【作者】 王澜；

【导师】 祝淑钗；

【作者基本信息】 河北医科大学 ， 肿瘤学， 2014， 博士

【摘要】 食管癌三维适形或调强放疗目前已普遍应用于临床，其治疗后的长期局控率和生存率与普通外照射相比也有了较大幅度的提高，但是在食管癌的精确放疗研究中，还存在诸多不确定之处，如食管癌大体肿瘤区(GrossTumor Volume—GTV)的确定、合理的临床靶区(Clinic Target Volume—GTV)外放范围、标准的放化综合治疗模式、非手术临床分期标准的完善及尚无统一、明确地量化近期疗效评价标准等。其中，在食管癌精确放疗的GTV勾画方面，目前仍需参考其他诊断信息在CT图像上进行靶区勾画，但CT扫描在判断食管癌浸润深度上并不准确，尤其是对于T1、T2期的原发灶很难界定，亦不能准确判断食管病变的确切长度。此外，恶性肿瘤经非手术治疗后通常用CT、常规MRI、X线造影等影像学检查进行疗效判断，但是肿瘤治疗后形态学的变化并不能完全反应其功能、代谢的变化，对近期疗效进行客观、量化、准确地评估具有重要意义，临床可依据其治疗反应情况制定后续方案并对患者预后进行评估。但目前学界仍缺乏明确的应用于食管癌的近期疗效评价的统一标准。磁共振弥散加权成像(DWI)检查是近年发展的较新的功能成像技术，初期研究显示该检查可对肿瘤组织进行早期诊断、临床分期及疗效监测。因磁共振常规T1WI、T2WI扫描可较清楚的现实肿瘤形态学特点，同时弥散加权成像检查又能提供肿瘤代谢功能方面的信息，并且其图像可经局域网传输至放疗科治疗计划系统与CT图像匹配融合直接指导靶区勾画，因此DWI研究近年已成为肿瘤放疗学者的研究热点之一。基于以上研究背景，本课题对食管癌患者进行了前瞻性入组和回顾性分析，从DWI技术指导食管癌病变长度的确定、磁共振DWI序列高信号表现及ADC值大小对食管癌治疗反应的判断和预后评估、DWI用于食管癌近期疗效评价的界定、比较和补充三方面进行了较为系统深入的研究，以期对临床靶区勾画、疗效评价、预后评估提供有价值的参考信息。第一部分：磁共振弥散加权成像及CT扫描对确定食管癌病变长度的价值研究目的：前瞻性入组根治性手术切除的胸段食管癌患者，对各种检查方法所确定的食管癌病变长度与术后病理进行对照分析，探讨利用DWI图像确定食管癌GTV长度的准确性，为食管癌精确放疗GTV的勾画提供优选的可供参考的影像学方法。方法：2012年10月至2013年5月，35例病理证实的胸段食管癌患者接受左后外开胸食管癌根治术，患者疗前完善食管内窥镜、胸腹CT扫描及MRI平扫(T1WI、T2WI)和弥散加权成像(DWI)平扫，CT图像经局域网传输至放疗科治疗计划系统(Pinnacle)并依据CT图像勾画食管大体肿瘤区(GTV)，将DWMRI图像上高信号区界定为肿瘤范围，在轴位图像上根据病变层数计算病变长度。患者手术切除标本用10%福尔马林液固定后沿纵轴解剖手术标本固定肿瘤边界，直尺测量肿瘤最长直径。按肿瘤组织标本固定后收缩比为(90±10)%取值，回推实体状态下食管癌病变长度。用组内相关系数(ICC)法及Bland-Altman法分析各种影像学方法所测量的食管病变长度与病理标准的符合程度。结果：⑴全组患者均顺利完成常规MRI平扫(T1WI、T2WI)及弥散加权成像(DWI)平扫，有4例患者于MRI图像上未见食管管壁增厚，DWI序列未见高信号表达，表现为假阴性结果(11.43%)。⑵在31例可供分析DWI病变长度的患者中，其大体肿瘤标本长度经直尺测量为4.12±1.81cm，依据公式回推实体状态下食管病变长度为4.58±2.01cm，其内窥镜检查、CT扫描、b值=600、800、1000s/mm2条件下DWI图像所测肿瘤长度分别为4.56±1.99cm，5.58±2.15cm，4.41±1.93cm，3.99±1.95cm和3.83±1.94cm，其与回推的实体肿瘤长度的差值分别为0.07±1.27cm、1.05±1.37cm、-0.27±0.64cm、-0.69±0.92cm和-0.85±0.95cm，经Pearson相关检验，对应的r值分别为0.802，0.786，0.946，0.890和0.833，P值均为0。⑶经组内相关系数(Intraclass Correlation Coefficient—ICC)法检验，内窥镜、CT扫描、b值等于600、800、1000s/mm2条件下DWI所测病变长度与病理比较的ICC值分别为0.703、0.764、0.946、0.890和0.882，P值均为0.000。内镜用于测量食管肿瘤长度的信度仅为中等(0.4至0.75之间)，CT的信度略优，不同b值条件下DWI所测肿瘤长度值的信度均为良好(明显大于0.75)，以b=600s/mm2条件下所测的病变长度值信度最高，接近于完全可信(即ICC=1)。在Bland-Altman法的分析结果中，以内窥镜和b=600条件下DWI所测肿瘤长度与病理标准下的差值的均值最接近于0，与病理标准的一致程度最高。综合衡量ICC法与Bland-Altma法的分析结果， b=600s/mm2条件下DWI所测得结果可能为与病理标准最接近的优选的食管癌病变长度评估方法。⑷亚组分析的结果显示T1期患者9例，4例于DWI图像未见高信号表达，假阴性率为44.4%，9例患者依据大体肿瘤标本推测的实体长度为3.05±2.12cm，内窥镜测量结果为3.67±1.41cm，经ICC法检验，其ICC值为0.636，P=0.024，内窥镜用于早期食管癌病变长度测量的信度为中等；CT扫描测量T1期患者的病变长度为5.22±2.11cm，ICC值为0.492，P=0.074，CT测量早期病变长度的参考价值不大。T3、4期患者病理实体长度为4.81±1.74cm，其余5种条件下测得的病变长度分别为4.80±1.91cm、5.78±1.97cm、4.54±1.68cm、4.11±1.75cm、3.93±1.72cm，其ICC值分别为0.734、0.744、0.916、0.829、0.810，P值均为0.000，仍以DWI所测得病变长度值的信度最高。⑸不同b值条件下所测得的病变ADC值分别为1.69±0.29×10-3mm2/s，1.54±0.30mm2/s，1.47±0.28mm2/s，正常食管的ADC值分别为3.12±0.31×10-3mm2/s，2.69±0.30mm2/s，2.34±0.20mm2/s，随着b值的增加，食管肿瘤及正常组织的ADC值均逐渐降低，食管肿瘤的ADC值均明显低于邻近正常食管组织，t=-15.535，-10.249，-11.892，P=0.000，0.000,0.000。不同性别、年龄、T分期、病理类型、GTV体积组患者的ADC值大小无统计学差异，P＞0.05。