【作者】 许茜；

【导师】 祝淑钗；

【作者基本信息】 河北医科大学 ， 肿瘤学， 2009， 博士

【摘要】 食管癌是发病率、死亡率均较高的一种恶性肿瘤,就诊患者中绝大多数为病期晚、预后差的病例。外科治疗和放射治疗是目前食管癌治疗的主要手段,虽然在治疗方法上不断取得新的进步,但总的5年生存率仍较低。准确的治疗前分期是制定合理的个体化治疗方案,提高治疗效果及评估预后的重要前提。目前,食管癌非手术分期方法多种多样,繁简各异,标准不统一,其原因即是没有一个理想的分期标准。食管癌的病理分期只考虑病变的外侵深度,适用于手术治疗的患者。而对于不适合手术的食管癌患者来说,除浸润深度外,肿瘤的体积大小也直接影响到放、化疗的治疗效果。本研究即是从肿瘤体积的角度来探讨其与病理分期及预后的关系,旨在找出其内在规律,为治疗前分期及评估预后提供依据。本研究首先用现有临床分期标准分析607例根治性切除食管癌患者的术前分期状况,与术后病理分期对照,评价现有临床分期标准在T分期、N分期及TNM分期上的准确性。然后,将该组病人术前的CT图像资料传输到三维放疗计划系统中,勾画出食管癌肿瘤的GTV-T靶区,计算出体积,分析(1)不同病理T分期GTV-T体积是否存在差异;(2)GTV-T体积与生存率的关系。最后在此基础上,寻找GTV-T分布规律,划分等级,制定新的临床分期方案;按照新方案重新给该组病人分期,并评价其与病理分期对照的准确性和一致性,及对预后生存的影响。并应用Cox逐步回归模型分析食管癌切除术后的预后影响因素。第一部分现有食管癌临床分期与病理分期的对照研究目的:分析现有食管癌临床分期标准与病理分期对照的准确性、符合程度及与预后的关系。方法:回顾性分析607例根治性切除的胸段食管癌患者的术前CT图像及钡餐造影图像,测量并纪录CT图像上病变最大层面的直径及钡餐造影片上显示的病变长度,观察并纪录CT图像上食管肿瘤与周围组织器官的关系(包括气管是否受侵、椎前脂肪间隙是否消失、病变与主动脉的夹角),锁骨上、纵隔和腹腔淋巴结肿大情况,以及术前各项检查所示其他组织脏器的情况。根据现有食管癌临床分期标准,对该组病例进行分期,并与病理分期进行对照研究,观察不同分期下的生存情况。结果:(1)607例患者中30例失访,随访率为95.06%,自手术之日起计算的术后1、3、5年生存率分别为83.84%、53.49%、36.42%,平均生存期42.46个月,中位生存期38.50个月。(2)依据CT图像诊断气管或主支气管受侵的准确性为90.12%,灵敏度为69.33%,特异度为92.08%,阳性预测值28.13%,阴性预测值98.53%。CT诊断胸主动脉受侵的准确性为84.84%,灵敏度为76.47%,特异度为85.08%,阳性预测值12.87%,阴性预测值99.21%。(3)术前CT示食管病变最大直径按≤2.0cm、2.1cm～3.0cm、>3.0cm分为三组,其5年生存率比较差异有显著性意义,x2=18.37,P=0.0001。术前钡餐造影示病变长度按≤3.0cm、3.1cm～7.0cm、>7.0cm分为三组,该三组的5年生存率比较差异有显著性意义,x2=17.02,P=0.0002。(4)607例食管癌临床T分期与术后病理T分期的符合率为36.41%。一致性分析显示,两者有一致性(P<0.001),但一致性较差(Kappa值为0.10)。临床分期T1、T2、T3、T4期的5年生存率比较有显著性差异(x2=13.34, P=0.004)。(5)该组食管癌临床N分期与术后病理N分期的符合率为62.44%,两者存在一致性(P<0.001),Kappa值为0.30。临床分期N0、N1、N2期的5年生存率比较亦有显著性差异(x2=17.94, P=0.0001)。(6)607例食管癌临床TNM分期与术后病理TNM分期的符合率为60.63%,两者存在一致性(P<0.001),Kappa值为0.29。非手术术前分期Ⅰ、Ⅱ、Ⅲ、Ⅳ期的5年生存率比较亦有显著性差异(x2=22.19, P=0.0001)。结论:(1)依据CT图像判断食管病变对气管或主支气管、胸主动脉受侵的准确性较高。(2)食管癌切除术后的5年生存率随着肿瘤直径、长度的增大而降低。(3)该食管癌术前临床分期与术后病理分期存在一致性,但一致性较差,尚需进一步完善。(4)该临床分期能较好地反映食管癌的预后。第二部分食管癌三维适形放疗GTV-T体积与病理T分期及预后的关系目的:探讨不同病理T分期GTV-T体积是否存在差异,及GTV-T体积与生存率的关系,初步制定出GTV-T体积的分级标准。方法:将607例食管癌患者术前的CT图像传输到三维放疗计划系统中,勾画出食管癌肿瘤的GTV-T靶区,计算出GTV-T体积。分析不同病理T分期GTV-T体积是否存在差异。以病理T1、T2、T3、T4各期的GTV-T体积平均值为依据,同时兼顾生存率曲线的分离度,选择出合适的GTV-T体积分级标准。结果:(1)食管癌GTV-T最大直径、长度和体积均与病理T分期呈正相关关系,Spearman相关系数r分别为0.466、0.376和0.464,P<0.001。除病理T3期与T4期的GTV-T最大直径、长度和体积平均值未见统计学差异(P>0.05)外,其它各病理T分期组间上述指标均有显著性差异(P<0.001)。(2)食管癌GTV-T长度与病变最大直径呈正相关关系,r =0.749,P<0.001。(3)依据病理T分期将GTV-T体积按≤5cm3,>5cm3～≤13cm3,>13cm3分为三级,与病理T1、T2、T3+4期进行对照,一致率达73.81%,两者存在一致性(P<0.001),且一致性较好,Kappa值为0.40,其5年生存率曲线分离度较好(x2=26.10, P<0.001)。(4)综合考虑预后生存情况将GTV-T体积分为四级:≤5cm3,>5cm3～≤13cm3,>13cm3～≤39cm3,>39cm3。该分级与病理T分期的总符合率为54.70%,两者存在一致性,但一致性较差,Kappa值为0.24;各分级间5年生存率比较均有统计学意义,P<0.05。(5)食管癌GTV-T三分级、四分级与术后病理淋巴结转移数均呈正相关,相关系数r值分别为0.209、0.230,P<0.001。GTV-T三分级、四分级与CT扫描显示的肿大淋巴结数目亦均呈正相关,r值分别为0.155、0.209,P<0.001。结论:(1)食管癌GTV-T最大直径、长度和体积均与病理T分期呈正相关关系,三者各自的平均值除在病理T3期和T4期间未见统计学差异外,在其它各病理T分期组间均有显著性差异。(2)按≤5cm3,>5cm3～≤13cm3,>13cm3将GTV-T体积分为三级,与病理T1、T2、T3+4期有较好的一致性,Kappa值为0.40。(3)将食管癌GTV-T按≤5cm3,>5cm3～≤13cm3,>13cm3～≤39cm3,>39cm3分为四分级,与病理T分期一致性稍差,Kappa值为0.24。(4)食管癌GTV-T三分级及四分级均能较好地反映预后生存。(5)食管癌GTV-T三分级、四分级与术后病理淋巴结转移数和CT扫描显示的肿大淋巴结数目均呈正相关关系。第三部分基于食管癌GTV-T体积的临床分期与病理分期及预后的关系目的:探讨基于食管癌GTV-T体积的临床TNM分期与病理TNM分期的符合程度,及与预后的关系。方法:按照GTV-T体积分级标准制定临床T分期标准,并结合区域淋巴结转移及远隔转移对607例食管癌病例重新进行分期。评价新分期与术后病理分期的符合程度,及与预后生存的关系。结果:(1)基于食管癌GTV-T体积的临床分期标准:T分期:T1期GTV体积≤5cm3,T2期5cm3<GTV-T体积≤13cm3,T3期13cm3<GTV-T体积≤39cm3,T4期GTV-T体积>39cm3。N分期:N0无区域淋巴结转移,N1有区域淋巴结转移。M分期:M0无远处转移,M1有远处转移。TNM分期标准:Ⅰ期T1N0M0,Ⅱa期T2N0M0 / T3N0M0,Ⅱb期T1N1M0 / T2N1M0,Ⅲ期T3N1M0 / T4任何NM0,Ⅳ期任何T任何N M1。