Preoperative PSMA-PET/CT as a predictor of biochemical persistence and early recurrence following radical prostatectomy with lymph node dissection.

2021 
Background This study aims to evaluate the predictive value of lymph nodes (LN) suspicious for metastases on preoperative prostate-specific membrane antigen (PSMA) PET/CT for biochemical persistence (BCP) and early biochemical recurrence (BCR) following robotic-assisted radical prostatectomy (RARP) with extended pelvic LN dissection (ePLND). Methods We evaluated 213 patients with intermediate and high-risk prostate cancer (PCa) who underwent clinical staging with preoperative 68Ga- or 18F-PSMA-PET/CT scan and subsequent RARP with ePLND. Patients were grouped as PSMA- or PSMA+ depending on their LN status on PSMA-PET/CT and subdivided according to histological LN status in pN0 or pN1. Diagnostic accuracy of PSMA-PET/CT for the detection of pN1 was evaluated. BCP was defined as a first postoperative serum PSA level ≥0.1 ng/mL 6-12 weeks following RP. Early BCR was defined as detectable PSA > 0.2 ng/mL within 12 months of follow-up. Univariable logistic regression analyses were used to evaluate the effect of PSMA+ on BCP and BCR. Results Forty patients (19%) were PSMA+. The overall incidence of pN1 was 23%. Sensitivity, specificity, PPV and NPV on a per patient level for the detection of pN1 was 29%, 84%, 35%, and 80% respectively. BCP was observed in 26 of 211 patients (12%) and early BCR in 23 of 110 patients (21%). The presence of PSMA+ was a significant predictor for BCP (OR 7.1, 2.9-17.1 95% CI) and BCR (OR 8.1, 2.9-22.6 95% CI). Conclusion Preoperative PSMA-PET/CT may be a valuable tool for patient counseling for RARP and ePLND as it is a significant predictor for the risk of postoperative BCP and early BCR. We conclude that an ePLND should not be avoided in men with intermediate or high-risk PCa and preoperative negative PSMA-PET/CT, as 20% have microscopic LN metastasis.
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    24
    References
    0
    Citations
    NaN
    KQI
    []