结论：DWI技术所测得食管肿瘤长度与病理实体长度较为接近，符合程度较高，以b=600s/mm2条件下的病变长度测量信度最高，建议作为DWI-CT图像融合的优选图像并指导食管癌靶区勾画。DWI技术不适合早期食管癌的诊断及病变长度测量。食管癌病变区域ADC值明显小于正常食管组织，但并未发现其与临床、病理因素存在关联性。第二部分：磁共振弥散加权成像(DWI)在食管癌疗效判断及预后评估中的应用价值目的：探讨磁共振弥散加权成像(DWI)在食管癌放疗疗效判断及预后评估中的应用价值，以指导临床治疗。方法：2010年3月至2011年12月，77例食管鳞癌患者接受三维适形或调强放疗,处方剂量54～61.2Gy，单次1.8~2Gy，1次/天，5次/周，其中33例接受了FP或TP方案的同期化疗，化疗于放疗开始后的第1、5周给予，剂量为顺铂12.5mg/m2×5天或25mg/m2×3天，5-FU450～500mg/m2×5天或紫杉醇135mg/m2，第1天，顺铂25mg/m2，第2、3、4天给予。患者于放疗前、放疗末及放疗后3个月行DWI检查，观察全组患者治疗前后的ADC值及DWI序列病变高信号表达情况，与其近期疗效及1、2、3年生存率相结合进行疗效判断和预后评估的分析。结果：⑴77例患者中有3例疗前DWI检查未见高信号表达，为假阴性(3.9%)。余74例放疗末DWI高信号消失者23例，高信号随访1~3月消失者40例，高信号始终未消失者11例，前两组1、2、3年生存率分别为82.6%、56.5%、56.5%和72.5%、48.5%、34.%，第3组至随访日已全部死亡，1年生存率9.1%，生存期最长17个月，χ2=38.070，P=0.000。⑵74例可测量ADC值患者放疗前后及正常食管组织的ADC值分别为(1.64±0.48)×10-3mm2/s，(2.65±0.58)×10-3mm2/s和(3.12±0.53)×10-3mm2/s，疗后ADC值明显高于疗前，t=-16.283，P=0.000，但二者均明显低于正常食管组织ADC值，t值分别为13.737，2.773，P值分别为0.000和0.009。⑶患者疗前低ADC值组(≤1.5×10-3mm2/s)和高ADC值组(＞1.5×10-3mm2/s)，疗后CR率分别为55.9%(19/34)和70.0%(28/40)，χ2=1.581，P=0.209。治疗后的1、2、3年生存率分别为为67.6%、40.2%、34.5%和65.0%、46.8%、36.7%，χ2=0.039，P=0.843。单因素的Cox回归模型分析显示疗前ADC值大小并非预后影响因素。⑷疗后低ADC值组(≤2.6×10-3mm2/s)和高ADC值组(＞2.6×10-3mm2/s)完全缓解率分别为45.9%和81.1%，χ2=9.855，P=0.002。两组1、2、3年生存率分别为48.5%、22.7%、13.6%和84.8%、66.0%、59.4%，χ2=14.101，P=0.000。Spearman等级相关分析显示疗前ADC值大小与近期疗效之间存在关联系，r=0.434，P=0.000。单因素的Cox回归模型分析提示疗后ADC值大小为预后影响因素。⑸Cox回归模型多因素分析结果显示疗末高信号表达情况和非手术T分期两项因素为独立预后影响因素，P值分别为0.000和0.041，OR值分别为为2.911和0.621。结论磁共振弥散加权成像(DWI)检查可以对食管癌放化疗后的疗效进行判断并且能够较准确地评估预后。相较于治疗前的ADC值而言，疗末ADC值大小可能与治疗反应和预后的关系更为密切，治疗末食管病变区域的DWI高信号表达情况为独立预后影响因素，高信号经复查始终未消失提示预后不良。结论：磁共振弥散加权成像(DWI)检查可以对食管癌放化疗疗效进行判断并较准确地评估预后。相较于治疗前的ADC值而言，疗末ADC值大小可能与治疗反应和预后的关系更为密切，治疗末食管病变区域的DWI高信号表达情况为独立预后影响因素。第三部分：磁共振弥散加权成像CT扫描及钡餐造影评价食管癌原发灶近期疗效的应用研究目的：依据不同的疗效评价标准对食管癌放化疗患者进行近期疗效评价和比较，同时依据患者疗末磁共振弥散加权成像高信号表达情况对疗效进行界定，以期寻找更全面、客观、准确的疗效评价方法，从而指导临床后续治疗及预后评估。方法：2012年9月至2013年6月，36例食管鳞癌患者接受三维适形或调强放疗,处方剂量54～64Gy/27~34次，中位60Gy，单次1.8~2.1Gy，1次/天，5次/周，其中11例接受了FP或TP方案的同期化疗。所有患者放疗末复查食管钡餐造影、胸及上腹CT、磁共振弥散加权成像(DWI)。依据1989版食管癌放射治疗后近期疗效评价标准和我院2013年提出的基于钡餐造影和CT扫描的食管癌近期疗效评价标准进行疗效评价和比较，同时依据患者疗末病变区域DWI高信号表达情况进行疗效界定和亚组分析。结果：⑴依据1989版近期疗效评价标准，全组完全缓解23例(63.89%)，部分缓解13例(36.11%)，1年局控率分别为70.6%和46.7%，χ2=1.135，P=0.287，1年生存率分别为66.9%和57.7%，χ2=0.498，P=0.480。⑵全组疗前肿瘤最大管壁厚度2.24±0.89cm，疗后为1.03±0.29cm，肿瘤最大层面管壁厚度收缩率为(52.3±0.21)%。依据2013基于钡餐造影和CT扫描的食管近期疗效评价标准，完全缓解组18例(50%)，部分缓解组18例(50%)，其1年局控率分别为82.4%和40.8%，χ2=4.219，P=0.040，1年生存率分别为80.0%和49.5%，χ2=2.514，P=0.113。⑶全组放疗后依据DWI扫描结果，病变区域DWI高信号完全消失者14例，呈稍高信号表达者15例，仍有高信号表达者7例，其1年局控率分别为83.3%，60.6%和14.3%，χ2=23.448，P=0.000，疗末仍有高信号表达的7例患者中，6例发生肿瘤未控或复发。三组的1年生存率分别为69.6%，71.1%和0，χ2=6.235，P=0.044。将疗后食管病变局部无高信号表达者界定为完全缓解(14例)，有信号表达者界定为部分缓解(22例)，与基于食管钡餐造影和CT检查的近期疗效评价标准进行Kappa一致性检验，结果Kappa值为0.444，P=0.006。⑷依据钡餐造影、CT扫描、DWI检查综合评价，三项均达完全缓解标准者11例，截止至随访日局控率100%。全组钡餐造影及CT扫描显示肿瘤缩小越明显者其DWI序列高信号消失率越高，但是肿瘤形态学评价结果并不能完全反应肿瘤DWI高信号表达情况。