中文摘要：

目的研究活体肝脏氢质子MRS（^1H—MRS）评价肝脏局灶性病变的价值。方法对53例肝肿瘤患者（54个直径〉4cm的肿瘤）和19名正常志愿者进行^1H-MRS检查，分别测量满足诊断需要的正常对照组（17名）和经病理证实的良性肝脏肿瘤组（8例）和肝细胞癌（HCC）组（25例）胆碱峰／甘油三酯亚甲基峰（Cho／Lip）比值，对结果进行方差分析和两样本间Dunnett—t检验，并通过受试者工作特征（ROC）曲线评价^1H-MRS诊断HCC的敏感性和特异性。结果正常肝脏、良性肿瘤及HCC组的Cho／Lip比值分别为0．07±0．04、0．11±0．06和0．55±0．17（F=6．58，P〈0．05），HCC组高于正常对照组和良性肝脏肿瘤组（t值分别为2．99和2．32，P值均〈0．05），正常肝脏组和良性肿瘤组Cho／Lip比值差异无统计学意义（t=1．53，P〉0．05）。ROC曲线下面积为0．77，^1H—MRS对诊断HCC有较高诊断价值，当Cho峰与Lip峰比值界值点为0．1时，诊断HCC的敏感性和特异性分别为80．0％和62．5％。结论肝脏局灶性病变^1H—MRS分析是可行的，对于鉴别HCC和良性肝脏肿瘤有一定的临床参考价值。

英文摘要：

Objective To investigate the value of in vivo proton MR spectroscopy （^1H-MRS） in the assessment of large focal hepatic tumors. Methods Fifty-three consecutive patients with 54 large （ no less than 4 cm in diameter） hepatic tumors and 19 normal volunteers were included in this study. MRS of the lesions was carried out using a whole-body 1.5 T MR scanner. In 17 normal volunteers, 8 benign tumors and 25 HCC confirmed by pathological examination, the choline-to-lipid ratios （Cho/Lip） were measured by dividing the peak area of choline at 3.2 ppm and lipid at 1.3 ppm. Differences in the ratios of normal liver, benign tumors and HCC were analyzed by Dunnett-t test. The sensitivity and specificity profiles of ^1H-MRS in the diagnosis of HCC were determined by plotting receiver operating characteristic （ROC） curves. Results The mean Cho/Lip ratios for normal liver （ n = 17 ） , benign tumor （ n = 8 ） , and HCC （ n = 25 ） were0.07±0.04, 0.11 ±0.06 and 0.55 ±0.17, respectively（F =6.58,P 〈0.05）. A significant statistical difference was found in the mean Cho/Lip ratios between HCC and benign liver tumors or normal liver（t =2. 99, 2. 32; P 〈0. 05）. But there was no difference between benign hepatic tumors and normal liver（ t = 1.53, P 〉 0.05 ）. The ROC curve showed proton MRS had moderate discriminating ability in diagnosing HCC. The area under the curve was 0. 77. If 0. 1 was chosen as the cut-off value for diagnosing HCC with MRS, the sensitivity and specificity for HCC were 80. 0% and 62. 5%, respectively. Conclusion In vivo proton MRS is technically feasible for the evaluation of focal hepatic lesions, and may be useful in the diagnosis and differential diagnosis of HCC by providing metabolic information .