917 NOVEL PREDICTORS OF BENIGN PATHOLOGY IN STAGE IA OR IB PATIENTS WITH NON-SEMINOMATOUS TESTICULAR GERM CELL MALIGNANCY UNDERGOING PRIMARY RETROPERITONEAL LYMPH NODE DISSECTION

Abstract

INTRODUCTION AND OBJECTIVES: There is no consensus on the optimal treatment for Stage I testicular non seminomatous germ cell tumors (T NSGCT), partly due to the inaccuracy of clinical staging methods. Prior reports have demonstrated that the presence of embyronal carcinoma (EC), lymphovascular invasion (LVI), and absence of yolk sac tumor (YST) in the orchiectomy specimen are important risk factors predicting occult metastatic disease. We assessed whether a novel combination of the above factors could predict the presence of benign retroperitoneal pathology. METHODS: We reviewed an institutional database of patients with T NSGCT and included all patients with Stage IA or IB NSGCT who underwent primary RPLND. Logistic regression analysis was used to compare patients with benign pathology versus germ cell tumor (GCT). A combination of significant pathologic factors was used as internal validation to predict the presence of benign or GCT pathology. Significance was determined at p 0.05. No financial funding was obtained. RESULTS: Among 55 patients with Stage IA or IB NSGCT managed with primary RPLND, 44(80%) had benign pathology and 11 (20%) had occult GCT in the retroperitoneum. No specimens revealed retroperitoneal teratoma in this cohort. Predictors of benign pathology were higher preorchiectomy alpha fetoprotein levels (p 0.03), larger percentage of YST (p 0.02), lower embryonal carcinoma to YST (E/Y) ratio (p 0.001), presence of teratoma and absence of LVI in the primary orchiectomy specimen (both p 0.01). Multivariate analysis demonstrated an E/Y ratio 4 to be a strong predictor of benign pathology in the retroperitoneum, with an area under the receiver operating curve of 0.71 0.08 (p 0.04), and a sensitivity and specificity of 37% (16/43) and 100% (12/12), respectively. In combination with absence of LVI, an E/Y ratio 4 had a 100%(10/10) positive predictive value for benign pathology on internal validation. Conversely, presence of LVI and an E/Y ratio of 4 revealed a 64% (7/11) predictive value for GCT. CONCLUSIONS: In our cohort, patients with Stage IA or IB NSGCT who had benign pathology on the primary RPLND had a significantly higher proportion of YST to EC in the preorchiectomy specimen. An E/Y cutoff value of 4 appears to be discriminatory for predicting benign pathology. Emphasis on the proportion of YST in the primary orchiectomy specimen in addition to other previously established risk factors may allow for improved risk stratification.