3.0T MRI在肝脏结节性病变及弥漫性疾病的临床应用研究

【作者】 林光武；

【导师】 周纯武；

【作者基本信息】 中国协和医科大学 ， 影像医学与核医学， 2008， 博士

【摘要】 第一部分3.0T MRI全肝十期动态增强扫描LAVA技术成像参数探讨目的:分析正常成人和肝硬化患者MR全肝灌注扫描的“时间-信号强度”曲线特点和肝血管重建期相,探讨全肝十期动态增强扫描LAVA技术的成像参数。材料与方法:对8例无任何肝脏疾病的正常成人和8例肝硬化患者行MR全肝灌注成像,探讨全肝十期动态增强扫描LAVA技术的成像参数。（1）扫描野包括全肝,采用GE公司的LAVA容积成像技术,平静呼吸,用高压注射器经肘前静脉注入非离子型对比剂GD-DTPA 20ml,注射速度为3ml/s,在注射对比剂后10s开始扫描;每期扫描40层,层厚4.4mm,扫描时间为7秒,每期重建72层;连续扫描12期,共重建864层图像,扫描时间为84s。（2）扫描图像经adw 4.2工作站处理,绘制腹主动脉、门静脉、肝静脉和肝实质的“时间-信号强度”曲线,分析正常成人和肝硬化患者腹主动脉、门静脉、肝静脉和肝实质的达峰时间和波动期时间,腹主动脉、门静脉、肝静脉及肝实质同时处于平衡期时间,制定正常成人及肝硬化患者全肝十期最佳扫描时间及血管重建期相。结果:（1）正常成人和肝硬化患者腹主动脉的“时间-信号强度”曲线呈“速升-速降”型,强化峰值时间分别为25.75秒、26.63秒,波动期时间段分别为17～31秒、17～38秒;门静脉的“时间-信号强度”曲线呈“速升-缓降”型,峰值出现较主动脉晚,幅度明显低于主动脉,峰值强化时间分别为45、53.75秒,波动期时间段分别为31～59秒、38～66秒;肝静脉的“时间-信号强度”曲线呈“缓升-缓降”型,峰值出现较门静脉晚,峰值强化时间分别为59.88、68.63秒,波动期时间段分别为45～73秒、52～80秒;肝实质的“时间-信号强度”曲线呈“缓升-缓降”型,动脉期上升缓慢,门静脉期上升较快并迅速达峰值,然后缓慢下降呈平稳状态,肝实质峰值期时间分别为52.88、59.88秒,波动期时间段最长,分别为38～66秒、45～73秒;肝动脉、门静脉和肝实质均处于平稳状态时间分别为68.25、77.38秒,平衡期时间段分别为66～94秒、73～94秒。（2）正常成人和肝硬化患者腹主动脉达峰时间无显著性差异（P＞0.05）,门静脉、肝静脉、肝实质达峰时间和进入真正意义平衡期时间点均有显著性差异（P＜0.05）,说明肝硬化患者门脉供血延迟。（3）肝动脉、门脉、肝静脉峰值期显示血管图像评分,较峰值前期及峰值后期均有显著性差异（P＜0.05）。结论:（1）采用平静呼吸,注射20ml对比剂,注射速度为3ml/s,在注射对比剂后10s开始连续扫描84s共重建864层图像的扫描方式,基本能反映正常肝脏及肝硬化组织的血液动力学变化过程,是较理想并较实用的预扫描方案,能完整包括肝动脉、门静脉、肝静脉和肝实质平衡期多时相图像。选取肝动脉、门静脉、肝静脉峰值期图像,能很好重建出肝动脉、门脉及肝静脉血缸管,是较理想的血管重建期相。（2）选用屏气方法,用四个时间段扫描,前三次每次扫描3期（21秒）,正常及肝硬化组分别以10～31秒、12～33秒作动脉三期扫描,40～61秒、45～66秒静脉三期扫描,70～91秒、75～96秒肝实质平衡三期扫描,最后加作一期3分钟延迟扫描,基本上能包括所有肝实质波动期、平衡期和延迟期图像。第二部分3.0T MRI全肝十期动态增强扫描在肝脏结节性病变的检出及诊断价值目的:探讨MR全肝十期动态增强扫描在肝脏结节性病变的检出和诊断价值,分析各期相组合对病变诊断的敏感性、特异性和准确性,缺期扫描可能造成的漏诊及误诊率,总结肝脏结节性病变各期相的强化特点和最佳动态增强扫描方案。材料与方法:对180例肝脏结节性病变（肝炎后肝硬化伴或不伴肝癌病例、肝内转移瘤病例及肝脏其它少见结节性病变）行LAVA全肝十期动态增强扫描。扫描图像经adw4.2工作站和GC RIS PACS系统处理,绘制腹主动脉、门静脉和肝实质的“时间-信号强度”曲线,分析扫描十期图像是否完整包括肝动脉三期、门静脉三期、肝静脉三期及平衡三期,统计各期相对病变的显示价值;分析各期相组合可能造成的漏诊及误诊率,总结肝脏结节性病变的各期相强化特点和最佳动态增强扫描方案。结果:本组180例纳入对象148例符合要求,扫描十期图像完整包括肝动脉三期、门静脉三期、肝静脉三期、平衡三期及延迟期图像,占82.2%。1各期相强化特点（1）肝炎后肝硬化56例,其中肝癌54例,共80个病灶。①病灶强化持续时间为2或3期,分别为65%和28.75%;②病灶以动脉期开始强化为主,动脉早、中、晚期分别为25%、62.5%、12.5%;③最佳观察强化期为动脉中、晚期,分别占87.5%、93.75%;④肿瘤信号下降最快（对比剂廓清最快期）为门脉中期,占72.5%;⑤最佳观察信号下降期（廓清期）为门脉中、晚期,肝静脉中期及平衡期,分别为75%、86.25%、96.25%、100%。（2）肝癌介入治疗后20例,共33个病灶。①病灶强化持续时间为2或3期,分别为75.76%、24.24%;②病灶以动脉期开始强化为主,动脉早、中、晚期分别为15.15%、63.63%、21.21%:③最佳观察强化期为动脉中、晚期,分别为78.79%、90.91%;④肿瘤信号下降最快（对比剂廓清最快期）为门脉中、晚期,分别为45.45%、36.36%;⑤最佳观察信号下降期（廓清期）为门脉中、晚期及肝静脉中期,分别为54.55%、90.90%、100%。（3）肝炎后肝硬化56例,其中再生结节（RN或DN）12例,共发现共29个病灶。①病灶强化持续时间为1或2期,分别为58.62%、34.48%;②病灶以动脉早、中期开始强化为主,分别为31.03%、62.07%;③最佳观察强化期为动脉早、中、晚期,分别为31.03%、82.76%、34.48%;④病变信号下降最快（对比剂廓清最快期）为动脉晚期及门脉中期,分别为55.17%、34.48%;⑤最佳观察信号下降期（廓清期）为动脉晚期及门脉中期,分别为65.52%、100%。（4）富血供转移（乳腺癌肝转移24例,70个病灶）。①病灶强化持续时间为2或3期,分别为57.14%、32.86%;②病灶以动脉期开始强化为主,动脉早、中、晚期分别为14.29%、51.43%、34.29%%;③最佳观察强化期为动脉中、晚期及门脉中、晚期,分别为65.71%、97.14%、47.14%、30%;④病变信号下降最快（对比剂廓清最快期）为门脉中、晚期及肝静脉中期,分别为44.29%、25.71%、24.29%;⑤最佳观察信号下降期（廓清期）为门脉中、晚期及肝静脉中期和平衡期,分别为44.29%、70%、94.29%、94.29%。（5）乏血供转移（消化道及肺部肿瘤转移43例,126个病灶）。①病灶强化持续时间为2或3期,分别为65.08%、30.95%;②病灶以动脉中、晚期开始强化为主,分别为20.63%、65.87%;③最佳观察强化期为动脉中、晚期及门脉中、晚期,分别为20.63%、86.51%、87.30%、34.92%:④病变信号下降最快（对比剂廓清最快期）为门脉晚期及肝静脉中期,分别65.08%、26.19%;⑤最佳观察信号下降期（廓清期）为门脉晚期及肝静脉中期和平衡期,分别为65.08%、91.27%、99.21%。