Detection rate and diagnostic accuracy of sentinel-node biopsy in early stage ovarian cancer: A prospective multicentre study (SELLY).

Abstract

e17579 Background: Comprehensive surgical staging of early epithelial ovarian cancer includes systematic paraaortic and bilateral pelvic lymphadenectomy but its therapeutic role is controversial and it is associated with potential severe morbidity. The feasibility of sentinel lymph-node (SLN) identification in early stage ovarian cancer has been shown only. We did a prospective multicenter study to assess the feasibility, detection rate and diagnostic accuracy of the SLN procedure in predicting the pathological pelvic-node status in patients with early stage ovarian cancer. Methods: This is a phase II single arm study (EUDRACT 2019-001088-58) including patients with presumed stage I-II epithelial ovarian cancer planned for immediate or delayed minimally-invasive comprehensive staging. The ovarian pedicle is injected with 2 mL of a 1.25 mg/mL indocyanine green (ICG) solution. The pelvic and lumbo-aortic retroperitoneum is then accessed and inspected to identify and remove SLNs. Staging is then completed including systematic pelvic and para-aortic lymphadenectomy and an histopathological examination of all nodes is performed. The primary endpoint was estimation of the sensitivity and sensibility of SLN in predicting nodal status. Assuming a sensitivity of 98.5% in predicting positive sentinel lymph nodes at histology, a pathological lymph node prevalence of 14.2%, a precision of estimate (i.e. the maximum marginal error) d=5%, a type I error α = 0.05, a sample size of 141 patients is needed to test the general hypothesis (i.e. to answer whether SLN(s) identified with ICG can accurately predict nodal status at histology of patients with apparently early EOC). Assuming a drop-out rate of 20%, a total of 176 patients will be enrolled in the study. Results: One-hundred and seventy-six patients were included with a drop-out rate of 22%. Sentinel node was identified in 102 patients (detection rate, 58%, [95% CI: 51-65]). There were no clinical differences between patients with successful or unsuccessful mapping. Both the infundibolopelvic ligament and the legamentum ovarii proprium were injected in 29.0% of cases. Twenty-one patients had positive nodes and 71% of them showed successful mapping. Among the latter, a negative sentinel node was identified in 3 patients (Sensitivity 78.6% [95% CI: 57-100]; Specificity 100%; Accuracy 97% [95% CI: 94-100]; PPV 100%; NPV 96,6% [95% CI: 93-100]). Twenty-two (12.5%) intra- and 25 (14.2%) postoperative complications occurred (12 G1; 9 G2; 2 G3, 2 G4). Clinical trial information: EUCTR2019-001088-58 .