INTRODUCTION: Nerve-sparing prostatectomy is recommended in cases of organ-confined prostate cancer but is generally contraindicated in patients with suspected extra-prostatic extension (EPE). PSMA ligand imaging has been shown to be valuable in predicting EPE when performed on a hybrid PET/MRI scanner; however, the majority of PSMA PET imaging is performed using PSMA-PET. To our knowledge, there are no established PET/CT criteria for assessing EPE. In this study, we aim to provide a reproducible method for evaluating EPE on PSMA-PET/CT imaging and assess its utility compared with MRI. METHODS: Imaging findings and histopathology were reviewed for 100 consecutive patients who underwent a radical prostatectomy after imaging with MRI and 18F-DCFPyL PSMA-PET/CT. RESULTS: A broad tumour-capsule interface measured using a standardised technique on fused PSMA-PET/CT imaging is associated with a higher risk for established EPE (P < 0.001). In our cohort, applying the criteria of tumour-capsule contact ≥ 10 mm measured on PET/CT was as sensitive as applying PI-RADS version 2 criteria to mpMRI imaging for predicting EPE (74% and 79%, respectively, P = 0.11) and had superior specificity (86% and 61%, respectively, P = 0.035). 93% of MRI-occult lesions were visualised on PSMA-PET/CT. Applying the proposed PET/CT criteria for EPE to this subgroup of 14 patients yielded a sensitivity of 67% and specificity of 92%. CONCLUSION: Our results suggest that tumour-capsule interface measured on fused F18-DCFPyL PSMA-PET/CT imaging is comparable to MRI criteria for predicting the presence of EPE. Applying PET/CT criteria may be of particular benefit in predicting EPE in patients with MRI-occult prostate cancer.