（6）肝脏局灶性结节增生5例,共8个病灶。①病灶强化持续时间为2或3期,分别为25%、62.5%;②病灶以动脉早、中期开始强化为主,分别为37.5%、62.5%;③最佳观察强化期为动脉早、中期,分别为37.5%、100%。④病变信号下降最快（对比剂廓清最快期）为门脉中、晚期,分别为50%、50%;⑤中央瘢痕开始强化为门脉晚期及肝实质平衡期,分别为25%、62.5%。2 LAVA十期及各期相组合诊断价值（1）各期相组合诊断肝癌的灵敏度、特异度、准确度、漏诊率及误诊率:十期分别为96.25%,96.57%,96.49%,3.75%,3.43%;九期分别为96.25%,96.57%,96.49%,3.75%,3.43%;八期分别为93.75%,96.57%,95.85%,6.25%,3.43%;六期分别为96.25%,96.57%,96.49%,3.75%,3.43%;四期分别为92.50%,94.85%,94.25%,7.50%,5.15%;三期分别为88.75%,92.27%,91.37%,11.25%,7.73%。（2）各期相组合经ROC检测诊断肝癌,其曲线下面积及P值:十期为0.985（p=0.000）,九期为0.984（0.009）,八期为0.980（0.009）,六期为0.983（0.009）,四期为0.969（0.000）,三期为0.955（0.000）。（3）各期相组合诊断肝硬化肝内再生结节的灵敏度、特异度、准确度、漏诊率及误诊率为:十期分别为93.10%,98.59%,98.08%,6.90%,1.41%;九期分别为86.21%,98.59%,97.44%,13.79%,1.41%;八期分别为68.97%,98.59%,95.85%,31.03%,1.41%;四期分别为58.62%,97.89%,94.23%,41.38%,2.11%;三期分别为51.72%,97.18%,92.97%,48.28%,2.82%。（4）各期相组合经ROC检测诊断肝内再生结节,其曲线下面积及P值:十期为0.980（p=0.000）,九期为0.961（0.000）,八期为0.921（0.000）,四期为0.881（0.000）,三期为0.839（0.000）。结论:肝内结节性病灶常呈多发性,完整意义十期扫描对肝内病变的检出及诊断具有重要价值,能较常规三期、四期扫描提供更多的检出及诊断信息。（1）肝癌组及肝癌介入治疗后病例以及肝内其它富血供病例,主要在动脉中、晚期强化,门脉中、晚期廓清,强化持续2～3期。因此,动脉中、晚期+门脉中、晚期+平衡期及延迟期共扫描6期应作为基本扫描方案。（2）肝炎后肝硬化肝内再生结节病例,主要在动脉早、中期强化,且强化期别非常短,故完整动脉三期+门脉中、晚期+平衡期共6期对其显示有重要价值。（3）肝乏血供转移瘤,大多接受门脉供血,动脉中、晚期,静脉三期,平衡期加延迟期扫描可满足诊断要求。第三部分3.0T MRI扩散加权成像在肝脏结节性病变及弥漫性疾病的量化诊断及鉴别诊断价值目的:探讨高场强MRI扩散加权成像对肝脏结节性病变及弥漫性疾病的量化诊断及鉴别诊断价值。材料与方法:2007.4～2008.3共431例肝脏MR扫描入选病例（正常肝脏、肝硬化、脂肪肝和肝脏结节性病变）进行常规MR扫描确定DWI扫描层面,采用EPI-DWI序列,b值取400,600,800s/mm²,行DWI扫描,观察图像质量并测量ADC值。结果:（1）DWI图像质量随着b值的增加而降低,在b值取600、800时,图像质量有显著性差异（t=1.97,p=0.001）。（2）正常肝脏及肝脏弥漫性疾病（b=600s/mm²）:正常肝脏ADC值=1.36±0.13×10^-3mm²/s,脂肪肝ADC值=1.18±0.13×10^-3mm²/s,肝硬化Child-Pugh A级ADC值=1.32±0.11×10^-3mm²/s,Child-Pugh B级ADC值=1.23±0.11×10^-3mm²/s,child-Pugh C级ADC值=1.11±0.11×10^-3mm²/s。正常肝脏与脂肪肝及肝硬化间ADC值均有显著性差异（P＜0.05）,肝硬化Child-Pugh A、B、C级间亦有统计学差异（P＜0.05）。（3）肝脏结节性病变（b=600s/mm²）:肝细胞癌ADC值=1.38±0.15×10^-3mm²/s,肝癌介入后ADC值=1.42±0.21×10^-3mm²/s,肝乏血供转移癌ADC值=1.36±0.19×10^-3mm²/s,肝富血供转移癌ADC值=1.49±0.19×10^-3mm²/s,胆管细胞癌ADC值=1.44±0.12×10^-3mm²/s,血管瘤ADC值=1.90±0.16×10^-3mm²/s,肝良性肿瘤ADC值=1.48±0.19×10^-3mm²/s。肝脏良性病灶的ADC值大于恶性病灶,但二者亦有交叉。b=600时,截止点ADC值取1.55×10^-3mm²/s,诊断的敏感性为80.18%,特异性为81.08%,准确性为80.41%,阳性预测值为92.71%,阴性预测值57.69%,阳性似然比4.24,阴性似然比为0.24。结论:（1）正常肝脏及病变DWI信号随着扩散敏感度b值的增大均匀降低,综合评价图像质量及ADC值鉴别价值,b值取600s/mm²较为合适。（2）ADC值对正常肝脏、脂肪肝及肝硬化Child-Pugh A、B、C级间有量化价值。（3）肝脏良性结节性病变的ADC值大于恶性病变,选择合适的b值测量ADC值,有助于良、恶性病变的鉴别。更多还原

【Abstract】 PartⅠTo Study the Technical Parameters of Hepatic 10-phase Dynamic Enhancement with Liver Acceleration Volume Acquisition（LAVA） at 3.0T MRIPurpose:To analyze the signal intensity time curves and the phase of hepatic vascular reconstruction at normal liver and hepatic cirrhosis by whole liver perfusion weighted imaging,in order to investigate the technical parameters of hepatic 10-phase dynamic enhancement with liver acceleration volume acquisition（LAVA） at 3.0 T MRI.Materials and Methods:Perfusion weighted images of whole liver were performed in 8 health adults and 8 patients with hepatic cirrhosis,in order to investigate the technical parameters of hepatic 10-phases dynamic enhancement with liver acceleration volume acquisition（LAVA） at high field 3.0 T MRI.