Prostate-specific membrane antigen (PSMA)-PET/CT has shown considerable promise in the evaluation of prostate cancer bone metastases; however, data utilizing a histopathologic reference standard in this setting are limited. We therefore sought to evaluate the diagnostic performance of PSMA-PET/CT using a consistent histopathologic gold standard in the form of bone biopsy.

In this single-center, retrospective study, we identified 80 patients with prostate cancer who underwent CT-guided bone biopsy of a tracer-avid osseous lesion on PSMA-PET/CT performed with F-piflufolastat. Concordance between PET/CT and histopathology and the positive predictive value of PSMA-PET/CT were determined. Factors predictive of positive biopsies were also evaluated.

PSMA-PET/CT and bone biopsy results were concordant in 55/80 patients (69%), and the positive predictive value of PSMA-PET/CT for osseous metastasis of prostate cancer was 66% (53/80). Positive predictive values for spine, pelvis, and rib biopsies were 82% (23/28), 72% (18/25), and 26% (5/19), respectively. Peak SUV and its ratio to liver mean SUV were significantly higher in biopsy-positive lesions compared to biopsy-negative lesions. A threshold peak SUV to liver mean SUV ratio of 1.7 had a sensitivity of 61% and a specificity of 92% for a histopathologic diagnosis of metastatic prostate cancer.

PSMA-PET/CT has a moderately high histopathologic concordance and positive predictive value for the diagnosis of osseous metastatic disease in prostate cancer. Peak SUV is useful for distinguishing biopsy-positive from biopsy-negative lesions. In keeping with prior investigations, a majority of biopsied rib lesions were negative for metastatic prostate cancer.