⑸全组放疗前后食管病变及正常食管组织的ADC值分别为1.67±0.44×10-3mm2/s，2.39±0.49×10-3mm2/s和3.04±0.30×10-3mm2/s，疗后ADC值明显高于疗前ADC值，t=-7.844，P=0.000，但二者均明显低于正常食管组织ADC值，t值分别为-20.339，-10.286，P值均为0.000。将患者分为疗末高ADC值组(＞2.60×10-3mm2/s)和低ADC值组(≤2.60×10-3mm2/s)，其1年局控率分别84.7%，37.5%，χ2=10.020，P=0.002，1年生存率为80%，44.9%，χ2=3.838，P=0.050。Cox回归模型的单因素分析提示疗末ADC值大小(≤2.60×10-3mm2/s VS..＞2.60×10-3mm2/s)为肿瘤局部控制的独立预后影响因素， Wald值，P值，OR值及95%CI分别为7.099，0.008，0.114和0.023~0.563。⑹将造影评价CR，CT扫描最大管壁厚度≤1.2cm，且疗末DWI扫描高信号完全消失三者全部满足界定为近期疗效达完全缓解的优选标准及观察目标进行ADC值的ROC曲线分析，结果得到的曲线下面积为0.737，寻找到的截点值为2.55×10-3mm2/s，对应的灵敏度为81.8%，特异度为58.3%。以此为界定值进行疗效观察，疗末ADC值≤2.55×10-3mm2/s组肿瘤完全缓解率仅11.8%(2/17)，疗末ADC值＞2.55×10-3mm2/s组肿瘤完全缓解率为47.4%(9/19)，χ2=4.900，P=0.027。结论：基于钡餐造影和CT扫描的近期疗效评价标准可较好地提示食管肿瘤局部控制情况；磁共振弥散加权成像(DWI)可为食管癌治疗反应评价提供直观、量化的参考信息，疗末食管病变局部仍有高信号表达预示极高的复发风险，疗末ADC值小于2.55×10-3mm2/s为预后不良表现；依据钡餐造影、CT扫描、DWI检查综合评价可更全面、客观、准确地评估疗效，钡餐造影及CT扫描显示肿瘤缩小越明显者其DWI序列高信号消失率越高，但是肿瘤形态学评价结果并不能完全反应肿瘤组织的增殖和代谢信息，三项检查依其标准均达完全缓解者提示较高的肿瘤控制可能。更多还原

【Abstract】 The3DCRT or IMRT treatment of esophageal carcinoma had been used inclinic generally recent years, and the long term local control rates and survivalrates had improved significantly as well, but there were still quite a fewuncertain questions exited in the study of esophageal cancer, such as thedetermination of accuracy GTV(Gross Tumor Volume), reasonable extra-boundary of CTV(Clinic Target Volume), the standard comprehensivetreatment model of chemoradiotherapy, the consummation of clinical stagingfor esophageal carcinoma treated with non-surgical methods and well-definedshort-term therapeutic effect evaluation criterion, et al. Till now, thedelineation of GTV for esophageal carcinoma still must be accomplished onCT images, but as reported of the related studies, the determination forinvasion depth of tumor was not accurate on CT images especially for lesionsof T1and T2stages, the real lesion length could not be measured on CTimages accurately too. What’s more, CT scan, convensional MRI scan andesophagogram were commonly used for the evaluation of therapeutic effect,there was importment significance to evaluate the short-term therapeuticeffect(TE) objective and accurate, further regimen of treatment could be madeaccording to TE. But definitely short-term TE evaluation criterion stillabsently in current study. The examination of DWI was a new functionaliconography technology, initial study indicated that DWI scanning could beused to diagnose tumor early, clinical stages dividing and TE detection.Because conventional MRI image could manifest the tumor modality clearly,DWI image could afford metabolic information of tumor, and DWI imagecould be tranferred to TPS in radiotherapy department and used for target delineation, DWMRI study had been a hot spot of oncologist. Perspective andretrospective study were performed based on the above background, thepurpose of the series studys was to investigate the concordance of lesionlength measured by DWI and