（1） After bolus injection 20 ml GD-DTPA（the injection rate was 3ml/s） of 10 seconds in anterior ulnar vein by high-pressure syringe,perfusion weighted images of whole liver were performed with LAVA technique at quiet breath by a 3.0 T MRI（GE medical system）. With section thickness of 4.4 mm,scan time of 7s,40 slices of whole liver were scan every phase.The total 864 slices（72 slices were reconstructed every phase） were reconstructed with 12 phases scan continuously at 84 seconds.（2） The scanned images were processing by adw4.2 workstation.The signal intensity-time curves of abdominal aorta,portal vein,hepatic vein and liver parenchyma were drawn and the peak time,fluctuant time and the simultaneously reach time of balanced phase were analyzed in both normal liver and hepatic cirrhosis. Results:（1） The signal intensity time curve of abdominal aorta was rapid ascending and rapid descending both at health adults and patients with hepatic cirrhosis.The peak time of the abdominal aorta enhancement was 25.75s,26.63s;Time segments of fluctuant state were 17～31s,17～38s respectively.The signal intensity time curve of portal vein was rapid ascending and slow descending.The peak time of portal vein enhancement were 45s,53.75s,Time segments of the fluctuant state were 31～59s and 38～66s respectively.The peak time of portal vein was longer than abdominal aorta and the intensity of enhancement was weaker than abdominal aorta obviously.The signal intensity time curve of hepatic vein was slow ascending and slow descending. The peak time of hepatic vein enhancement was 59.88s,68.63;Time segments of the fluctuant state were 45～73s and 52～80s respectively.The signal intensity time curve of liver parenchyma was slow ascending and slow descending.The signal intensity of liver parenchyma increases slowly in arterial phase and rapidly reached the peak in portal phase.And then,it decreases slowly and keeps relative stationary state.The peak time of hepatic parenchyma enhancement was 52.88s,59.88s;Time segments of the fluctuant phase were 38～66s and 45～73s respectively.The simultaneously reached equilibrium phase time of abdominal aorta,portal vein,hepatic vein and liver parenchyma was 68.25s,77.38s;the time segments of equilibrium phase were 66～94s, 73～94s respectively.（2） The peak time of abdominal aorta has no significant difference between health adult and patients with hepatic cirrhosis（P＞0.05）.But the peak time of portal vein,hepatic vein and liver parenchyma and the real time of equilibrium phase have significant difference（P＜0.05）.It illustrates the blood supply of portal vein in patients with hepatic cirrhosis longer than the healthy.