Sentinel lymph node staging with superparamagnetic iron oxide nanoparticles (spio) and magnetic resonance imaging: mag-stage
Pontus Zaar, Olivia Näslund, Oscar Jalnefjord, Jenny Nyqvist-Streng, Per Nyman, Helena Sackey, Erica de Coursey, Iva Johansson, Roger Olofsson Bagge, Henrik Leonhardt, Fredrik WärnbergAbstract
Introduction
Sentinel lymph node (SLN) status is essential for staging in breast cancer (BC). SLN-detection with technetium has logistical and regulatory limitations. Superparamagnetic iron oxide nanoparticles (SPIO) combined with magnetic resonance imaging (MRI) offer a radiation-free alternative with a potential of nodal staging. The SPIO-MRI performance for SLN-staging was evaluated in BC-patients.
Methods
Clinically node-negative (N0) patients scheduled for SLN-biopsy received 0.1 ml Magtrace intradermally. SPIO-MRI was performed to localize SLNs and assess nodal status based on morphology and SPIO uptake patterns. Histopathology served as reference. Diagnostic performance was assessed at nodal and patient levels.
Results
A total of 283 nodes were identified on SPIO-MRI of which 206 were removed. Node-level analysis included 194 SLNs, of which 18 (9.3%) were metastatic. At nodal level, sensitivity and specificity were 38.9% (95% c.i. 17.3–64.3) and 76.7% (69.6–82.9%), with a positive (PPV) and negative predictive value (NPV) of 14.6 and 92.5%. SPIO-MRI detected metastases were significantly larger than those not detected: median 11mm (true-positive) versus 2mm (false-negative). At patient level, sensitivity was 53.3% (26.6–78.7%) and specificity 66.7% (55.4–76.7%), with a PPV and NPV of 22.2% (11.7–38.1%) and 88.9% (78.4–95.4%), respectively.
Discussion
SPIO-MRI demonstrated moderate diagnostic performance for SLN-staging in N0-patients, with a high NPV suggesting potential as a rule-out tool. The low PPV limits SPIO-MRI to reliably confirm metastases. Detection was strongly influenced by metastatic size. These findings support a complementary role for SPIO-MRI in SLN staging, but further refinement and validation in larger cohorts are required.