节点文献
食管癌预后和单核苷酸多态性的关系
The Relations between Prognosis of Esophageal Cancer and SNPs
【作者】 袁翎;
【导师】 王立东;
【作者基本信息】 郑州大学 , 内科学, 2011, 博士
【摘要】 1.研究背景食管癌是世界上最常见的六大恶性肿瘤之一,其发病率位于全世界恶性肿瘤发病率的第八位,死亡率居全球癌症死亡率的第六位。我国是世界上食管癌发病率和死亡率最高的国家,在中国最常见的十大恶性肿瘤死亡率中,食管癌占第四位。中国每年食管癌发病和死亡人数占全世界的一半以上食管癌预后很差,中晚期食管癌的自然生存7-8个月,治疗后的生存情况也很不令人满意,5年生存率在20%左右,虽然治疗方式有了很大的改进,然而其5年生存率几十年来未见明显提高。“三早”工作已引起普遍重视,但食管癌患者就诊时80-90%属中晚期是目前的现实,提高这一大组病人的生存自然也就成为了改善食管癌肿瘤防治现状的一个决定性环节。中晚期患者生存普遍较差,但其中经治疗干预后仍有20%能长期生存,这就提出了一个科学问题,是什么原因造成了相似人群,却有不同的预后?无论在发病机制、自然生存、临床预后等方面,食管癌都呈现出明显的个体化特征。是环境因素的作用,还是先天遗传的差异?比较一致的看法是遗传和环境相互影响共同作用的结果,但二者的作用,孰轻孰重,以及所占比重尚不完全清楚。预后研究历来都是广大科研和医务工作者关注的焦点,寻找预后的内在机制、建立预后客观的评价体系,进而能够对预后风险及早进行干预,对改善患者的生存、延长病人的生命,具有现实而重要的意义。既往的研究多集中在外界环境、表观症状、临床病理特征等相关因素和预后的关系,如TNM分期、生活习惯、生存环境、治疗方式等。已取得了一些共识和成果,但还存在很多的争议和疑问,离临床实际工作要求尚有很大距离,足可见该问题的复杂性。“所有医学都是遗传学”,任何疾病的发生、发展、转归及临床表型的背后都需要一定的遗传物质基础来支配。随着分子生物学和遗传学的发展,分子检测、基因筛选等成为了现实,人们在注意多种临床表观现象研究的同时,也更加重视这些现象背后的分子机制、遗传背景。基因和疾病的关系成为目前研究的热点,并取得了可喜的成果,如癌基因、抑癌基因、药物敏感基因、预后基因等等。预后因素的研究也进入到了一个新的领域,深入到了细胞甚至蛋白、基因等水平。食管癌是一种复杂疾病,又称多基因病或多因子病。目前,人类基因组计划的完成和单核苷酸多态的单体型图谱构建完成(HapMap计划),为肿瘤的遗传分子基础研究提供了大量的数据支持。同时,经济、高效大通量基因分型技术,在一个反应内可同时检测到几万、几十万个SNPs(单核苷酸多态性)位点成为可能,即全基因组关联分析技术(Genome-wide Associate Study, GWAS)。本项目组运用GWAS技术发现了18个和食管癌密切相关的SNPs位点,在此基础上,将18个SNPs位点及其相关的基因和食管癌预后联系起来,对食管癌预后的分子机制进行研究。与以往有关食管癌预后分子机制方面的研究相比,本研究具有以下特点:1.检测的SNPs位点多。18个SNPs位点是本项目组利用GWAS技术对中国汉族1077食管癌病例和1733正常对照的506,666个位点中筛选发现的。和既往只限于个别基因和预后关系的研究相比,其覆盖面广、代表性强,具有更明显的种族特色和针对性。2.样本量大,临床病理资料全,进入观察项目多。首先从9703病例中随访出有明确生存期的病人共有2290例,然后从21个预后相关项目中选择了13个常见的临床病理因素进行分析。3.临床和基础相结合。预后看似个临床问题,其背后隐藏着复杂的分子生物学机制,把预后及相关临床病理因素和单核苷酸多态性研究相结合,既注重临床表型,又注重分子机制探讨。4.流调工作和实验室验证相结合。把经过流调和统计分析后得到的食管癌预后相关基因在癌组织中进行蛋白表达检测,在分子水平验证基因对预后的影响。初步进行了基因的功能研究。5.具有良好的前期工作基础。易感基因对肿瘤发生的作用和对肿瘤的进展、预后作用具有一致性,二者具有相似的生物学基础和研究方法。本次研究是本项目组“食管癌易感基因GWAS研究”的一个延伸,GWAS前期工作结果已发表在了Nature Genetics杂志上。