（3）There was significant difference between the peak time phase and other phase of showing the quality of hepatic artery,portal vein and hepatic vein imaging（P＜0.05）.Conclusion:（1） The scan mode can reflect the hemodynamic alteration procedure at both normal liver and liver cirrhosis;it is an ideal and practical pre-scan procedure, which included hepatic artery,portal vein,hepatic vein,liver parenchyma and equilibrium phase entirely.The optimal phases to observe and reconstruct vessel of hepatic artery,portal vein and hepatic vein are the peak time phase images correspondingly.（2） The optimal scan mode include with breath holding method,the time of arterial 3 phases at 10～31s,12～33s,vein 3 phases at 40～61s,45～66s, equilibrium 3 phases at 70～91s,75～96s for normal adults and patients with hepatic cirrhosis respectively,which could include all the fluctuant phases and equilibrium phases of whole liver entirely. PartⅡThe Detection and Diagnosis of Hepatic Nodular Lesions with 3.0T MRI Ten-phase Dynamic ScanningObjective:To assess the value of MRI ten-phase dynamic scanning in the detection and diagnosis of hepatic nodular lesions,to analyze the sensitivity,specificity, accuracy,false negative rate and false positive rate of multi-phase combination scans, and to summarize the enhancement features of nodular lesions in different phase and their optimal enhancing scan projects.Materials and Methods:Ten-phase dynamic enhanced images of whole liver were performed with LAVA volume technique in 180 cases with hepatic nodular lesions （hepatic cirrhosis after hepatitis with or without hepatic cancer,Hepatic Metastases and other rare nodular lesions）.All images were processed with adw4.2 workstation and GC RIS PACS,time-intensity curves for abdominal aorta,portal veins and hepatic parenchyma were obtained.Ten-phase images including arterial three phases, portal vein three phases,hepatic venous three phases and balanced 3 phases were analyzed.The values of different phase to demonstrate lesions were assessed,as well as multi-phase combination.The dynamic enhancement features of hepatic nodular lesions and their optimal scan projects were summarized.Results:148 cases（82.2%） completed scans of ten phases including arterial three phases,portal vein three phases,hepatic venous three phases,balanced three phases and delayed phase.1 Enhancing features in different phase（1） 80 lesions in 54 cases of hepatic cirrhosis after hepatitis with liver cancer:① Enhancement persists for 2 or 3 phases in 65%and 28.75%respectively.②Enhancement mainly begins at arterial phase,arterial early phases in 25%,arterial middle phase in 62.5%and arterial late phase in 12.5%.③Optimal phases to observe enhancement（enhancing phase） were arterial middle phase in 87.5%and arterial late phase in 93.75%.④Fastest signal attenuation of tumors was observed at portal middle phase in 72.5%（fastest attenuation phase）.⑤The optimal phases to observe signal attenuation（attenuation phase） were portal middle,late phase,hepatic venous late phase and balanced phase in 75%,86.25%,96.25%and 100%respectively.（2） 33 lesions in 20 cases with liver cancer after intervention treatment:①Enhancement persists for 2 or 3 phases in cases of 75.76%and 24.24%respectively.