目的： 分析BCLC分期A期原发性肝细胞肝癌破裂（rupture of hepatocellular carcinoma,rHCC）的危险因素及建立预测破裂的模型。方法： 回顾性分析2002年3月至2019年11月我院25例BCLC分期A期的rHCC临床资料,以及对照组25例未破裂的HCC病人。通过单因素及多因素Logistic回归分析,研究rHCC的独立危险因素,建立预测模型。结果： 多因素回归分析证实,与BCLC分期A期rHCC相关的独立危险因素为肿瘤凸出肝表面（OR=38.333,P<0.01）。成功得出预测rHCC的回归模型,绘制受试者工作特征（receiver operating characteristic,ROC）曲线。曲线下面积为0.820,约登指数为0.640,预测灵敏度为0.720,特异度为0.920。结论： 肿瘤凸出肝包膜表面是BCLC分期A期rHCC的独立危险因素。Logistic回归模型能有效预测该期HCC发生破裂的风险。

Objective To explore the risk factor and prediction of rupture of Barcelona Clinic Liver Cancer (BCLC) stage A primary hepatocellular carcinoma (HCC). Methods We analyzed retrospectively the clinical data of 25 patients with spontaneous rupture of BCLC stage A HCC between March 2002 and November 2019 in our hospital with 25 patients of HCC without rupture of HCC (rHCC) by random as control. Independent risk factor of rHCC was studied by both univariate and multivariate Logistic regression. Logistic regression model to predict rHCC was studied. Results Multivariate regression analysis identified tumor protrusion from liver surface as independent risk factor associated with BCLC stage A rHCC (OR=38.333, P<0.01). Logistic regression model to predict rHCC was established successfully. Receiver operating characteristic (ROC) curves was set up. The area under the ROC curve was 0.820 with Youden index 0.640, sensitivity 0.720 and specificity 0.920. Conclusions It was shown tumor protrusion from liver surface as an independent risk factor of BCLC stage A rHCC. Logistic regression model could be used to predict effectively the risk of rupture in patients with BCLC stage A HCC.