Characteristics of lung metastasis in testicular cancer: A large-scale population analysis based on propensity score matching.

Abstract

This study aims to systematically evaluate predictive factors for lung metastasis (LM) in patients with testicular cancer (TC) and to investigate cancer-specific survival (CSS) and overall survival (OS) of LM in TC patients based on a large population-cohort. A total of 10,414 patients diagnosed with TC during 2010-2015 were adopted from the Surveillance, Epidemiology, and End Results (SEER). After propensity score matching (PSM), 493 patients with LM were included for subsequent analysis. Univariate and multivariate logistic regression analyses were employed to identify risk factors, a nomogram was developed, and the receiver operating characteristic (ROC) curve was utilized to confirm the validation of the nomogram. Prognostic factors for OS and CSS among TC patients with LM were estimated via Cox proportional hazards models. Postmatching indicated that 11 parameters were successfully balanced between both groups (P > 0.05). After PSM, TC patients with LM presented an undesirable prognosis in both CSS and OS than those without LM (P < 0.001). The logistic regression model showed that tumor size; T stage; N stage; liver, brain, and bone metastases; and histology were positively associated with LM (P < 0.05). A nomogram was developed to predict diagnostic possibilities based on the independent risk variables, and the ROC curve verified the predictive capacity of the logistic regression model [area under the curve (AUC) = 0.910]. The selected variates in the nomogram can be predictive criteria for TC patients with LM. Brain metastasis, liver metastasis, and larger tumor size were prognostic factors for CCS and OS among TC patients with LM.