Impact of number of positive lymph nodes on prognostic stratification in renal cell carcinoma: Analysis of the National Cancer Database.

Abstract

736 Background: Lymph node positivity in Renal Cell Carcinoma (RCC) is associated with worsened oncologic outcomes. However, the actual prognostic significance of node positivity is poorly understood. Currently, American Joint Committee on Cancer (AJCC) Stage III RCC includes both node-positive pN1 and node-negative pN0 disease. We hypothesize that (1) there is a threshold in number of pathologic node positivity that distinguishes favorable risk from poor risk nodal disease, and (2) current categorization of pN1 can be subdivided into pN1 and pN2 based this threshold. We tested our hypothesis using the National Cancer Database (NCDB). Methods: From 2004-2019, all cases of RCC were queried in patients age ≥18. Patients with pathologic node positive disease and without synchronous metastasis were selected for analysis to minimize confounding from metastatic burden. Multivariable Cox Proportional-Hazards regression tested association between number of pathologically positive lymph nodes and all-cause mortality (ACM), adjusting for clinical and pathologic co-variables. Receiver Operator Characteristic (ROC) Curve analyses employing the concordance probability method evaluated performance of potential cut-points for pN2 node-positivity. Kaplan-Meier analyses (KMA) compared these thresholds against overall survival (OS) in non-metastatic Stage IV RCC. Results: 28,590 patients with above criteria were identified, of which 13.6% had pN1. On multivariable analyses, increased pathologic node positivity was associated with increased hazard of ACM (HR 1.19, 95% Confidence Interval [CI] 1.17-1.20, p<0.001). ROC mapping of all possible lymph node thresholds from ≥2 to ≥10, with stage IV as the highest point, showed comparable concordance probability among these cutoffs 0.26-0.33, AUC=0.656. On KMA, when threshold was set at ≥3, 5-year OS was no longer significantly different from non-metastatic Stage IV RCC as illustrated by overlapping confidence intervals. We designated pN1 as 1-2 pathologic positive nodes, and pN2 as ≥3 pathologic positive nodes. 5-year OS for Stage III pN0 was 69.4% (95% CI 68.4-70.5), for Stage III pN1 was 41.4% (95% CI 39.0-43.8), for Stage III pN2 was 31.8% (95% CI 28.2-35.9%), and for non-metastatic Stage IV was (30.0%, 95% CI 28.2-32.0%). On multivariable analyses, pN2 exhibited 38% greater hazard of ACM (HR 3.31, 95% CI 2.24-2.54, p<0.001) compared to pN1 (HR 2.39, 95% CI 2.24-2.54, p<0.001). Conclusions: These findings represent one of the largest characterizations of impact of positive nodal counts on prognostic stratification in RCC. Pathologic node positivity could be stratified to pN1 and pN2, with pN2 conferring poor prognostic risk comparable to non-metastatic Stage IV pN0 disease. Although further validation studies are warranted, consideration should be given towards stratifying Stage III pN2 patients to a higher risk group.