surgical specimen, the application value of DWIin short-term TE and prognosis evaluation.Part1: The Application Value Study of Diffusion-weighted MagneticResonance Imaging(DWI) and CT Scan in the Determination ofLeision Length for Esophageal carcinoma.Objective: Esophageal cancer patients who would receive radicaloperation were perspectively enrolled in this study and the lesion lengthsmeasured by endoscopy, CT scan and DWI at different b values werecompared with pathological length, the objective was to investigate theaccuracy of DWI in determination the length of GTV and affording optimizediconography method which could be referenced in the delineation of GTV foresophageal cancer.Methods: From October2012to May2013,35patients with thoracicesophageal carcinoma received radical operation, all the patients made theexaminations of endoscopy, CT scan of thorax and abdomen, conventionalMRI scan(T1WI、T2WI) and DWI scan before the operation, CT images weretransferred to treatment planning systems(TPS) through local area networkand esophageal gross tumor volume(GTV) were delineated based on CTimages. The region with hyperintense expression on DWI image were definedto the tumor and lesion length were calculated according to the amout sliceswith tumor on axial images. The tumor border were immobilized alonglongitudinal axis after surgical specimen were immobilized by10%formalinand the maximum long diameter of tumor were measured by ruler. Value of(90±10)%was adopted as shrinkage ratio of tumor specimen afterimmobilization, and the real tumor length were calculated by the formula oftumor surgical specimen length/0.9. Mehtod of Intraclass Correlation Coefficient(ICC) and Bland-Altman were used to analyzed the consistency oflesion lengths measured by iconography methods and pathological length.Results:⑴All the patients accomplished the examination ofconventional MRI scan(T1WI、T2WI) and DWI scan, there were4patientshad no esophagus wall thickening on MRI image and had no hyperintenseexpression on DWI image, the false-negative rate of DWI scan was11.4%(4/35).⑵There were31patients who had lesion length measured fromDWI, the length of tumor specimen calculated by ruler was4.12±1.81cm, thecomputative real tumor length was4.58±2.01cm, the lesion length of31patients measured from endoscopy, CT scan, DWI of b=600,800,1000s/mm2were4.56±1.99cm,5.58±2.15cm,4.41±1.93cm,3.99±1.95cm and3.83±1.94cm, respectively, the difference value compared with real tumor lengthwere0.07±1.27cm,1.05±1.37cm,-0.27±0.64cm,-0.69±0.92cm and-0.85±0.95cm. According to Pearson correlatived test, r=0.802,0.786,0.946,0.890and0.833, respectively, P=0.000.