A New Nomogram for Predicting Extrahepatic Metastases in Patients With Hepatocellular Carcinoma： A population-based study of the SEER database and a Chinese single-institutional cohort

Abstract

Abstract Purpose This study aimed to identify risk factors associated with the occurrence of extrahepatic metastases (EHM) in patients with hepatocellular carcinoma (HCC) and to establish an effective predictive nomogram. Methods We extracted eligible data of HCC patients from the Surveillance, Epidemiology, and End Results (SEER) database. This study also included 196 HCC patients from the Zhejiang Cancer Hospital in China. A nomogram for predicting extrahepatic metastases in patients with hepatocellular carcinoma was developed according to the independent variables that were found by univariate and multivariate logistic analysis analyses. The effective performance of the nomogram was evaluated using the areas under the curves (AUC), receiver operating characteristic curve (ROC), and calibration curves. The clinical practicability was evaluated using decision curve analysis (DCA). Results Sex, N stage, histological grade, tumor size, AFP, vascular Invasion (VI), and surgery were all included as independent predictors in a nomogram to predict HCC patients for extrahepatic metastases. In the training cohort, internal validation cohort, and external validation cohort, the AUC of the prediction model were 0.830, 0.834, and 0.831, respectively, while the AUC of the AJCC Stage were 0.692, 0.693, and 0.650. Among patients with extrahepatic metastases, the most common metastasis site was lung (37.38%), followed by bone (36.0%), and lymph nodes (30.6%). Conclusion Based on the SEER database and the Chinese single-institutional cohort, we have developed and validated a nomogram to forecast EHM in HCC patients. The AUC indicated that the nomogram showed adequate accuracy in discriminating EHM. Additionally, the nomogram fared well in the validation cohort and could support clinical decision-making.