②Enhancement mainly begins at arterial phase,arterial early phase in 15.15%, arterial middle phase in 63.63%and arterial late phase in 21.21%.③The optimal phases to observe enhancement were arterial middle phase in 78.79%,and arterial late phase in 90.91%.④Fastest signal attenuation of tumors（fastest attenuation phase） were observed at portal middle phase in 45.45%and portal late phase in 36.36%.⑤The optimal phases to observe signal attenuation（attenuation phase） were portal middle,late phase and hepatic venous middle phase in 54.55%,90.90%and 100% respectively.（3） 29 lesions in 56 cases with hepatic cirrhosis after hepatitis accompanying liver cancer or not,including 12 cases with regenerative nodules or dysplastic nodular（RN or DN）:①Enhancement persists for 2 or 3 phases in 58.62%and 34.48%respectively.②Enhancement begins at arterial phase for most lesions,arterial early phases in 31.03%and arterial middle phase in 62.07%.③The optimal phases to observe enhancement were arterial early,middle and late phase in 31.03%,82.76%and 34.48%respectively.④Fastest signal attenuation of tumors were observed at arterial late phase in 55.17%and portal middle phase in 34.48%.⑤The optimal phases to observe signal attenuation（attenuation phase） were arterial late phase in 65.52%and portal middle phase in 100%.（4） Hyper-vascular Hepatic Metastases（70 lesions in 24 cases with primary breast carcinoma）:①Enhancement persists for 2 or 3 phases in cases of 57.14%and 32.86%respectively.②Enhancement begins at arterial phase for most lesions,arterial early,middle and late phase in 14.29%,51.43%and 34.29%,respectively.③The optimal phases to observe enhancement（enhancing phase） were arterial middle,late phase,portal middle and late phase in 65.71%,97.14%,47.14%and 30%respectively.④Fastest signal attenuation of tumors were observed at portal middle,late phase and hepatic venous phase in 44.29%,25.71%and 24.29%respectively.⑤Optimal phases to observe signal attenuation（attenuation phase） were portal middle,late phase, hepatic venous middle phase and balanced phase in 44.29%,70%,94.29%and 94.29%respectively.（5） Hypo-vascular Hepatic Metastases（126 lesions in 43 cases with primary digestive tract or lung tumors）:①Enhancement persists for 2 or 3 phases in cases of 65.08% and 30.95%respectively.②Enhancement begins at arterial phase for most lesions, arterial middle and late phase in 20.63%and 65.87%,respectively.③The optimal phases to observe enhancement（enhancing phase） were arterial middle,late phase, portal middle and late phase in 20.63%,85.61%,87.30%and 34.92%,respectively.④Fastest signal attenuation of tumors were observed at portal late phase in 65.08% and hepatic venous middle phase in 26.19%.⑤The optimal phases to observe signal attenuation（attenuation phase） were portal late phase,hepatic venous middle phase and balanced phase in 65.08%,91.27%and 99.21%respectively.（6） 8 lesions in 5 cases with focal nodular hyperplasia:①Enhancement persists for 2 or 3 phases in cases of 25%and 62.5%respectively.②Enhancement mainly begins at arterial early and middle phases in 37.5%and 62.5%respectively.