⑶According to test of ICC method,the ICC values of lesion lengths measured from endoscopy, CT scan, DWI ofb=600,800,1000s/mm2comparing with real tumor length were0.703,0.764,0.946,0.890and0.882, respectively, P=0.000for all. The reliability of tumorlength measured by endoscopy was only moderately(0.4~.075), the reliabilityof CT was slightly better, and that of DWI at differrent b values wassatisfactory(much greater than0.75). The reliability of lesion length measuredby DWI of b=600s/mm2was maximum and nearly closed to absolutelybilievable(ICC=1). In the results of Bland-Altman test, the mean value ofdifference calculated from endoscopy and DWI of b=600s/mm2compared topathological length was most close to0, which indicated that the resultsmeasured by endoscopy and DWI of b=600s/mm2had a higher concordancewith pathological results. Basing on the comprehensive weight of ICC methodand Bland-Altman menthod, tumor length measured by DWI of b=600s/mm2may be more close to pathogogical result.⑷The results of sub-group analysisshowed that4of9patients of T1stage had no hyperintense expression onDWI image, the false-negative rate was44.4%. The real tumor length of T1 stage was3.05±2.12cm, the resut measured by endoscopy was3.67±1.41cm, the reliability was moderate according to test of ICC method,ICC=0.636, P=0.024. The lesion length of T1stage measured by CT scan was5.22±2.11cm, ICC=0.492, P=0.074, the reference value was not so high. Thereal tumor length of T3and T4stages was4.81±1.74cm, the lesion lengthsmeasure by iconography method were4.80±1.91cm,5.78±1.97cm,4.54±1.68cm,4.11±1.75cm and3.93±1.72cm, respectively, the correspondingICC values were0.734,0.744,0.916,0.829and0.810, respectively, P=0.000for all, the reliability of DWI method was better.⑸The ADC values of tumorat different b values (600,800,1000s/mm2) were (1.69±0.29)×10-3mm2/s,(1.54±0.30) mm2/s,(1.47±0.28)mm2/s, respectively, the correspondingnormal esophageal tissues were(3.12±0.31)×10-3mm2/s,(2.69±0.30) mm2/s,(2.34±0.20)mm2/s, respectively. The higher of b value, the lower of ADCvalue of tumor and normal esophageal tissues. The ADC values of tumor weresignificantly lower than that of normal esophageal tissues according topared-samples T test, t=-15.535,-10.249,-11.892, respectively, P=0.000,0.000,0.000, respectively. There was no significant differrence of ADC valueamong sub-groups of different genders, ages, T stages, pathological categoriesand GTV volumes, P＞0.05.Conclusions: The esophageal tumor lengths measured by DWI scan wasclosely to real tumor length based on surgical specimen and had a highconcordance with pathology. The results measured by DWI scan of b=600s/mm2had the highest reliability and was a optimized reference for tumordelineation. However, the technology of DWI scan did not fit for the diagnosisand lesion length measurement of pristine esophageal carcinoma. The ADCvalue of esophageal cancer was significantly lower than that of normalesophageal tissues, but no relevance was observed between clinical,pathological parameters and ADC values. Part2: