【作者】 袁正；

【导师】 肖湘生；

【作者基本信息】 第二军医大学 ， 影像医学与核医学， 2008， 博士

【摘要】 研究背景及时、有效评价肿瘤治疗疗效在判断治疗是否成功和指导后续治疗方面是相当重要的。目前应用于临床的监测肿瘤治疗反应的方式较多,包括组织学检查、肿瘤标志物监测和影像学方法等。其中,组织学检查是评价的金标准,然而使用组织学评价只有当取得的标本是活性肿瘤组织时才能得到肯定性的结果,因此,即使反复出现的阴性活检结果也不能排除肿瘤残留的可能;而在评价肿瘤的治疗反应中,单独使用肿瘤标记物的临床应用价值也非常有限。各种影像学成像方法,包括多普勒超声（Doppler ultrasonography）、血管造影（angiography）、CT（computedtomography）和磁共振成像（magnetic resonance imaging,MRI）都可以用于评价治疗反应。目前临床上应用的实体瘤疗效评价标准（WHO标准和RECIST标准）是基于影像学上病灶治疗前后的大小变化来评价肿瘤疗效,因治疗后病灶的形态学在一定时间内可能并没有发生改变,所以这种形态学评价方法结果具有滞后性和欠准确性。临床工作者开始将注意力集中到开发一些新的更能真实反映疗效的影像学评价方法。这些新的方法包括评价肿瘤血管和细胞的完整性,水分子的运动和肿瘤内化合物的浓度变化等。本研究运用磁共振功能成像技术预测和早期评价肝癌化疗栓塞疗效。研究目的1、评价不同扩散梯度系数（b值）条件下肝癌化疗栓塞术后病灶磁共振扩散加权成像（DWI）的图像质量,筛选肝癌化疗栓塞术后磁共振扩散加权成像的最佳扩散梯度系数/1)值。2、评价肝癌化疗栓塞术后病灶的DWI图像的信号特点,定量分析肿瘤内部各组织成分的ADC值,探讨ADC值鉴别肿瘤内部不同组织成分的能力。3、定量分析肝癌化疗栓塞术前、术后病灶ADC值,探讨病灶术前ADC值在预测肝癌化疗栓塞疗效中的价值和病灶术后ADC变化值早期评价肝癌化疗栓塞疗效价值。4、初步观察原发性肝癌³¹P-MRS的特点,并初步探讨³¹P-MRS在早期评价原发性肝癌化疗栓塞治疗疗效的临床价值。研究方法1、采用SS-SE-EPI序列、弥散系数b分别为150、350、500、800 s/mm²对肝脏进行扩散加权成像,计算T2WI和不同b值DWI中病灶信号-噪声比（signal-noise-ratio,SNR）,病灶-肝脏对比信噪比（contrast signal-noise-ratio,CNR）,不同b值DWI中病灶ADC值,通过观察CT和DSA造影图像,分析病灶内部碘化油沉积、液化坏死和肿瘤残留或复发的活组织情况,定性分析在不同b值下DWI图像上区分病灶内部各种组织成分的能力。2、采用SS-SE-EPI序列、弥散系数b=0和500s/mm²对肝脏进行扩散加权成像,计算DWI、T2WI序列中观察到的新发肿瘤病灶与周围肝组织之间的CNR,观察上腹部CT增强检查和肝脏DSA造影图像,分析病灶内部碘化油沉积情况,肿瘤血供、肿瘤染色和肿瘤坏死情况,并与DWI图像相对照,分析碘化油沉积区、肿瘤坏死区和肿瘤复发/残留区等在DWI图像上的信号特点,并测量上述区域ADC值,用ROC曲线分析（receiver operating curve analysis）ADC值鉴别肿瘤活性和坏死组织的价值。3、采用SS-SE-EPI序列、弥散系数b=0和500s/mm²对肝脏进行扩散加权成像,分别于化疗栓塞术前24小时内和化疗栓塞术后24-48小时内完成DWI检查,计算病灶术前ADC值和术后%ADC值,与栓塞效果比较,用ROC曲线分析术前病灶ADC值预测肝癌化疗栓塞治疗疗效和术后%ADC值早期评价肝癌化疗栓塞治疗疗效的临床价值。4、初步观察正常肝脏组织和肝癌组织的³¹P-MRS主要特点,比较肝癌组织和正常肝组织中的代谢物变化情况;定量分析原发性肝细胞肝癌化疗栓塞治疗前后³¹P-MRS中各化合物变化情况;同时,通过与这些病例的临床随访结果对照,初步评价³¹P-MRS在早期评价肿瘤疗效中的价值。研究结果1、化疗栓塞后病灶在T2WI、DWI₁₅₀、DWI₃₅₀、DWI₅₀₀和DWI₈₀₀的SNR分别为:（69.81±18.99）、（59.33±32.66）、（59.23±32.94）、（54.25±19.71）和（39.43±11.67）,差异有统计学意义（F=4.43,P=0.0024＜0.05）;病灶.肝脏CNR分别为:（19.11±11.33）、（17.69±9.20）、（21.38±10.10）、（19.90±13.75）和（13.24±11.02）,差异没有统计学意义（F=1.70,P=0.1556＞0.05）;b=150,350,500和800s/mm²的ADC图中测得病灶ADC值分别为:（2.35±0.80）×10^-3mm²/s、（1.95±0.59）×10^-3mm²/s、（1.78±0.44）×10^-3mm²/s和（1.54±0.37）×10^-3mm²/s,差异有统计学意义（F=21.96,P=0.0001＜0.05）。使用b=500和800s/mm²的DWI图像能有效区分肿瘤内部坏死和活性肿瘤组织。2、新发病灶在DWI图像中的CNR为46.36±19.49,在T2WI图像上的CNR为33.24±17.26,两者差异有统计学意义（t=2.400,P=0.025＜0.05）;在DWI图像上,肿瘤坏死组织呈低信号,肿瘤活性组织呈高信号,而碘油沉积区呈低、等或稍高信号。DWI和ADC值能有效区分肿瘤内的坏死组织和活组织,用ROC曲线分析,以ADC值为1.84×10^-3mm²/s为域值,诊断坏死组织的敏感度为92.3%,特异度为100%。3、栓塞良好组术前病灶ADC值中位数为1.278×10^-3mm²/s(范围:0.989×10^-3mm²/s～1.768×10^-3mm²/s);栓塞不良组中位数为1.687×10^-3mm²/s(范围:1.24×10^-3mm²/s～2.20×10^-3mm²/s)（P＜0.05）。用ROC曲线分析,以ADC值为1.618×10^-3mm²/s为预测栓塞不良域值,敏感度为96%,特异度为77.8%;栓塞良好组治疗前后病灶%ADC值中位数为32.63%（范围:11.15%～182.16%）,栓塞不良组%ADC中位数为5.24%（范围:-11.06%～15.50%）（P＜0.05）;用ROC曲线分析,以%ADC值为16.21%为预测栓塞效果良好域值,敏感度为72%,特异度为100%。非治疗叶段肝实质ADC值（t=0.174,P=0.862）和脾脏治疗前后ADC值（t=1.964,P=0.052）变化无统计学意义。4、在原发性肝癌中,磷酸单酯（PME）水平中位数为2.98×10⁷（范围:0.846×10⁷～102.5×10⁷）,高于正常肝实质中PME水平（中位数:1.81×10⁷;范围:0.734×10⁷～2.38×10⁷）:HCC中磷酸二酯（PDE）水平中位数为1.94×10⁷（范围:0.384×10⁷～5.24×10⁷）,低于正常肝实质（中位数:3.39×10⁷;范围:0.661×10⁷～3.55×10⁷）;治疗后PME水平（中位数:1.38×10⁷;范围:0.665×10⁷～6.21×10⁷）低于治疗前（中位数:2.98×10⁷;范围:0.846×10⁷～102.5×10⁷）（P＜0.05）。临床随访发现治疗有效病灶,其³¹P-MRS中,PME/NPT（P＜0.01）和PDE/NPT（P＜0.01）比值治疗后与治疗前相比均下降;而临床随访中疗效归为治疗无效的病灶,治疗后24-48h内与治疗前相比,PME/NPT比值下降（P＜0.05）,而PDE/NPT比值轻度上升,但差异无统计学意义（P＞0.05）。结论1、在肝癌化疗栓塞后DWI中,选择合适的扩散梯度系数能够得到较佳的信号对比,同时也能较准确地区分坏死组织和活性肿瘤。2、磁共振扩散加权成像能敏感检出肝癌化疗栓塞术后病人肝内新发肿瘤,ADC值能有效区分肿瘤内部坏死和活性肿瘤成分,可用于肝癌化疗栓塞术后病人的临床随访。3、肝癌化疗栓塞术前病灶ADC值能预测栓塞效果,化疗栓塞治疗后癌灶ADC值有不同程度上升并能反应癌灶的栓塞效果,对指导临床治疗方式的选择和个性化治疗方案的制定有一定意义。4、在原发性肝癌化疗栓塞术前和术后48h内³¹P-MRS中PME/NPT和PDE/NPT的比值变化对早期评价疗效和后续治疗有指导意义。更多还原