③The optimal phases to observe enhancement were early arterial and peak phase in 37.5%,100% respectively.④Fastest signal attenuation of tumors was observed at portal peak phase for 50%and late portal phase for 50%.⑤Central scars enhancement begins at late portal phase and balanced liver parenchyma phase in 25%and 62.5%respectively.2 The diagnosis value of LAVA ten phases and multi-phase combination（1）The sensitivity,specificity,accuracy,false negative rate and false positive rate in diagnosis of liver cancer with ten-phase combination were 96.25%,96.57%,96.49%, 3.75%and 3.43%respectively;nine-phase were 96.25%,96.57%,96.49%,3.75%and 3.43%respectively;eight-phase were 93.75%,96.57%,95.85%,6.25%and 3.43% respectively;six-phase were 96.25%,96.57%,96.49%,3.75%and 3.43%respectively; four-phase were 92.50%,94.85%,94.25%,7.50%and 5.15%respectively; three-phase were 88.75%,92.27%,91.37%,11.25%and 7.73%respectively.（2）The diagnosis of liver cancer with multi-phase combination through ROC:Area under curves was 0.985（p=0.000） for ten-phase combination,0.984（p=0.009） for nine-phase,0.980（p=0.009） for eight-phase,0.983（p=0.009） for six-phase,0.969 （p=0.000） for four-phase and 0.955（p=0.000） for three-phase.（3） The sensitivity,specificity,accuracy,false negative rate and false positive rate in diagnosis of regenerative nodules or dysplastic nodular lesions of liver with ten-phase combination were 93.10%,98.59%,98.08%,6.90%and 1.41%respectively; nine-phase were 86.21%,98.59%,97.44%,13.79%and 1.41%respectively; eight-phase were 68.97%,98.59%,95.85%,31.03%and 1.41%respectively; six-phase were 96.25%,96.57%,96.49%,3.75%and 3.43%respectively;four-phase were 58.62%,97.89%,94.23%,41.38%and 2.11%respectively;three-phase were 51.72%,97.18%,92.97%,48.28%and 2.82%respectively.（4） The diagnosis of liver regenerative nodular or dysplastic nodular lesions with multi-phase combination through ROC:Area under curves was 0.980（p=0.000） for ten-phase combination,0.961（p=0.000） for nine-phase,0.921（p=0.000） for eight-phase,0.881（p=0.000） for four-phase and 0.839（p=0.000） for three-phase.Conclusions:Nodular lesions are often multiple in the liver.Ten-phase dynamic scanning is important in the detection and diagnosis of hepatic lesions,which can provide more characteristic and definite information than route three-phase or four-phase scans.（1）Enhancing at arterial peak or late arterial phase,signal attenuation at portal peak or late portal phase and persistent enhancement for two or three phases are the features of liver cancer with or without intervention treatment and other lesions with hyper-vascular.Six-phase scan including arterial peak,late phase,portal peak and late phase,balanced and delayed phase is the basic scan project for such lesions.（2）Regenerative nodular or dysplastic nodular lesions in patients with hepatic cirrhosis after hepatitis often show enhancement at early arterial and peak phase,with short enhancement and attenuation phase.Six-phase scan including arterial three phases,portal peak and late phase,and balanced one phase is important to detect those lesions.