(2)基于食管癌GTV-T体积的临床分期与病理分期对照:其T分期的符合程度见第二部分,N分期的符合率为66.56%,Kappa值为0.30,P<0.001;M分期的符合率为97.20%,Kappa值为0.55,P<0.001;TNM分期的符合率为62.93%,Kappa值为0.31,P<0.001。(3)基于食管癌GTV-T体积的T、N、M分期的各期之间5年内生存率曲线均有显著性差异(P<0.05)。GTV-TNM分期中Ⅱa期与Ⅱb期、Ⅲ期与Ⅳ期之间的5年生存率未见统计学差异,P>0.05,其余各期之间的5年生存率比较差异均有统计学意义,P<0.05。(4)术后病理T分期中只有T2期与T3期的5年生存率未见统计学差异(P = 0.36),其余各期的5年生存率比较均有显著性差异。术后病理N0期与N1期的5年生存率比较有显著性差异。而术后病理M0期与M1期的5年生存率比较未见统计学差异(P = 0.097)。术后病理TNM分期中Ⅱb期与Ⅲ期、Ⅳ期以及Ⅲ期与Ⅳ期之间的5年生存率未见统计学差异,P>0.05,其余各期之间的5年生存率比较差异均有统计学意义,P<0.05。(5)影响食管癌预后的单因素分析显示:病理类型、分化程度、钡餐示病变长度、CT示病变最大直径、病理T分期、病理N分期、病理TNM分期、GTV-T分期、GTV-N分期、GTV-M分期、GTV临床分期对预后都有影响,统计学上有显著性意义(P<0.05)。Cox多因素分析显示年龄、病变部位、病理类型、病理N分期、GTV-T分期为食管癌根治术后的独立预后因素(P<0.05)。结论:(1)基于食管癌GTV-T体积的临床分期与术后病理分期存在一致性,其中M分期的一致性最好,T分期的一致性亦较好。(2)基于食管癌GTV-T体积的临床分期能较好地预示食管癌的预后。其中,GTV-T分期和M分期对预后的判断可能优于病理T分期和M分期。(3)经Cox多因素分析显示GTV-T分期、病理N分期、病变部位、病理类型、年龄为食管癌根治术后的独立预后因素。(4)基于食管癌GTV-T体积的临床分期是依据病理分期制定的一套简单、实用的分期系统,适于临床医师应用。更多还原

【Abstract】 Esophageal carcinoma is a malignant tumor with high incidence and mortality. Most of patients suffered from advanced stage and bad prognosis. Surgery and radiotherapy are the main treatment methods of esophageal carcinoma. Although the methods in the treatment have made continuously progress, overall 5-year survival is still low. Accurately staging before treatment is an important prerequisite for developing reasonable individual treatment, improving the therapeutic effect and assessment of prognosis. At present, the non-surgical staging methods of esophageal carcinoma are various, and have no unified standards.Pathological staging of esophageal cancer only considered the depth of lesions invasion. It fits patients cured by surgical operation. In addition to the depth of invasion, the tumor size have a direct impact on radiotherapy and chemotherapy treatment to patients who not able for surgery of esophageal cancer. In this study, the relationship between tumor volume and pathological stage and prognosis was explored. Aims to identify the internal rules of pre-treatment staging and provide the basis for assessment of prognosis.Firstly, this study analysis preoperative staging of the 607 cases of radical resection in patients with esophageal cancer used the current clinical staging method. Evaluate the accuracy of the current standards in T stages, N stages and TNM stages compared with postoperative pathological staging. Secondly, transfer preoperative CT image data of this group of patients to the 3D radiotherapy planning system. Outline esophageal tumor GTV-T target, measure the volume of target and analyze (1) whether there are differences of GTV-T volume in different pathologic T stages; (2) relationship between GTV-T volume and survival rate. Finally, look for distribution of GTV-T, plot clinical staging, and develop a new clinical staging program. Re-phase this group of patients under the new program, and evaluate its accuracy and consistency compare with the pathological staging, and the impact on the prognosis for survival. Study factors that influencing survival in patients with cancer of the esophageal after resection by Cox regression model.PartⅠComparison of current clinical staging with pathological staging of esophageal carcinomaObjective:Explore the accuracy and consistency of current clinical staging compare with pathological staging of esophageal cancer, and relationship with prognosis.Methods:Retrospective analysis of 607 cases of radical resection of