The Investigation of Using Diffusion-weighted MagneticResonance Imaging(DWI) to Evaluate Therapeutic Effect ofEsophageal Carcinoma Treated with3DCRT or IMRT.Objective: To investigate the value of DWI in the evaluation oftherapeutic effect of esophageal carcinoma treatment with3DCRT or IMRT,and to direct the clinic treatment program or regimen selection.Methods: From March2010to December2011,77patients withesophageal carcinoma received3DCRT or IMRT treatment. The prescribeddoses were ranged from54Gy-61.2Gy with median dose of60Gy andconventional fraction，33of them received concurrent chemotherapy of FP orTP regimen. All the patients received examination of Diffusion-weightedMagnetic Resonance Imaging(DWI) before, at the end of and3months afterradiotherapy. Therapeutic effect was evaluated by esophagogram, the status ofhyperintense expression in DWI and the value of apparent diffusioncoefficient(ADC). Prognosis analysis was performed associated with theabove facors.Results:⑴There were3patients diagnosed by pathology had nohyperintense expression in the examination of DWI, the false negative ratewas3.90%. In the other74patients, when finished radiotherapy the DWIhyperintense disappeared immediately was23patients, disappeared in3months after radiation was40patients, did not disappeared all the time was11patitents, the1-,2-,3-year survival rates of the anterior two groups were82.6%,56.5%,56.5%and72.5%,48.5%,34.5%， respectively, the1-yearsurvival rate of the third group only was9.1%, all the11patients had died tillthe date of follow up, χ2=38.070，P=0.000。⑵The ADC values of esophagealcarcinoma before radiotherapy,at the end of radiotherapy and normalesophageal tissues were1.64±0.48×10-3mm2/s,2.65±0.58×10-3mm2/s and3.12±0.53×10-3mm2/s，respectively, the ADC value of post-radiotherapy washigher than that of pre-radiotherapy, but they were both lower than that ofnormal esophageal tissues, t=13.737,2.773, and P=0.000,0.009.⑶The CR rates of low value ADC group(≤1.5×10-3mm2/s) and high value ADCgroup(＞1.5×10-3mm2/s) before radiotherapy were55.9%(19/34)and70.0%(28/40), respectively, χ2=1.581, P=0.209. The1-,2-,3-year survivalrates of the two groups were67.6%,40.2%,34.5%and65.0%,46.8%,36.7%,and no existed the significant difference, χ2=0.039, P=0.843. Univariateanalysis of Cox regression indicated the ADC value of pre-radiotherapy wasnot a significant prognosis factor.⑷The CR rates of low value ADCgroup(≤2.6×10-3mm2/s) and high value ADC group(＞2.6×10-3mm2/s) afterradiotherapy were45.9%and81.1%, respectively, χ2=9.855，P=0.002。Therewas a correlathion relationships between the ADC value and therapeuticeffects according to Spearman Bivariate Correlations analysis, r=0.434,P=0.000. The1-,2-,3-year survival rates of the two groups were48.5%,22.7%,13.6%and84.8%,66.0%,5.4%, respectively, survived significantdifference between the two groups, χ2=14.101, P=0.000. Univariate analysisof Cox regression further indicated the ADC value of post-radiotherapy was asignificant prognosis factor.