Prognostic Impact of Lymphadenectomy in Different Stages of Malignant Germ Cell Tumor of the Ovary Based on Propensity Score Matching

Background: Lymphadenectomy has been widely used in the treatment of malignant germ cell tumor of the ovary (OGCT), which is a kind of ovarian cancers occurred mostly in young women and adolescent girls. But the clinical decision mainly depends on the doctor’s experience without a well-defined guideline. This population-based study aimed to evaluate the prognostic impact of lymphadenectomy in different stages of malignant germ cell tumors of the ovary. Methods: Patients with known status of lymphadenectomy in different stages of OGCT were explored from the Surveillance, Epidemiology, and End Results (SEER) program database from 1973 to 2013. We used propensity score matching algorithm to reduce the selection bias between the two study groups. Survival curves, univariate and multivariate Cox proportional hazards model were applied to evaluate the prognostic impact of lymphadenectomy in different stages of OGCT. Results: We included 1,996 OGCT patients in the study, and 818 (41%) of them had lymph node resection. Compared to the LND- group, patients with lymph node resection tended to be at stage II and III, had larger tumor sizes and diagnosed as dysgerminoma. The influence of diagnosis ages, marital status and tumor grades were significantly decreased by applying the propensity score matching. Lymphadenectomy-positive (LND+) group demonstrated significantly worse survival than the lymphadenectomy-negative (LND-) group in later stages (stage III, overall, P=0.027, cancerspecific, P=0.006; stage IV, overall, P=0.034, cancer-specific, P=0.037). While, both the overall and cancer-specific survival showed no significant differences between LND+ and LND- in stage I (overall, P=0.411, cancer-specific, P=0.876) and stage II (overall, P=12, cancer-specific, P=0.061). Univariate (overall, HR=1.497, CI=1.010-2.217, P=0.044; cancer-specific, HR=1.524, CI=1.067- 2.404, P=0.050) and multivariate (overall, HR=1.580, CI=1.046-2.387, P=0.030; cancer-specific, HR=1.661, CI=1.027-2.686, P=0.039) Cox proportional model both verified the association between the lymph node resection and better survival in the whole cohort. Conclusion: Lymphadenectomy significantly increased the survival probability of OGCT patients in stage III and IV, but had no significant influence on early-stage patients (stage I and II), indicating lymphadenectomy should be performed in a stage-dependent manner in clinical utility.