综上观点,本论文将分为三大部分:第一部分:食管癌预后和临床表型的关系;第二部分:食管癌预后及相关表型和单核苷酸多态性的关系;第三部分:预后相关基因在食管癌组织中的不同表达。2.材料及方法2.1研究对象2.1.1研究对象入选标准本研究主要是对病人进行生存分析,入选标准如下:1)全部研究对象均为中国汉族人群;2)均经手术治疗,排除放化疗因素的干扰;3)有相关的临床病理资料(TNM分期、大体类型、食管癌部位、分化程度、浸润程度、淋巴结转移、血型等);4)家庭住址或电话登记详细并能随访到患者本人或家属;5)登记有详细的病案号和病理号,有完整的病史资料。6)所有研究对象之间均无亲缘关系。本研究共调查食管癌患者9703例,部分来自医院住院患者,部分来自新农合患者,经上述标准严格筛选后,通过电话和家访等形式共随访6793例,得到随访资料齐全并且是经过手术治疗的病例2290例,随访截止日期为2011年3月1同。所有患者未接受任何放疗和化疗。为了排除由于患者在不同地区或者医院就诊可能引起的重复调查,调查时保证每个家庭只有一位先证者。2.1.2研究对象一般情况本次研究系观察生存期及相关因素,所以所统计资料均来源于有明确生存期的2290例患者。其中男性1447例,年龄最小22岁,最大81岁,平均年龄54.69±8.746岁;女性843例,年龄最小29岁,最大86岁,平均年龄54.31±8.478。2.1.3研究对象病理学诊断由至少两名病理科医师确诊,全部符合食管鳞癌的标准。并采用到医院查阅病历的方式,追踪到有完整的临床TNM分期、淋巴结转移、浸润程度、分化程度、肿瘤类型、食管癌病变部位、血型、手术方式并有明确的手术日期等记录。2.2流行病学调查本研究主要采用问卷调查的方式,调查内容主要包括:一般情况(年龄、性别、籍贯等),食管癌肿瘤家族史,初次诊断时间及治疗方式等。回顾性调查病人既往住院病历档案,补充患者的食管癌临床病理资料(食管癌部位,大体类型、分化程度,浸润程度和淋巴结转移等)。然后80%采用电话追踪随访和20%的家访完成生存期随访,随访周期为3个月,随访截止时间为2011年3月。所有调查员均经过严格培训,问卷标准统一。2.3 18个SNPs位点基因分型采用Qiagen DNA试剂盒提取血样中DNA,测浓度,电泳,标化至15-20ng/μl,根据本研究团队在前期利用GWAS技术对内镜和病理确诊的1077食管癌病例和1733正常对照做了506,666个SNPs位点全扫,发现和食管癌密切相关的18个SNPs位点,然后进行引物设计,用Sequenom MassArray技术对472例有完整随访资料的患者在18个SNPs位点进行基因分型,经过人群吻合度检验,观察值和预期值符合Hardy-Weinberg平衡,具有人群代表性。然后通过官方HapMap网站或NCBI网站查询18个SNPs位点所在的染色体及所对应的基因,如rs12263737和rs2274223位点对应于PLCE1基因、rs6140125和rs13042395位点对应于C20orf54基因等。2.4免疫组织化学69例食管癌手术切除标本来自各个医院,其中浸润程度达到固有层、粘膜下层和浅肌层者25例、达到纤维膜或邻近器官者44例,所有患者术前未接受任何放、化疗,并均经组织学证实为食管鳞癌。其中,28例有完整的临床随访资料,并且69例组织均有对应患者的18个SNPs位点的分型。免疫组织化学实验方法采用卵白素-生物素-辣根过氧化氢酶复合物(ABC)法完成。采用本实验室已建立的ABC法,进行定性和定量分析。2.5统计学分析采用SPSS17.0统计软件分析,Cox-regression分析判断影响食管癌患者预后的年龄、性别、家族史、高低发区、食管癌部位、大体类型、病理分化程度、淋巴结转移、TNM分期等临床因素,用Hardy-Weinberg平衡检验18个SNPs位点的人群吻合度,Cox回归、Kaplan-Meier和Log-rank检验分析18个SNPs位点与生存的关系,卡方检验比较18个SNPs位点在不同的临床表型中的遗传差别和免疫组化结果,检验标准取α=0.05。3结果3.1食管癌患者临床表型和预后的关系3.1.1年龄分布与生存时间总体之间无统计学差异(Log-rank P=0.061)。但Cox回归分析结果显示65岁以上年龄组患者较35岁(包括35岁)以下年龄组死亡风险降低了45.7%(HR=0.643),同时差异达到统计学意义(P=0.023)。