【Abstract】 BackgroundAssessment of tumor response after therapies is important in determining treatment success and in guiding future therapy.Several monitors of tumor response have been used,including histology,tumor markers,and imaging.However,histologic evaluation using tissue biopsy can only be conclusive when it shows viable malignancy.Therefore, repeated negative biopsies do not exclude the presence of residual tumor.Tumor markers sorely are of limited use in assessing tumor response.Various imaging modalities,including Doppler ultrasonography,angiography,computed tomography （CT）,and magnetic resonance imaging（MRI） have been used to evaluate therapeutic response.In clinic,standardized criteria for measuring therapeutic response have been established in 1981 by the World Health Organization（WHO）.In 2000,the Response Evaluation Criteria in Solid Tumors（RECIST） was introduced to unify response assessment criteria.However,these methods rely on size change of lesions to assess response.Therefore,the early effects of some therapies would not be detected and occasionally patients may not be considered to have exhibited a response despite the presence of tumor necrosis.This realization that anatomy may not change after therapies moved the focus toward new evaluation methods.These include assessment of tumor vascular and cellular integrity,motion of water molecules,and biochemical concentration.In this article we discuss the role of functional MRI in assessing treatment response after transarterial chemoembolization in patients with liver cancer.Objective 1.To assess the diffusion-weighted MR image quality with different b values in the follow-up after transarterial chemoembolization for liver cancer.2.To evaluate the feasibility and capability of diffusion-weighted MR imaging technique in follow-up after transarterial chemoembolization for liver cancer.3.To determine whether the pretreatment apparent diffusion coefficients（ADCs） of liver cancer are predictive of response to transarterial chemoembolization and to compare the ADCs of tumors before and after transarterial chemoembolization.4.To investigate the value of ³¹p MR spectroscopy in monitoring the early response of therapy for hepatocellular carcinoma.Methods1.Diffusion-weighted MR imaging and routine non-enhanced MR imaging were performed in patients with liver cancer after transarterial chemoembolization at 1.5T MRI.DWI were performed with the same SS-SE-EPI sequence and different b value, b＝0 and b≠0（b＝150,350,500 and 800s/mm²,respectively）.The signal-noise-ration （SNR） of the lesions after transarterial chemoembolization and the contrast noise ratio （CNR） between the lesion and liver tissue on diffusion-weighted images with different b values and T2-weighted images were measured.The lesion apparent