PSMA-RADS (PSMA reporting and data system)

PSMA-RADS version 1.0, proposed by Rowe, Pienta, Pomper and Gorin at Johns Hopkins in 2017-2018, is a structured reporting system for PSMA-targeted PET imaging modelled on other "reporting and data systems" such as BI-RADS. It assigns each lesion a category from 1 (benign, subdivided into 1A without uptake and 1B with uptake but known-benign) through 2 (likely benign, low-level uptake atypical for prostate cancer), 3A/3B/3C/3D (equivocal, warranting further work-up such as biopsy or 3-6 month follow-up imaging), 4 (intense uptake typical of prostate cancer but without a corresponding finding on conventional imaging) to 5 (intense uptake with a corresponding conventional-imaging finding, prostate cancer almost certainly present). Reports give up to five target lesions a PSMA-RADS category plus an overall scan score equal to the highest individual lesion score.

Interobserver agreement studies on 18F-DCFPyL PET/CT found good-to-excellent agreement (intraclass correlation coefficients of 0.60-0.84 depending on lesion type), supporting its reproducibility across readers. A revised PSMA-RADS 2.0 has since been proposed, folding the 1A/1B split into a single benign category, allowing PSMA-RADS 3A/3B/3D lesions to be reclassified as 4 in patients with widespread metastatic disease, and adding a 5T subcategory for effectively treated metastases after systemic prostate cancer therapy.

The system is complementary to PSMA PET-guided treatment decisions (for example selecting patients for lutetium-177 or actinium-225 PSMA radioligand therapy) but is not itself a treatment-eligibility threshold; eligibility criteria used in trials such as VISION relied on separate PSMA-positivity rules.