thoracic esophageal carcinoma in patients with pre-operative CT images and barium esophagogram. Measure and record the maximum diameter of lesion in CT images and the length of lesion in barium esophagogram. Observe and record the invasion of esophageal tumor to the surrounding tissues and organs in CT images (Including whether the tracheal was involved, whether the prevertebral space was disappeared and the angle between lesions and aorta). Observe the situation of supraclavicular, mediastinal and abdominal lymph nodes. Record the cases of other organizations shown in preoperative examination. Plot this group of cases in stages using current clinical staging of esophageal cancer, compare with the pathological staging, and observe the survival of different clinical stages.Results:(1) 30 cases were out of contact among these 607 cases. The follow-up rate was 95.15%. The 1, 3, 5-years survival rates since the surgery date were 83.32%,53.33%,36.02% respectively. The average survival time and median survival time were 42.26 months and 38.33 months. (2) Based on CT images, the diagnosis of tracheal involvement for the accuracy of 90.12%, the sensitivity of 69.33%,the specificity of 92.08%,the positive predictive value of 28.13%,the negative predictive value of 98.53%. Diagnosis of thoracic aorta involvement in CT for the accuracy of 84.84%,the sensitivity of 76.47%,the specificity of 85.08%,the positive predictive value of 12.87%,,the negative predictive value of 99.21%. (3) Divided the largest diameter of esophageal lesions in preoperative CT into three groups:≤2.0cm、2.1cm～3.0cm、>3.0cm. Its 5-year survival rate difference was significant, x2=18.37,P=0.0001. Divided the lesion length in pre-operative barium esophagogram into three groups:≤3.0cm、3.1cm～7.0cm、>7.0cm. Its 5-year survival rate difference was significant, x2=17.02,P=0.0002. (4) The 607 cases clinical T staging of esophageal cancer were concordant with postoperative pathological T staging in 36.41%. Consistency analysis shows that,there was consistency between the clinical T staging and the pathological T staging (P<0.001),but less consistency (Kappa = 0.10). Clinical staging T1, T2, T3, T4 period compared 5-year survival rate were dramatically different (x2=13.34, P=0.004). (5) This esophageal groups clinical N staging were concordant with postoperative pathological N staging in 62.44%,there is consistency between the two (P<0.001, Kappa = 0.30). Clinical staging N0、N1、N2 period compared 5-year survival rate were significantly different (x2=17.94, P=0.0001). (6) The 607 cases clinical TNM staging of esophageal cancer were concordant with postoperative pathological TNM staging in 60.63%,there is consistency between the two (P<0.001, Kappa = 0.29). Clinical stagingⅠ、Ⅱ、Ⅲ、Ⅳperiod compared 5-year survival rate also were dramatically different(x2=22.19, P=0.0001)。Conclusions:(1) The accuracy is higher to determine whether the tracheal and aorta were involved by esophageal lesions based on CT images. (2) After resection of esophageal cancer 5-year survival rate reduces with tumor diameter and length increasing. (3) There was consistency between the preoperative clinical staging and postoperative pathological staging, however, less consistency, subject to further refinement. (4) This clinical staging can better reflect the prognosis of esophageal cancer.PartⅡRelationship among three-dimensional conformal radiotherapy GTV-T