Utilization Patterns of Lymph Node Dissection in Endometrial Cancer Patients Without Distant Metastasis in the United States

Abstract

Endometrial cancer is the most common gynecologic cancer in the United States, and patients with early-stage endometrioid adenocarcinoma have a favorable prognosis. Over the past decade, the gynecologic oncology community has debated whether potential harms of systematic lymph node dissection (LND) outweigh potential benefits for these patients. To minimize number of nodes removed, sentinel lymph node dissection (SLND) is under investigation as an alternative. However, ongoing uncertainty of LND/SLND best practices may result in variations in disease management and discrepant outcomes. Methods Three retrospective cohort studies examined LND/SLND use in patients with endometrioid adenocarcinoma. Two examined temporal and geographic variations, respectively, utilizing the Surveillance, Epidemiology, and End Results (SEER) 18 dataset for the years 2004 through 2015. The third used the SEER-Medicare dataset from 2003 through 2016 to quantify and compare the risk of developing 6-month post-surgical lymphedema, lymphocele, hemorrhage, ileus, infection, thrombosis, and all-cause death by number of lymph nodes removed (0, 1-4, 5-9, or 10+). Results Time trend analyses found LND increased from 2004 through 2008, followed by a significant decline through 2015. SLND was rare and did not increase significantly. Significant geographic variation existed for LND use but not SLND. Per 1,000 patients, analyses of 6-month post-surgical complications found 6.5 experienced lymphedema, 3.9 experienced lymphocele, 15.7 experienced hemorrhage, 28.7 experienced ileus, 37.1 experienced infection, 18.6 experienced thrombosis, and 19.8 died. Controlling for size of primary tumor, tumor grade, comorbidities, race/ethnicity, age at diagnosis, adjuvant chemotherapy, and radiotherapy, adjusted hazard ratios (HRs) and 95% confidence intervals (CIs) showed greater risk for ileus (HR: 1.53; 95% CI: 1.24-1.90), infection (HR: 1.52; 95% CI: 1.25-1.83), and thrombosis (HR: 1.41; 95% CI: 1.09-1.82) when comparing removal of 10+ nodes versus 0 nodes. Conclusion Overall, these studies found significant temporal and geographic variation in LND, as well as increasing risk of post-surgical complications associated with increasing numbers of lymph nodes removed. Should continued research into SLND find strong evidence that it effectively detects cancer spread, patients may benefit through decreased risk of post-surgical ileus, infection, and thrombosis.