⑸Multivariate analysis of Cox regressionindicated that hyperintense expression of DWI after radiotherapy andnon-surgical T stages were independent prognosis factors with P values of0.000,0.041, OR values of2.911,0.621.Conclusions: The examination of DWI could accurately evaluatetherapeutic effect and prognosis of esophageal carcinoma treated with3DCRTor IMRT. Compared to ADC value of pre-radiotherapy, the ADC value afterradiotherapy had more closed relationships with therapeutic effect andsurvival rate. The DWI hyperintense expression of esophageal carcinoma afterradiotherapy was an independent prognosis factor, and the undisappear ofhyperintense on DWI sequence indicated a poor prognosis. Part3: The Investigation of Using Diffusion-weighted MagneticResonance Imaging(DWI), CT Scan and Esophagogram toEvaluate the Therapeutic Effect of Esophageal PrimaryCarcinoma Treatment with3DCRT or IMRT.Objective: To compare the therapeutic effect(TE) of esophageal primarycancer evaluated by the examination of DWI, CT scan and esophagram, findmor objective and accurate TE evaluation method and therefore direct theclinical further treatment.Methods: From September2012to June2013,36patients withesophageal squamous carcinoma received3DCRT or IMRT treatment. Theprescribed doses were ranged from54Gy-64Gy with median dose of60Gy andconventional fraction，11of them received concurrent chemotherapy of FP orTP. All the patients performed the examinations of DWI, CT scan andesophagogram at the end of treatment. TE was evaluated by short-termtherapeutic effect evaluation criterion of1989version and2013version andhyperintense expression on DWI sequence.Results:⑴According to the short-term therapeutic effect evaluationcreterion of1989version,23patients achieved complete remission(CR) aftertreatment(63.89%),13achieved partly remission(PR)(36.11%), the1-yearlocal control rates of the two groups were70.6%and46.7%, respectively,χ2=1.135, P=0.287, the1-year survival rates were66.9%and57.7%,respectively, χ2=0.498, P=0.480.⑵The maximum tumor wall thicknessesof pre-radiotherapy(RT) and post-RT were2.24±0.89cm and1.03±0.29cm,the contraction rate of tumor wall thickness was (52.3±0.21)%. According tothe TE eavluation criterion of2013version which based the examination ofesophagogram and CT scan,18patients achieved CR(50.0%) and18achievedPR(50.0%), the1-year local control rates were82.4%and40.8%, respectively,χ2=4.219, P=0.040, the1-yeat survival rates were80.0%and49.5%,respectively, χ2=2.514, P=0.113.