Development of a novel nomogram incorporating platelet-to-lymphocyte ratio for the prediction of lymph node involvement in prostate carcinoma

Abstract

Objective: The available nomograms used to predict lymph node involvement (LNI) are not comprehensive. We sought to derive a novel nomogram incorporating the platelet to lymphocyte ratio (PLR) to predict LNI and compare its performance to validated preoperative risk nomograms in a cohort of men undergoing robotic-assisted radical prostatectomy at our institution. Methods: Our electronic health record was queried for patients who underwent robotic-assisted radical prostatectomy with bilateral pelvic lymphadenectomy between 2013 and 2019. A bootstrapped multivariate logistic regression model was constructed for the predictors of LNI while adjusting for other covariates. Then, we used the derived logistic regression formula to estimate each patient's risk (%) for LNI. Individualized risks were also calculated using the following verified nomograms: Briganti-2012, Cagiannos, Godoy, and Memorial Sloan Kettering Cancer Center. Subsequently, we plotted the risks for our nomogram and the 4 verified nomograms into receiver operating characteristics curves. We reported the area under the curve (AUC) for each of the 5 nomograms and the corresponding 95% confidence interval (CI). Results: The cohort included 173 patients, of which 13.9% demonstrated LNI. LNI was associated with higher preoperative prostate-specific antigen (PSA) ≥ 10 [odds ratio [OR] = 4.89; 95% confidence interval [CI] (1.42–16.83)], higher grade (WHO group ≥ 3)[19.21; (2.23–195.25)], and higher percentage of positive biopsy cores (≥60%) [3.38, (1.04–11.00)]. With every 30-unit increase in PLR the risk of LNI increased by 47%. The nomogram derived from our data had the highest AUC [(AUC 0.877; 95% CI (0.806–0.947)]. The Memorial Sloan Kettering Cancer Center and Briganti 2012 displayed almost congruent ability [0.836; 95% CI (0.758–0.915)] and [0.827; (0.752–0.902)] to identify patients with positive nodes in our cohort with perfect sensitivity and negative predictive value. Conclusion: The nomogram incorporating PLR demonstrated 94.7% sensitivity to predict LNI and avoided pelvic lymphadenectomy in half of the patients at a cut-off between 6.5% and 8.5%. A prospective study with a larger sample is needed to validate our findings. © 2020 Elsevier Inc.