Diagnostic Accuracy of Magnetometer-Guided Sentinel Lymphadenectomy After Intraprostatic Injection of Superparamagnetic Iron Oxide Nanoparticles in Intermediate- and High-Risk Prostate Cancer Using the Magnetic Activity of Sentinel Nodes

Due to the high morbidity of extended lymph node dissection (eLND) and the low detection rate of limited lymph node dissection (LND), targeted sentinel lymph node dissection (sLND) was implemented in prostate cancer (PCa). Subsequently, nonradioactive sentinel lymph node (SLN) detection using magnetic resonance imaging (MRI) and a magnetometer after intraprostatic injection of superparamagnetic iron oxide nanoparticles (SPIONs) was successfully applied in PCa. To validate the reliability of this approach, considering the magnetic activity of SLNs or whether it is sufficient to dissect only the most active SLNs as shown in other tumor entities for radio-guided sLND, we analyzed magnetometer-guided sLND results in 218 high- and intermediate-risk PCa patients undergoing eLND as a reference standard. Using a sentinel nomogram to predict lymph node invasion (LNI), a risk range was determined up to which LND could be dispensed with or sLND only would be adequate. In total, 3,711 LNs were dissected, and 1,779 SLNs (median, 8) were identified. Among 78 LN-positive patients, there were 264 LN metastases (median, 2). sLND had a 96.79% diagnostic rate, 88.16% sensitivity, 98.59% specificity, 97.1% positive predictive value (PPV), 93.96% negative predictive value (NPV), 4.13% false-negative rate, and 0.92% additional diagnostic value (LN metastases only outside the eLND template). For intermediate-risk patients only, the sensitivity, specificity, PPV, and NPV were 100%. Magnetic activities of SLNs were heterogeneous regardless of metastasis. The accuracy of predicting the presence of metastases for each LN from the proportion of activity was only 57.3% in high- and 65% in intermediate-risk patients. Patients with LNI risk of less than 5% could have been spared LND, as no positive LNs were found in this group. For patients with an LNI risk between 5% and 20%, sLND-only would have been sufficient to detect almost all LN metastases; thus, eLND could be dispensed with in 36% of pat