68Ga-PSMA-11 PET has the potential to improve patient selection for extended pelvic lymph node dissection in intermediate to high-risk prostate cancer

Introduction Radical prostatectomy with extended pelvic lymph node dissection (ePLND) is a curative treatment option for patients with clinically significant localised prostate cancer. The decision to perform an ePLND can be challenging because the overall incidence of lymph node metastasis is relatively low and ePLND is not free of complications. Using current clinical nomograms to identify patients with nodal involvement, approximately 75-85% of ePLNDs performed are negative. The aim of this study was to assess the added value of Ga-68-PSMA-11 PET in predicting lymph node metastasis in men with intermediate- or high-risk prostate cancer. Methods Ga-68-PSMA-11 PET scans of 60 patients undergoing radical prostatectomy with ePLND were reviewed for qualitative (visual) assessment of suspicious nodes and assessment of quantitative parameters of the primary tumour in the prostate (SUVmax, total activity (PSMA(total)) and PSMA positive volume (PSMA(vol))). Ability of quantitative PET parameters to predict nodal metastasis was assessed with receiver operating characteristics (ROC) analysis. A multivariable logistic regression model combining PSA, Gleason score, visual nodal status on PET and primary tumour PSMA(total) was built. Net benefit at each risk threshold was compared with five nomograms: MSKCC nomogram, Yale formula, Roach formula, Winter nomogram and Partin tables (2016). Results Overall, pathology of ePLND specimens revealed 31 pelvic metastatic lymph nodes in 12 patients. Ga-68-PSMA-11 PET visual analysis correctly detected suspicious nodes in 7 patients, yielding a sensitivity of 58% and a specificity of 98%. The area under the ROC curve for primary tumour SUVmax was 0.70, for PSMA(total) 0.76 and for PSMA(vol) 0.75. The optimal cut-off for nodal involvement was PSMA(total) > 49.1. The PET model including PSA, Gleason score and quantitative PET parameters had a persistently higher net benefit compared with all clinical nomograms. Conclusion Our model combining PSA, Gleason score and visual lymph node analysis on Ga-68-PSMA-11 PET with PSMA(total) of the primary tumour showed a tendency to improve patient selection for ePLND over the currently used clinical nomograms. Although this result has to be validated, Ga-68-PSMA-11 PET showed the potential to reduce unnecessary surgical procedures in patients with intermediate- or high-risk prostate cancer.