PD60-02: Can pelvic lymph node dissection be omitted at the time of radical prostatectomy in patients with a negative preoperative 68Ga-PSMA PET/CT? The importance of patient risk stratification according to individual risk of lymph node metastases

Elio Mazzone*, Milan, Italy, Daniele Robesti, Giorgio Gandaglia, Armando Stabile, Carlo Andrea Bravi, Andrea Necchi, Daniele Raggi, Laura Marandino, Milano, Italy, Giuseppe Cirulli, Milan, Italy, Eleonora Balestrazzi, Stefano Fanti, Bologna, Italy, Vito Cucchiara, Milan, Italy, Andrea Farolfi, Lorenzo Bianchi, Riccardo Schiavina, Bologna, Italy, Francesco Montorsi, Alberto Briganti, Milano, Italy

Introduction: The role of PSMA PET/CT in the primary staging for patients with prostate cancer (PCa) is debated. Although previous studies suggested a high negative predictive value (NPV) in intermediate risk patients, the variation of false negative (FN) rate according to the prevalence of lymph node invasion (LNI) based on preoperative LNI risk has never been tested. We explored the variation of FN rates in a consecutive cohort of men receiving preoperative PSMA PET.

Methods: We identified 92 patients evaluated with 68Ga-PSMA PET/CT before RP and extended lymph node dissection (LND) between 2019 and 2021 at two referral centers. The main outcome was FN rate, defined as presence of LNI at pathology in patients with negative preoperative scan. To test the correlation between LNI prevalence and FN rate, individual preoperative LNI risk was calculated based on the Briganti nomogram. We investigated the association between the preoperative LNI risk and FN using logistic regression. Thereafter, the relationship between preoperative LNI risk and FN rate was graphically explored using the Lowess function.

Results: Overall, 30 patients had positive spots at preoperative scan. Median LNI risk was 31% (IQR 17-52%). Overall, 13 patients (22%) had FN findings. On a patient-based analysis, PSMA PET/CT showed 59% sensitivity, 81% specificity, 78% NPV and 63% positive predictive value in the detection of LNI. At logistic regression, higher LNI risk was significantly associated with increased risk of missing positive nodes at PSMA PET (HR 1.36, p=0.02). At Lowess analysis, the FN rate progressively increased with higher LNI risk, with rates ranging from 7 to 35% between a preoperative LNI risk of 5 and 90%. Notably, FN rates remained equal or lower 10% in patients with a low risk of LNI ( < 15%) (Fig. 1).

Conclusions: We confirmed that the NPV of preoperative PSMA PET/CT is strongly correlated with LNI prevalence. Notably, in patients with low LNI risk at nomograms and negative PSMA PET, the omission of LND could be considered in the light of the very low rate of FN. This evidence reinforces the need to prospectively explore the diagnostic role of PSMA PET compared to LND in patients with intermediate risk disease.

Source of Funding: None