Adam C Calaway1, Sean Q Kern2, David Crook2, Yan Tong3, Timothy A Masterson2, Nabil Adra4, Lawrence H Einhorn4, Richard S Foster2, Clint Cary2. 1. University Hospitals/Case Western Reserve University, Cleveland, Ohio. 2. Department of Urology, School of Medicine, Indiana University,Indianapolis, Indiana. 3. Department of Biostatistics, Indiana University,Indianapolis, Indiana. 4. Melvin and Bren Simon Comprehensive Cancer Center, Indiana University,Indianapolis, Indiana.
Abstract
PURPOSE: Presence of teratoma in the orchiectomy and residual retroperitoneal mass size are known predictors of finding teratoma during postchemotherapy retroperitoneal lymph node dissection (PC-RPLND). We sought to determine if the percentage of teratoma in the orchiectomy specimen could better stratify the risk of teratoma in the retroperitoneum. MATERIALS AND METHODS: The Indiana University Testis Cancer Database was reviewed to identify patients who underwent PC-RPLND for nonseminomatous germ cell tumors from 2010 to 2018. A logistic regression model was fit to predict the presence of retroperitoneal teratoma using teratoma and yolk sac tumor in the orchiectomy, residual mass size and log transformed values of prechemotherapy alpha-fetoprotein (AFP) and beta-human chorionic gonadotropin. The study cohort was split into 60% training and 40% validation sets using 200 bootstraps. A predictive nomogram was developed for predicting teratoma in the retroperitoneum. RESULTS: A total of 422 men were included. Presence of teratoma in the orchiectomy (OR 1.02, p <0.001), residual mass size (OR 1.16, p <0.001) and log transformed prechemotherapy AFP (OR 1.12, p=0.002) were predictive factors for having teratoma in the retroperitoneum. The C-statistic using this model demonstrated a predictive ability of 0.77. Training set C-statistic was 0.78 compared to 0.75 for the validation set. A nomogram was developed to aid in clinical utility. CONCLUSIONS: The model better predicts patients at higher risk for teratoma in the retroperitoneum following chemotherapy, which can aid in a more informed referral for surgical resection.
PURPOSE: Presence of teratoma in the orchiectomy and residual retroperitoneal mass size are known predictors of finding teratoma during postchemotherapy retroperitoneal lymph node dissection (PC-RPLND). We sought to determine if the percentage of teratoma in the orchiectomy specimen could better stratify the risk of teratoma in the retroperitoneum. MATERIALS AND METHODS: The Indiana University Testis Cancer Database was reviewed to identify patients who underwent PC-RPLND for nonseminomatous germ cell tumors from 2010 to 2018. A logistic regression model was fit to predict the presence of retroperitoneal teratoma using teratoma and yolk sac tumor in the orchiectomy, residual mass size and log transformed values of prechemotherapy alpha-fetoprotein (AFP) and beta-human chorionic gonadotropin. The study cohort was split into 60% training and 40% validation sets using 200 bootstraps. A predictive nomogram was developed for predicting teratoma in the retroperitoneum. RESULTS: A total of 422 men were included. Presence of teratoma in the orchiectomy (OR 1.02, p <0.001), residual mass size (OR 1.16, p <0.001) and log transformed prechemotherapy AFP (OR 1.12, p=0.002) were predictive factors for having teratoma in the retroperitoneum. The C-statistic using this model demonstrated a predictive ability of 0.77. Training set C-statistic was 0.78 compared to 0.75 for the validation set. A nomogram was developed to aid in clinical utility. CONCLUSIONS: The model better predicts patients at higher risk for teratoma in the retroperitoneum following chemotherapy, which can aid in a more informed referral for surgical resection.