Significance of lymph node ratio on survival of women with borderline ovarian tumors.

Abstract

To assess the qualitative and quantitative measures of the effect of pelvic lymph node involvement on survival of women with borderline ovarian tumors (BOTs). This is a retrospective study examining the Surveillance, Epidemiology, and End Results Program between 1988 and 2003. Women with stage T1-3 BOTs who had results of pelvic lymph node status at surgery were included. The effect of lymph node involvement on cause-specific survival (CSS) was evaluated using multivariable analysis with the following approaches: (1) any involvement, (2) involvement of multiple nodes (≥ 2 nodes), and (3) lymph node ratio (LNR), defined as the ratio of the number of tumor-containing lymph nodes to the total number of harvested lymph nodes. A total of 1524 women were examined for analysis. Median count of sampled nodes was 8 (interquartile range 3-15), and there were 81 (5.3%, 95% confidence interval [CI] 4.2-6.4) women who had lymph node involvement. Median follow-up was 15.8 (interquartile range 13.8-18.9) years, and 83 (5.4%) women died of BOTs. After controlling for age, histology, stage, and tumor size, only LNR remained an independent prognostic factor for decreased CSS (adjusted hazard ratio [HR] per percentage unit 1.015, 95% CI 1.003-1.026, P = 0.014), whereas any involvement (adjusted HR 1.700, 95% CI 0.843-3.430, P = 0.138) and involvement of multiple nodes (adjusted HR 1.644, 95% CI 0.707-3.823, P = 0.249) did not. On cutoff analysis, LNR ≥ 13% had the largest magnitude of significance on multivariable analysis of CSS (adjusted HR 2.399, 95% CI 1.163-4.947, P = 0.018). Our study suggests that high pelvic LNR may be a prognostic factor associated with decreased CSS in women with BOTs.