不典型强化原发性肝癌的MRI影像学特征

MRI features of atypical enhanced primary hepatocellular carcinoma

目的分析不典型强化的原发性肝细胞癌(HCC)的MRI影像学特征,提高不典型强化HCC的诊出率。方法回顾性纳入2019~2021年于深圳市人民医院行MRI检查并经病理证实的142例HCC患者,按照MRI诊断方式将患者分为两组:符合“快进快出”为典型强化组(n=97),不符合“快进快出”为不典型强化组(n=45)。由2位影像医师分别阅片并记录肿瘤的强化方式、瘤体直径、非边缘动脉期强化、强化包膜、非边缘强化洗褪、晕状强化、瘤内脂肪或出血、瘤内结节、马赛克征、肿瘤的对血管的浸润、动脉期病灶信号强度、门脉期病灶信号强度、表观弥散系数值及门脉期强化率。对单因素分析有统计学意义的资料纳入多因素Logistic回归分析,建立不典型强化HCC的诊断模型。绘制ROC曲线,计算曲线下面积,评估判定模型的敏感度与特异性。结果两组间非边缘动脉期强化、强化包膜、肿块内脂肪、出血、结中结、马赛克征、血管浸润、肿瘤直径及表观弥散系数的差异无统计学意义(P>0.05),而非边缘强化洗褪(P=0.005)、晕状强化(P=0.005)以及门脉期强化率(P=0.001)的差异有统计学意义。将上述3种差异有统计学意义的变量纳入多因素Logistic回归分析显示:缺乏非边缘强化洗褪征象(OR=27.995,95%CI:3.910~200.462,P=0.001)和门脉期强化率(OR=1.034,95%CI:1.018~1.051,P<0.001)是HCC发生不典型强化的独立影响因素。进一步基于非边缘强化洗褪和门脉期强化率,建立联合判定模型,ROC曲线分析显示非边缘强化洗褪以及门脉期强化率联合观察对不典型强化HCC具有较好的判定效能(曲线下面积为0.874)。结论晕状强化、门脉期强化率及缺乏非边缘强化洗褪与HCC发生不典型强化具有相关性,而门脉期强化率和缺乏非边缘强化洗褪特征则是HCC发生不典型强化的独立影响因素,两者联合有助于判定不典型HCC。

Objective To analyze the MRI features and improve the diagnostic rate of atypical enhanced primary hepatocellular carcinoma(HCC).Methods A total of 142 patients with pathologically confirmed HCC at Shenzhen People's Hospital from 2019 to 2021,who concurrently received preoperative contrast-enhanced MRI examination,were included in this retrospective study.Patients were divided into two groups based on MRI features:Typical enhancement group(n=97,those who met the"fast in and fast out"criteria)and atypical enhancement group(n=45,those who did not meet the"fast in and fast out"criteria).A series of imaging parameters,including enhancement pattern,tumor diameter,non-rim arterial phase enhancement,enhancement capsule,non-rim washout,halo enhancement,intratumoral fat or bleeding,intratumoral nodules,mosaic sign,tumor infiltration into blood vessels,lesion signal intensity on arterial phase,lesion signal intensity on portal phase,apparent diffusion coefficient,and portal phase enhancement rate,were independently assessed by two expert radiologists.Variables with statistical significance in the univariate analysis were further incorporated in the multivariate Logistic regression model to establish a prediction model for identifying atypical enhanced HCC.The ROC curve was used to evaluate the sensitivity and specificity of the prediction model.Results There was no significant difference between the two groups in non-rim arterial phase enhancement,enhanced capsule,intratumoral fat or bleeding,intratumoral nodules,mosaic sign,vascular infiltration,tumor diameter and apparent diffusion coefficient value(P>0.05),while there were statistical differences in non-rim washout(P=0.005),halo enhanced(P=0.005)and portal phase enhanced rate(P=0.001).Multivariate logistic regression analysis showed that the absence of non-rim washout(OR=27.995,95%CI:3.910-200.462,P=0.001)and portal phase enhancement rate(OR=1.034,95%CI:1.018-1.051,P<0.001)were independent influencing factors for atypical enhancement in HCC.A combined prediction model was established based on the non-rim washout and portal phase enhancement rate.ROC analysis showed that the prediction ability of the combined observation(non-rim washout plus portal phase enhancement rate)for atypical enhanced HCC was overall satisfying with the AUC of 0.874.Conclusion Non-rim washout enhancement and portal venous phase enhancement rate are independent influencing factors for atypical enhancement HCC,and the combination of these two parameters can improve the diagnostic efficacy of HCC.