Fellow, Urologic Oncology
Department of Urology
Weill Cornell/New-York Presbyterian Hospital
Utilization of lymphadenectomy and predictors of lymph node yield in upper tract urothelial carcinoma
Siv Venkat1, Patrick Lewicki1, Spyridon Basourakos1, Douglas Scherr1.
1Department of Urology, Weill Cornell/New-York Presbyterian Hospital, New York, NY, United States
Introduction: Lymph node (LN) dissection (LND) in high-risk upper tract urothelial carcinoma (UTUC) currently remains underperformed compared to other cancers, partly due to a lack of high-quality evidence. We sought to examine current trends in LND for UTUC, as well as patient factors contributing to the performance and omission of LND. We further sought to examine factors predicting higher LN yields.
Methods: A large national hospital-based database (The National Cancer Database) was used to identify all patients from 2004 to 2016 with nonmetastatic UTUC with urothelial histology who underwent extirpative surgery in the form of either radical nephroureterectomy (RNU) or segmental ureterectomy (SU). Rates of LND were examined, with subgroup analysis performed based on high risk features (high grade tumor, tumor over 2cm, pT2-4 disease). Univariate linear regression and multivariate logistic regression was performed to identify clinical variables that predict LN yield.
Results: 30861 patients with UTUC were identified from 2004 to 2016 who underwent extirpative surgery. Overall, 7377 (23.9%) of these patients had a LND performed, and 23484 (76.1%) had no LND (Table 1). Only 27.3% of high grade tumors, 24.0% of tumors over 2cm, 29.4% of pT2 tumors, 32.6% of pT3 tumors, and 44.7% of pT4 tumors had LND performed (Table 2).
LND was more likely to be omitted in older patients, higher Charlson Comorbidity Index (CCI) score, Community Hospitals, lower clinical T and N stage disease, lower pathological T stage disease, low grade tumor, laparoscopic surgery, and with RNU (p<0.001 for all). Patients with LND performed were more likely to have positive margins and LVI present, and to undergo chemotherapy and radiation (p<0.001 for all).
In multivariate analysis, predictors of higher LN yield were time from diagnosis to definitive surgery, Academic/Research Hospital, open surgery, and high grade disease (p<0.05 for all). The strongest predictor was Academic/Research Hospital (OR 3.45, p=0.002). Predictors of lower LN yield were increased age, CCI 1/2, and laparoscopic surgery (Table 3).
Conclusions: LND remains underperformed by surgeons for high risk UTUC tumors (high grade tumor, tumor over 2cm, pT2-4 disease). LND was more likely to be omitted in older patients with increased comorbidities, lower stage disease, low grade tumor, with laparoscopic approaches, with RNU, and at Community Hospitals. Predictors of higher LN yield were time from diagnosis to definitive surgery, Academic/Research Hospital, open surgery, and high grade disease, with the strongest predictor being Academic/Research Hospital setting.