Follow-up of lesions with equivocal radiotracer uptake on PSMA-targeted PET in patients with prostate cancer: Predictive values of the PSMA-RADS-3A and PSMA-RADS-3B categories

Abstract

Prostate-specific membrane antigen (PSMA)–targeted PET imaging has become commonly used in patients with prostate cancer (PCa). The PSMA reporting and data system version 1.0 (PSMA-RADS version 1.0) categorizes lesions on the basis of the likelihood of PCa involvement, with PSMA-RADS-3A (soft-tissue) and PSMA-RADS-3B (bone) lesions being indeterminate for the presence of disease. We retrospectively reviewed the imaging follow-up of such lesions to determine the rate at which they underwent changes suggestive of underlying PCa. Methods: PET/CT imaging with 18F-DCFPyL was performed in 110 patients with PCa, and lesions were categorized according to PSMA-RADS version 1.0. The study reported herein is a retrospective analysis of those patients. Fifty-six of 110 (50.9%) patients were determined to have indeterminate PSMA-RADS-3A or PSMA-RADS-3B lesions, and 22 of 56 (39.3%) patients had adequate follow-up to be included in the analysis (median follow-up time was 10 mo [range, 3–22 mo]). The SUVmax of the lesions was obtained, and the ratios of SUVmax of the lesions to SUVmean of blood pool (SUVmax-lesion/SUVmean-bloodpool) were calculated. Predetermined criteria were used to evaluate the PSMA-RADS-3A and PSMA-RADS-3B lesions on follow-up imaging to determine whether they demonstrated evidence of underlying malignancy. Results: A total of 46 lesions in 22 patients were considered indeterminate for PCa (i.e., PSMA-RADS-3A [32 lesions] or PSMA-RADS-3B [14 lesions]) and were evaluable on follow-up imaging. Twenty-seven of 46 (58.7%) lesions demonstrated changes suggesting they were true-positive for PCa. These lesions included 24 of 32 (75.0%) PSMA-RADS-3A lesions and 3 of 14 (21.4%) lesions categorized as PSMA-RADS-3B. The ranges of SUVmax and SUVmax-lesion/SUVmean-bloodpool overlapped between those lesions demonstrating changes consistent with malignancy on followup imaging and those lesions that remained unchanged on follow-up. The presence of additional definitive sites of PCa (PSMA-RADS-4 and PSMA-RADS-5) increases the likelihood that indeterminate lesions will manifest as true-positive on follow-up imaging. Conclusion: PSMA-RADS-3A and PSMA-RADS-3B lesions are truly indeterminate in that proportions of findings in both categories demonstrate evidence of malignancy on follow-up imaging. Overall, PSMA-RADS-3A lesions are more likely than PSMA-RADS-3B lesions to represent sites of PCa, and this information should be considered when guiding patient therapy.