Multicenter validation study of the sentinel lymph node concept in cervical cancer:: AGO study group

Purpose Single-institution case series have demonstrated the feasibility of the sentinel concept in cervical cancer. However, the diagnostic accuracy remains to be validated. We evaluated detection rate and diagnostic accuracy to predict the histopathologic pelvic nodal status in patients with cervical cancer of all stages. Patients and Methods In a hypothesis-based, prospective, multicenter cohort study, patients underwent lymph node detection after labeling with technetium, patent blue, or both. After systematic pelvic and, if indicated, para-aortic node dissection, all lymph nodes were histopathologically examined. Detection rate, sensitivity, and negative predictive value (NPV) were calculated. Results According to the protocol, 590 patients were eligible. Detection rate of pelvic sentinel nodes was 88.6% (95% CI, 85.8% to 91.1%) and was significantly higher for the combination of technetium and patent blue (93.5%; 95% CI, 90.3% to 96.0%). Of 106 patients with pelvic lymph node metastases, 82 had pelvic sentinel node metastases. The overall sensitivity was 77.4% (95% CI, 68.2% to 85.0%), which was lower than 90%, the predefined noninferiority margin (P < .001). Sensitivity in women with tumors <= 20 mm (90.9%), with bilateral detection (87.2%), or with both substances applied (80.3%) was higher compared with the total population. The overall NPV was 94.3% (95% CI, 91.6% to 96.4%) and was higher in patients with tumors <= 20 mm (99.1%; 95% CI, 96.6% to 100%) compared with patients with tumors more than 20 mm (88.5%; 95% CI, 82.9% to 92.8%; P < .001). Conclusion In our cohort (all stages), sensitivity of the sentinel concept was low. However, patients with tumor diameter <= 20 mm may profit from this concept.