diffusion coefficient（ADC） values were measured on functional diffusion maps.Qualitative evaluations the capability of identification the viable and necrotic tissue in tumor were performed on diffusion-weighted imaging with different diffusion coefficients comparing with DSA and enhanced-CT.2.Diffusion-weighted MR imaging and routine non-enhanced MR imaging were performed in patients with liver cancer after transarterial chemoembolization with the same SS-SE-EPI sequence（b=500s/mm² and 0） at 1.5T MRI.Contrast noise ratio（CNR） between new lesions and liver for diffusion-weighted MR image and T₂WI were measured.The characteristics of the treated lesion on diffusion-weighted image were evaluated.And the capability of discriminating the remaining or recurrent viable tumor and necrotic tissue were evaluated and compared with DSA and enhanced-CT.3.Liver cancers were prospectively evaluated with diffusion-weighted imaging at two b values before and after transarterial chemoemblization.Quantitative ADC maps were calculated with images with b values of 0,and 500s/mm².The mean ADC values of lesions before and after transarterial chemoemblization were compared according to response defined by the effect of transarterial chemoembolization.4.Phosphorus-31 MR spectra of hepatocellular carcinoma obtained before and at 24-48h after transarterial chemoemblization in 15 patients with 17 hepatocellular carcinomas.Alterations of phosphorus metabolism between before and after transarterial chemoembolization were compared with tumor responses evaluated by means of long-term follow-up.Results1.The SNR of the treated lesions were 69.81±18.99,59.33±32.66,59.23±32.94, 54.25±19.71,39.43±11.67 on T2WI,DWI（b＝150s/mm²）,DWI（b=350s/mm²）, DWI（b=500s/mm²） and DWI（b=800s/mm²）,respectively（F=4.43,P=0.0024＜0.05）. And CNR between the treated lesion and liver were 19.11±11.33,17.69±9.20, 21.38±10.10,19.90±13.75 and 13.24±11.02 on T2WI,DWI（b=150s/mm²）, DWI（b=350s/mm²）,DWI（b=500s/mm²） and DWI（b=800s/mm²）,respectively（F=1.70, P=0.1556＞0.05）.The treated lesion ADC values were（2.35±0.80）×10^-3mm²/s,（1.95±0.59）×10^-3mm²/s,（1.78±0.44）×10^-3mm²/s and（1.54±0.37）×10^-3mm²/s on functional diffusion maps with b=150,350,500 and 800s/mm²,respectively（F=21.96, P=0.0001＜0.05）.The viable and necrotic tissue in treated lesion can be identified more clearly on diffusion-weighted images with b=500s/mm² and with b=800s/mm² than diffusion-weighted images with b=150s/mm² and with b=350s/mm².2.There was significant difference in CNR between new cancer lesion and liver tissue on diffusion-weighted and T2-weighted images（46.36±19.49 vs.33.24±17.26, t=2.400,P=0.025）.Liver cancer after transarterial chemoemblization has variable signal intensity on diffusion-weighted MR image.On diffusion-weighted image,the hypo-intensity regions in the treated lesions correspond to regional no enhancement on enhanced-CT or no tumor stain tissue on angiogram that means necrotic tissue because of decreased cellular density;and the hyper-intensity regions in the treated lesions mean the viable tissue because of higher cellular density and hyper-vascular which enhancement on enhanced-CT and tumor stain