size and pathological T staging and prognosis of esophageal carcinoma Objective:To explore whether there are differences of GTV-T volume among different pathologic T stages. Analyze the relationship between GTV-T volume and survival rate. We preliminarily work out the volume of GTV-T standard classification.Methods:The CT scans of 607 cases of esophageal carcinoma can be transmitted to the three-dimensional treatment planning system by the network at digital format. Esophageal GTV-T targets were outlined and their volumes of GTV-T were measured. To analysis that whether there are differences of the GTV-T volume among different pathological T stages. Take into account the average volume of GTV-T in different pathological T stages and the separation of survival curve at the same time. Select a suitable classification standard of GTV-T volume.Results:(1) Esophageal carcinoma GTV-T maximum diameter, length and volume are related to pathological T staging with a positive correlation, Spearman correlation coefficient (r) is 0.466、0.376 and 0.464 respectively,P<0.001. In addition to the maximum diameter, length and volume of GTV-T in pathological T3 and T4 period has no significant difference (P> 0.05), other pathological T stages of these indicators are inter-group significant difference (P<0.001). (2) There is a positive correlation between GTV-T length and maximum diameter of esophageal lesion, r =0.749,P<0.001. (3) Divided the GTV-T volume into there grades:≤5cm3,>5cm3～≤13cm3,>13cm3,compared with pathology T1、T2、T3+4 stage,the coincidence rate is 73.81%. Consistency between the GTV-T volume level and pathology staging is good (P<0.001, Kappa = 0.40). Its 5-year survival rate difference was significant, x2=26.10, P<0.001. (4) Considering the prognosis of survival the volume of GTV-T is divided into four grades:≤5cm3,>5cm3～≤13cm3,>13cm3～≤39cm3,>39cm3. The coincidence rate of GTV-T volume grade and pathology staging is 54.70%,consistency between the two is poor, Kappa = 0.24. The 5-year survival rate compared in different grades was statistically significant, P<0.05. (5) GTV-TⅢandⅣgrading of esophageal cancer has a positive correlation with the number of pathological lymph node metastasis, Spearman correlation coefficient (r) is 0.209 and 0.230,P<0.001. GTV-TⅢandⅣgrading of esophageal cancer has a positive correlation with the number of swollen lymph nodes showed by CT scan, r is 0.155 and 0.209,P<0.001。Conclusions:(1) GTV-T maximum diameter, length and volume has a positive correlation with pathological T staging. In addition to during pathological T3 and T4 stage the average of GTV-T maximum diameter, length and volume has no significant difference, in other pathological T staging groups were significantly different. (2) Divided GTV-T volume into three grades:≤5cm3,>5cm3～≤13cm3,>13cm3,the coincidence rate is good with pathological T1、T2、T3+4 stage,Kappa = 0.40. (3) Divided GTV-T volume into four grades:≤5cm3,>5cm3～≤13cm3,>13cm3～≤39cm3,>39cm3,the coincidence rate is poor with pathological T staging, Kappa = 0.24. (4) GTV-TⅢandⅣgrading of esophageal cancer can better reflect the prognosis of survival. (5) GTV-TⅢandⅣgrading of esophageal cancer has a positive correlation with the number of pathological lymph node metastasis and the number of swollen lymph nodes showed by CT scan.PartⅢRelationship among clinical staging based on GTV-T volume and pathological staging and