（3）Hepatic metastasis tumors with hypo-vascular were often supplied by portal vein. Multi-phase scan including arterial peak,late phase,venous three phases,balanced and delayed phase can give enough information for diagnosis. PartⅢEvaluation of Diffusion Weighted Imaging on Differential Diagnosis and Quantitative Diagnosis of Nodular Hepatic Lesions and Diffused Liver Lesions at 3 Tesla Magnetic ResonanceObjective:To investigate the value of Diffusion-Weighted Imaging（DWI） on differential diagnosis and quantitative diagnosis of liver diffuse lesions and nodular hepatic lesions with 3.0 T MRI.Materials and Methods:Diffusion weighted imaging（DWI） was obtained in a total of 431 cases of nodular hepatic lesions and liver diffuse lesions including normal liver, liver cirrhosis,fatty liver,hepatocellular carcinoma（HCC）,cholangiocellular carcinoma（CC） and secondary tumors of liver from Apr,2007 to Mar,2008.The quality of the DWI and apparent diffusion coefficient（ADC） values were analyzed.Results:（1） The quality of DWI attenuated homogeneously according to the gradient factor（b value）.There was a significant difference between the quality of b=600 and b=800 DWI（p＜0.05）.（2） The ADC value of the normal liver,fatty liver,liver cirrhosis Child-Pugh A,B,C was 1.36±0.13×10^-3mm²/s,1.18±0.13×10^-3mm²/s, 1.32±0.11×10^-3mm²/s,1.23±0.11×10^-3mm²/s,1.11±0.11×10^-3mm²/s respectively. There was a significant difference between normal liver tissue,fatty liver and the liver cirrhosis,（P＜0.05）,and significant difference were found among Child-Pugh A,B,C （P＜0.05）.（3） The ADC value of the hepatocellular carcinoma（HCC）,the HCC after transcatheter arterial chemoembolization（TACE） treatment,hypovascular Hepatic Metastases,hypervascular Hepatic Metastases,cholangiocarcinoma,haemangioma, and benign lesions was 1.38±0.15×10^-3mm²/s,1.42±0.21×10^-3mm²/s,1.36± 0.19×10^-3mm²/s,1.49±0.19×10^-3mm²/s,1.44±0.12×10^-3mm²/s,1.90±0.16×10^-3mm²/s,1.48±0.19×10^-3mm²/s respectively,and the ADC value was higher in benign lesion than that in malignant tumor.With ADC value of 1.55×10^-3mm²/s as the threshold for malignant tumors and benign lesions,the sensitivity,specificity, accuracy,positive predictive value,negative predictive value,positive likelihood ratio, and negative likelihood ratio were 80.18%,81.08%,80.41%,92.71%,57.69%,4.24, 0.24,respectively.Conclusion:（1） With b value increased,the ADC value and quality of DWI decreased at the same time.According to both the value of differential diagnosis and the quality of DWI,b=600 is a free parameters in DWI scanning to get appropriate imaging on 3.0T MRI.（2） DWI is relevant to the severity of hepatic cirrhosis and fatty liver,and may be a useful quantitative tool for evaluating the severity of hepatic cirrhosis and fatty liver.（3） ADC values of malignant tumors were lower than that of benign lesions.Proper threshold of ADC value was helpful to differentiating malignant tumor from benign lesion.更多还原