⑶According to the examination of DWI,14patients’ hyperintense disappeared completely at the end of radiotherapy,15 patients’ had a slightly hyperintense expression and7still had hyperintenseexpression on DWI sequence, the1-year local control rates were83.3%,60.6%and14.3%, respectively, χ2=23.448, P=0.000. In the group of still havinghyperintense expression(7pts), tumor control failure happened in6patients.The1-year survival rates of the three groups were69.6%,71.1%and0,respectively, χ2=6.235, P=0.044. The group whose hyperintense disappearedcompleted of DWI was defined to CR, and the others were defined to PR, theTE results evaluated by DWI and TE evaluation criterion of2013version werecompared according to Kappa test, as a result, the Kappa corfficient was0.444,P=0.006.⑷According to the examination of esophagogram, CT scan andDWI,11patients acieved CR in all exams, the local control rate was100%forthis group. The more obvious of tumor contraction according to esophagogramand CT scann, the more patients’ hyperintense disappeared completely aftertreatment, but the tumor change of morphology could not react thehyperintense expression on DWI sequence.⑸The ADC value of pre-RT,post-RT and normal esophageal tissues were1.67±0.44×10-3mm2/s,2.39±0.49×10-3mm2/s and3.04±0.30×10-3mm2/s, respectively, the ADC valueof post-RT was higher than that of pre-RT, t=-7.844, P=0.000, but the twoformers were both lower than that of normal esophageal tissues, t=-20.339,-10.286, P=0.000,0.000. The1-year local control rates of high ADC group(＞2.60×10-3mm2/s) and low ADC group(≤2.60×10-3mm2/s) were84.7%,37.5%, respectively, χ2=10.020, P=0.002, the1-year survival rates of thetwo groups were80.0%,44.9%, respectively, χ2=3.838, P=0.050. Univariatanalysis of Cox regression model indicated that ADC value of post-RT was ainfluence factor of local control, Wald=7.099, P=0.008, OR=0.114, and withthe95%CI of0.023to0.563.⑹The status of CR evaluated by esophagogram,CT scan and DWI was made a observe event in ROC(Receiver Operatingcharacteristic Curve) analysis, as a result, the area under the cure was0.737,the finded cut-off point was2.55×10-3mm2/s, with the correspondingsensitivity of81.8%, specificity of58.3%. For the group of ADC value≤2.55×10-3mm2/s after treatment, the CR rate of primary was only11.8%(2/17), for the group of ADC value＞2.55×10-3mm2/s, the CR rate was47.4%(9/19),χ2=4.900, P=0.027.Conclusions: The esophagogram and CT based TE evaluation criterioncould indicate local control status of esophageal cancer well, the examinationof DWI could afford visualized and quantifying reference information aboutthe TE of esophageal cancer, the expression of hyperintense at the end of RTindicated a high risk of recurrence, the post-RT ADC value of less than2.55×10-3mm2/s suggested a worse prognosis. The therapeutic effect evaluated byesophagogram, CT scan and DWI maybe more objective and more accurate,the therapeutic effect of CR evaluated congenerously by esophagogram, CTscan and DWI indicated a high local control probability.更多还原