prognosis of esophageal carcinomaObjective:To explore the coincidence of clinical TNM staging based on GTV-T volume and pathological TNM staging, the relation of clinical staging and prognosis.Methods:The clinical staging standards are formulated according to the volume of GTV-T. Re-phase 607 esophageal carcinoma patients under the new staging program, considering local lymph node metastasis and distant metastasis. The coincidence between new staging and postoperative pathological staging was evaluated. The relation between new staging and prognosis was analyzed.Results:(1) Clinical staging standards based on GTV-T volume:T-staging:T1 GTV-T≤5cm3,T2 5cm3<GTV-T≤13cm3,T3 13cm3<GTV-T≤39cm3,T4 GTV-T>39cm3。N-staging:N0 no local lymph node metastasis,N1 had local lymph node metastasis. M-staging:M0 no distant metastasis,M1 had distant metastasis. TNM-staging:ⅠT1N0M0,Ⅱa T2N0M0 / T3N0M0,Ⅱb T1N1M0 / T2N1M0,ⅢT3N1M0 / T4 any N M0,Ⅳany T any N M1. (2) Comparison of clinical staging based on GTV-T with pathological staging: The coincidence of T staging was in partⅡ.,the coincidence rate of N-staging is 66.56%,Kappa = 0.30,P<0.001;the coincidence rate of M-staging is 97.20%,Kappa = 0.55,P<0.001;the coincidence rate of TNM-staging is 62.93%,Kappa = 0.31,P<0.001. (3) 5-year survival rate difference was significant in different T、N、M staging based on GTV-T volume (P<0.05). In GTV-TNM staging, there was no significant difference betweenⅡa andⅡb stage,ⅢandⅣstage of 5-year survival, P>0.05. Significant difference was existent in other stages of 5-year survival, P<0.05. (4) There was no significant difference between pathological T2 and T3 stage of 5-year survival (P = 0.36). Significant difference was existent in other pathological T stages and pathological N staging of 5-year survival. But no significant difference was showed in pathological M staging of 5-year survival (P = 0.097). In postoperative pathological TNM staging, no significant difference was showed betweenⅡb withⅢ/ⅣandⅢwithⅣof 5-year survival,P>0.05. Significant difference was existent in other pathological TNM stages of 5-year survival,P<0.05. (5) The single factor analysis of impact on the prognosis of esophageal cancer showed that the impacting factors are:pathological type, differentiation degree, the length of lesion in barium esophagogram, the maximum diameter of lesion in CT images, pathological T staging, pathological N staging, pathological TNM staging, GTV-T staging, GTV-N staging, GTV-M staging, GTV clinical staging (P<0.05). Cox regression model showed that the independent prognostic factors are: age, lesion region, pathological type, pathological N staging, GTV-T staging (P<0.05).Conclusions:(1) There is coincidence between clinical staging based on GTV-T and postoperative pathological staging of esophageal carcinoma. The consistency of M staging was the best, T staging was better. (2) The clinical staging based on GTV-T staging can better reflect the prognosis of esophageal cancer. GTV-T stage and M stage to determine the prognosis may be better than the pathological T stage and M stage. (3) Cox regression model showed that the independent prognostic factors are age, lesion region, pathological type, pathological N staging, GTV-T staging. (4) The clinical staging based on GTV-T volume of esophageal cancer is a simple and practical staging system based on pathological stage, suitable for clinical application.更多还原