Risk factors for lymph node metastasis in apparent early-stage epithelial ovarian cancer : implications for surgical staging

Objectives: The extent of lymphadenectomy to be performed in apparent early-stage epithelial ovarian cancer (EOC) is not well defined. We evaluated the patterns of lymphatic spread in apparent early-stage EOC and risk factors for lymph node metastasis, as these have potential implications for clinical decision making. Methods: All cases of apparent early-stage EOC diagnosed at our institution between January 1994 and December 2003 were retrospectively identified. Apparent early-stage EOC was defined as gross disease that appeared confined to the pelvis without abdominal spread at the time of initial exploration. Demographics, pathologic findings, staging procedures performed, and clinical impression at surgery were analyzed. Patterns of lymph node positivity and risk factors associated with upstaging were assessed. Results: One hundred and ninety patients with apparent early-stage EOC undergoing primary surgical staging met criteria for inclusion. All patients had at least some pathologic assessment of lymph nodes, with 115 having both bilateral pelvic and paraaortic lymphadenectomy performed. After review of pathology and operative reports, the final FIGO staging within the cohort was 54 IA (28.4%), 10 IB (5.3%), 51 IC (26.8%), 1 IIA (0.5%), 4 IIB (2.1%), 37 IIC (19.5%), 8 IIIA (4.2%), 25 IIIC (13.2%). Overall 25/190 (13%) had lymph nodes metastasis as follows: 8 (32%) had positive pelvic nodes, 12 (48%) had positive paraaortic nodes, and 5 (20%) had both positive pelvic and paraaortic lymph nodes. Significant risk factors for lymph node metastasis included bilateral vs. unilateral primary lesion (26.8% vs. 7.5%, p < 0.001), positive cytologic washings vs. negative (22.4% vs. 9.1%, p = 0.012), ascites vs. no ascites (28.2% vs. 9.3%, p = 0.002), serous vs. other histology (28% vs. 9%, p = 0.001), grade 1 vs. grade 2 vs. grade 3 disease (2.7% vs. 1.9% vs. 23.2%, p < 0.001), and preoperative CA 125 levels of > 35 vs. ≤ 35 U/ml (22.4% vs.0% p = 0.006). No patients with mucinous cancers (n = 29) had lymph node metastases. Patterns of LN metastases were largely independent of laterality of primary lesions: among those with unilateral lesions and positive nodes (n = 10), 5 (50%) had ipsilateral lymph node involvement, 4 (40%) had bilateral involvement, and 1 (10%) had isolated contralateral lymph nodes positive. Conclusions: Complete surgical staging in EOC patients with gross disease confined to the ovaries and pelvis should include bilateral pelvic and paraaortic lymphadenectomy. Even in patients with unilateral lesions, lymph node metastases are commonly bilateral. Bilateral ovarian lesions, positive cytology, presence of ascites, high grade histology, and serous histology are risk factors for lymph node involvement. This information may be helpful in counseling patients presenting for consideration of re-staging after unexpected findings of malignancy.