BACKGROUND AND OBJECTIVE

Positron emission tomography (PET) with prostate-specific membrane antigen (PSMA) in the diagnosis and primary staging of patients with prostate cancer (PCa) has an established role, but recent summative evidence on its actual diagnostic and staging value is still missing. We aimed to collect and analyze published studies reporting the accuracy of PSMA PET for the diagnosis of clinically significant prostate cancer (csPCa) and detection of distant metastases at primary staging before definitive treatment.

METHODS

We performed a systematic review of the literature, by searching the PubMed/MEDLINE, Cochrane library's CENTRAL, EMBASE, and Scopus databases, from inception to April 2024. Two coprimary outcomes were assessed: first, to evaluate the sensitivity, specificity, positive (PPV) and negative (NPV) predictive values of PSMA PET in detecting intraprostatic csPCa on a per-patient level, and second, to assess the positivity rates of metastatic disease in the primary staging, prior to definitive therapy. As a secondary outcome, the diagnostic accuracy of PET PSMA for the detection of lymph nodal invasion (LNI) was tested in a per-patient-level analysis of studies where pelvic lymph node dissection (PLND) was available as the reference standard. Positivity and detection rates were pooled using random-effect models. Preplanned subgroup analyses tested the diagnostic accuracy of PET PSMA across different study cohorts. Variation in PPV and NPV over csPCa and LNI prevalence was evaluated.

KEY FINDINGS AND LIMITATIONS

In total, 12 and 99 studies, with a total of 1533 and 18 649 participants, respectively, were included in the quantitative synthesis for intraprostatic diagnosis and staging. For intraprostatic disease, the sensitivity, specificity, PPV, and NPV of PSMA PET for csPCa were 82% (95% confidence interval [CI] 73-90%), 67% (95% CI 46-85%), 77% (95% CI 63-88%), and 73% (95% CI 56-87%), respectively. At a bivariate analysis, the diagnostic accuracy of PSMA PET estimated through a summary receiver operating characteristic curve-derived area under the curve was 84%, increasing up to 88% when combined with magnetic resonance imaging (MRI). On staging level, PSMA PET results were positive outside the prostate in 23% of the patients, with substantial variation in positivity rates between high-risk (31%) and intermediate-risk (12%) subcohorts. When using PLND as the reference standard (51 studies, 7713 patients), the sensitivity, specificity, PPV, and NPV of PSMA PET were, respectively, 54%, 94%, 77%, and 86%. With higher csPCa and LNI prevalence, a similar increase in PPV and a decrease in NPV were observed.

CONCLUSIONS AND CLINICAL IMPLICATIONS

The current updated systematic review and meta-analysis provides updated evidence on the diagnostic and staging accuracy of PSMA PET in PCa. We reported good accuracy of PSMA PET to discriminate csPCa, particularly when added to MRI, but NPV alone is insufficient to omit a biopsy. Regarding staging, PSMA PET cannot be used alone to determine the need for lymph node dissection (LND) and should be combined with additional clinical information within predictive tools. As such, further research should develop and validate models that incorporate PSMA PET to reliably inform biopsy or LND.