Disparities in the surgical staging of high-grade endometrial cancer in the United States

Jonathan R. Foote, Stephanie Gaillard, Gloria Broadwater, Julie A. Sosa, Brittany Davidson, Mohamed A. Adam, Angeles Alvarez Secord, Monica B. Jones, Junzo Chino, Laura J. Havrilesky

The National Comprehensive Cancer Network (NCCN) and the Society of Gynecologic Oncology (SGO) recommend lymph node sampling (LNS) as a key component in the surgical staging of high-grade endometrial cancer. Our goal was to examine surgical staging patterns for high-grade endometrial cancer in the United States. The National Cancer Data Base (NCDB) was searched for patients who underwent surgery for serous, clear cell, or grade 3 endometrioid endometrial cancer. Outcomes were receipt of LNS and overall survival (OS). Multivariate logistic regression was used to examine receipt of LNS in Stage I-III disease based on race (White vs. Black), income, surgical volume, and distance traveled to care. Multivariate Cox proportional hazards regression modeling was used to assess OS based on stage, race, income, LNS, surgical volume, and distance traveled. Forty-two thousand nine hundred seventy-three patients were identified: 76% White, 53% insured by Medicare/Medicaid, 24% traveled >30 miles, and 33% stage III disease. LNS was similar among White and Black women (81% vs 82%). LNS was more common among >30 miles traveled (84% vs 81%, p < 0.001), higher surgical volume (83% vs 80%, p < 0.001), and academic centers (84% vs 80%, p < 0.001). In multivariate analysis, higher income, higher surgical volume, Charlson-Deyo score, and distance traveled were predictors of LNS. Stage III disease (HR 3.39, 95% CI 3.28-3.50), age (10-year increase; HR 1.63, 95% CI 1.61-1.66), lack of LNS (HR 1.64, 95% CI 1.56-1.69), and low income (HR 1.20, 95% CI 1.14-1.27) were predictors of lower survival. Surgical care for high-grade endometrial cancer in the United States is not uniform. Improved access to high quality care at high volume centers is needed to improve rates of recommended LNS.