Julian Chavarriaga1, Diego Camacho2, Daniel Suso-Palau2, Fabián Godoy2, Marino Cabrera2, Jorge Forero2, Byron López-de-Mesa2, Rodolfo Varela2. 1. Department of Urologic Oncology, Instituto Nacional de Cancerología. Bogotá, Colombia; Division of Urology, Clínica Imbanaco -Grupo Quirón Salud. Cali, Colombia; Division of Urology, Pontificia Universidad Javeriana. Bogotá, Colombia. Electronic address: chavarriagaj@javeriana.edu.co. 2. Department of Urologic Oncology, Instituto Nacional de Cancerología. Bogotá, Colombia.
Abstract
INTRODUCTION: Penile cancer (PC) is an aggressive malignancy in which the most important prognostic factor for cancer specific survival (CSS) is the involvement of regional lymph nodes (LNs). Lymph node density (LND) could become a superior prognostic tool for CSS, by accounting for both extent of dissection and nodal disease burden. We aim to validate LND as a prognostic factor for CSS in a contemporary series of patients with PC treated and followed at a single high-volume center, treating more than 25 PC patients per year, over a 13-year period. METHODS: Clinical charts of all patients with PC who underwent surgical treatment between 2007 and 2020 were reviewed. Clinicopathological data was collected and analyzed retrospectively. We only included patients with ≥ 8 LNs removed in a unilateral ILND or ≥16 LNs when a bilateral approach was used. We attempted to find an optimal threshold for LND, capable of maximizing effect difference in terms of CSS and RFS between dichotomized groups. To determine this threshold, we used the chi-squared and the Mann-Whitney tests, and it was required to fulfill the proportional hazards assumption. We assessed different thresholds previously reported in the literature. In our study the optimal threshold for LND was determined to be ≤ 20% Descriptive statistics were used to summarize patient characteristics, CSS and RFS were graphically represented by Kaplan-Meier estimates. Harrell's C index for CSS and RFS were calculated for LND and pN stage, to determine which variable has a superior predictive capacity RESULTS: We identified 110 patients with PC who underwent ILND at our institution, of these, 87 were node-positive and were included in the final analysis. Overall estimates of CSS showed a 3-year CSS of 43% (95% CI: 32-54), the estimated 3-year CSS for the patients with a LND ≤ 20% was 69% (95% CI: 50-82) and 26% (95% CI: 14-39) in the group with a LND >20% (Log-rank P = 0.001). The estimated 3-year RFS for the patients with LND ≤ 20% was 61% (95% CI: 42-76) and 30% (95% CI: 16-44) in the group with a LND >20% (Log-rank P = 0.009). The results of univariate analysis indicate that in patients with a LND >20% the risk for cancer specific mortality was increased (HR 2.68; 95% CI: 1.45-4.98, P = 0.002) compared with LND ≤ 20%. In the and Cox multivariate analysis after Adjusting for age and pN stage the association increased (HR 2.73; 95%, CI 1.38-5.40, P = 0.004). Harrell´s C index for CSS was 0.63 for LND vs. 0.54 for pN stage, suggesting a 9% higher concordance for LND and CSS. CONCLUSIONS: Lymph node density stands as a promising tool for risk-stratifying patients with node-positive PC after ILND. In this retrospective study, LND was a significant predictor of CSS and RFS when using a LND >20% threshold, and also showed a superior predictive ability than pN stage. These results support the use of the LND parameter in clinical practice with a final goal to improve risk stratification, and individualized adjuvant treatment decision-making to patients with high-risk of cancer specific mortality.
INTRODUCTION: Penile cancer (PC) is an aggressive malignancy in which the most important prognostic factor for cancer specific survival (CSS) is the involvement of regional lymph nodes (LNs). Lymph node density (LND) could become a superior prognostic tool for CSS, by accounting for both extent of dissection and nodal disease burden. We aim to validate LND as a prognostic factor for CSS in a contemporary series of patients with PC treated and followed at a single high-volume center, treating more than 25 PC patients per year, over a 13-year period. METHODS: Clinical charts of all patients with PC who underwent surgical treatment between 2007 and 2020 were reviewed. Clinicopathological data was collected and analyzed retrospectively. We only included patients with ≥ 8 LNs removed in a unilateral ILND or ≥16 LNs when a bilateral approach was used. We attempted to find an optimal threshold for LND, capable of maximizing effect difference in terms of CSS and RFS between dichotomized groups. To determine this threshold, we used the chi-squared and the Mann-Whitney tests, and it was required to fulfill the proportional hazards assumption. We assessed different thresholds previously reported in the literature. In our study the optimal threshold for LND was determined to be ≤ 20% Descriptive statistics were used to summarize patient characteristics, CSS and RFS were graphically represented by Kaplan-Meier estimates. Harrell's C index for CSS and RFS were calculated for LND and pN stage, to determine which variable has a superior predictive capacity RESULTS: We identified 110 patients with PC who underwent ILND at our institution, of these, 87 were node-positive and were included in the final analysis. Overall estimates of CSS showed a 3-year CSS of 43% (95% CI: 32-54), the estimated 3-year CSS for the patients with a LND ≤ 20% was 69% (95% CI: 50-82) and 26% (95% CI: 14-39) in the group with a LND >20% (Log-rank P = 0.001). The estimated 3-year RFS for the patients with LND ≤ 20% was 61% (95% CI: 42-76) and 30% (95% CI: 16-44) in the group with a LND >20% (Log-rank P = 0.009). The results of univariate analysis indicate that in patients with a LND >20% the risk for cancer specific mortality was increased (HR 2.68; 95% CI: 1.45-4.98, P = 0.002) compared with LND ≤ 20%. In the and Cox multivariate analysis after Adjusting for age and pN stage the association increased (HR 2.73; 95%, CI 1.38-5.40, P = 0.004). Harrell´s C index for CSS was 0.63 for LND vs. 0.54 for pN stage, suggesting a 9% higher concordance for LND and CSS. CONCLUSIONS: Lymph node density stands as a promising tool for risk-stratifying patients with node-positive PC after ILND. In this retrospective study, LND was a significant predictor of CSS and RFS when using a LND >20% threshold, and also showed a superior predictive ability than pN stage. These results support the use of the LND parameter in clinical practice with a final goal to improve risk stratification, and individualized adjuvant treatment decision-making to patients with high-risk of cancer specific mortality.