99.3%, whereas SUVmax cutoff of 5.7 yielded a sensitivity of 72.6% and a specificity of 85.8%. When PSMA-RADS-4 was used as the positivity threshold, PET/CT showed higher sensitivity than CT (81.3% vs. 48.6%) and BS (84.5% vs. 65.7%), as well as higher accuracy (83.2% vs. 56.2% and 87.6% vs. 74.5%, respectively). The true-positive rate of PSMA-RADS-3 lesions was 69.7%, which was substantially lower than that of PSMA-RADS-4 lesions (96.7%). PSMA-RADS-4 represents an optimal positivity threshold for [¹⁸F]PSMA-1007 PET/CT, providing superior diagnostic performance compared with CT and BS and supporting its clinical utility for standardized image interpretation.