on angiogram.The results of receiver operator characteristic（ROC） analysis for differentiation of viable and necrotic tumor tissue with mean apparent diffusion coefficients for images with b values of 0 and 500 s/mm² showed that threshold ADC value of 1.84×10^-3mm²/s had 92.3%sensitivity and 100%specificity for identification of necrotic tumor tissue.3.Nonresponding lesions had a significantly higher pretreatment mean ADC than did responding lesions(1.687×10^-3 mm²/s vs.1.278×10^-3 mm²/s,P＜0.05).The results of receiver operator characteristic（ROC） analysis for differentiation of nonresponding and responding lesions with mean apparent diffusion coefficients for images with b values of 0 and 500 s/mm² showed that threshold ADC value of 1.618×10^-3 mm²/s had 96.0%sensitivity and 77.8%specificity for identification of nonresponding lesions. After transarterial chemoembolization,responding lesions had a significant increase in %ADC values than did nonresponding lesions（32.63%vs.5.24%,P=0.025）.The results of receiver operator characteristic（ROC） analysis for differentiation of nonresponding and responding lesions with mean apparent diffusion coefficients changes for images with b values of 0 and 500 s/mm² before and after transarterial chemoembolization showed that threshold%ADC value of 16.21%had 72%sensitivity and 100%specificity for identification of responding lesions.No significant ADC value change was observed in normal liver parenchyma（t=0.174,P=0.862） and spleen （t=1.964,P=0.052） after transarterial chemoembolization.4.For alterations in phosphorus metabolism in HCCs,the median phosphomonoesters（PME） level in 17 HCCs after treatment was 1.38×10⁷（range, 0.665×10⁷-6.21×10⁷）,which was significantly decreased than that before treatment （median,2.98×10⁷;range,0.846×10⁷-102.5×10⁷）（P＜0.05）.In long-term follow-up studies,the ratio of phosphomonoester to nucleoside triphosphate（PME/NPT）（P＜0.01） and the ratio of phosphodiester to nucleoside triphosphate（PDE/NPT）（P＜0.01） was remarkable decreased before and after treatment in the HCCs（11/17） which showed a response to therapy;in other 6 HCCs classified as no response,each PME/NPT ratio significantly diminished（P＜0.05）,while the PDE/NPT ratio rose slightly,but no significant（P＞0.05）. Conclusions1.In the follow-up after transarterial chemoembolization for liver cancer with DWI, diffusion coefficient influences the image quality.The diffusion-weighted MR imaging with suitable diffusion coefficient can achieve high image quality and identify viable and necrotic tissue in tumor more clearly.2.Diffusion-weighted MR imaging is a sensitive and useful method for detecting new lesion and the remaining or recurrent viable tumor for liver cancer after transarterial chemoembolization.3.High pretreatment mean ADC of liver cancers was predictive of poor response to chemoembolization.A significant increase in mean ADC was observed in lesions that responded to chemoembolization.These findings may have implications for development of individualized therapy.4.The results demonstrate ³¹P MR spectroscopy is a suitable method for monitoring the early response of HCC to transarterial chemoembolization.更多还原