The estimated prevalence of missed positive lymph nodes based on extent of lymphadenectomy at radical prostatectomy

Purpose: To determine practice patterns for the extent of lymphadenectomy at radical prostatectomy and associations with detection of pN1 prostate cancer, as well as the impact of lymphadenectomy extent on underdetection of pN1 disease and overall survival. Materials and methods: Prostatectomy cases in the NCDB from 2004 to 2013 were included. Lymphadenectomy extent was defined by the number of nodes examined. Logistic regression was used to identify risk factors for the top quartile of lymph node count and pN1 disease. This model was created to estimate the expected prevalence of pN1, and generated observed over expected ratios. A Cox regression model was used to evaluate the effect of lymph node count on overall survival. Results: Lymphadenectomy was performed in 209,789 (60%) of 358,522 surgeries, with pN1 in 6,428 (3.08%). Increasing quartiles for lymph node count was associated with pN1 (3–5 nodes OR 2.11; 6–8 nodes OR 3.12; ≥9 nodes OR 5.91, all P< 0.001). The logistic regression model suggested that 59% of pN1 cases are missed due to low lymph node count. Increased lymph node count was associated with increasing pN1 detection (O/E: 1–2 nodes = 0.18; 3–5 nodes = 0.37; 6–8 nodes = 0.56; ≥9 nodes = 1.01). Cox proportional hazards modeling demonstrated that the top quartile for lymph node count had improved overall survival (HR 0.93, CI 0.87–0.99, P= 0.03). Conclusions: Increasing lymphadenectomy extent was associated with pN1 disease on multivariate analysis, and logistic regression modeling suggested a substantial proportion of pN1 were missed due to low lymphadenectomy extent across all risk groups.