男、女生存时间没有明显差异(Log-rank P=0.104),男性患者较女性患者死亡风险增加(HR=1.078),但是差异未达到统计学意义(P=0.137)。3.1.2三种不同的好发部位在总体生存时间上差异并无统计学意义(Log-rankP=0.342)。食管中段癌较下段癌死亡风险增加(HR=2.691),但差异未达到统计学意义(P=0.098);食管上段癌较下段癌死亡风险增加(HR=2.008),差异有统计学意义(P=0.045)。3.1.3食管癌总体生存在四种大体形态之间无统计学差异(Log-rank P=0.114),蕈伞型、溃疡型、髓质型较缩窄型死亡风险均降低(HR=0.588;HR=0.65;HR=0.838),)但差异只有溃疡型有统计学意义(P=0.05)。3.1.4各种分化程度的患者生存时间无显著性差异(Log-rank P=0.542)。中分化和低分化患者较高分化患者死亡风险均增加(HR=1.762;HR=1.413),但是差异均未达到统计学意义(P=0.277;P=0.204)。3.1.5早、中、晚期患者之间生存上存在显著的差异(Log-rank P=0.000),中、晚期患者较早期患者死亡风险增加(HR=1.702;HR=2.154),晚期患者与早期之间死亡风险差异有显著性统计学意义(P=0.004)。3.1.6浸润程度达到肌层和纤维膜层的患者较粘膜层的患者生存时间较短(Log-rank P=0.028)。并且浸润程度达到肌层和纤维膜层的患者较粘膜层得患者死亡风险增加(HR=1.676;HR=1.752),且纤维膜层与粘膜层之间死亡风险差异有统计学意义(P=0.016)。3.1.7无淋巴结转移的患者较有淋巴结转移的患者生存时间长(Log-rank P=0.000),Cox回归分析结果显示有淋巴结转移的患者较无淋巴结转移的患者死亡风险增加(HR=2.159),差异有显著性统计学意义(P=0.000)。3.1.8高发区患者生存时间优于中高发区,中高发区患者生存期优于低发区患者(Log-rank P=0.003), Cox回归结果显示,高发区患者死亡风险较低发区患者有所下降(HR=0.647),且差异有显著性统计学意义(P=0.001)。3.1.9食管癌家族史阳性的患者生存时间优于食管癌家族史阴性的患者(Log-rankP=1.18E-06),Cox回归结果显示,食管癌家族史阳性的患者死亡风险较食管癌家族史阴性的患者低(HR=0.64),差异有统计学意义(P=2.53E-06)。3.1.10手术年代分布与患者的生存时间之间存在统计学差异(Log-rankP=3.49E-52)。手术方式与患者的生存时间之间也存在统计学差异(Log-rankP=2.31E-21)。90年代、2000年及以后两组死亡风险较70年代手术的患者均降低(HR=0.666;HR=0.432),差异均有统计学意义(P=6.52E-1和P=2.85E-2),弓下吻合和颈部吻合较弓上吻合的死亡风险均降低(HR=0.943;HR=0.814),且颈部吻合与弓上吻合之间死亡风险差异有统计学意义(P=0.046)。3.1.11不同血型分布与患者的生存时间之间不存在统计学差异(Log-rankP=0.329);Cox回归分析发现:B型、O型两组死亡风险和A型的患者基本相似。(HR=0.988和HR=0.967),差异均无统计学意义(P=0.882和P=0.692)。3.2食管癌预后和18个SNPs位点基因型的关系3.2.1 AHDC1基因rs4908343位点A>G基因型分布中,与突变纯合基因型GG携带者相比,野生纯合基因型AA携带者死亡风险降低(HR=0.54,95%CI=0.30-0.99)。与携带AA/GA基因型者相比,GG基因型携带者生存时间较短(Log-rank P=0.048)3.2.2 PPP1R14A基因rs8102476位点T>C基因型分布的三种基因型携带者,携带TT和TC基因型的患者生存时间较短(Log-rank P=0.043)3.3食管癌临床表型和18个SNPs位点基因型的关系3.3.1 C20orf54基因rs6140125位点C>A基因型分布在食管癌不同的浸润程度(粘膜层、肌层、纤维膜)之间有统计学差异(P=0.007)。3.3.2 IRX1基因rs11134032位点T>C基因型分布与食管癌发生部位有统计学差异(P=0.014)。3.3.3 PPP1R14A基因rs8102476位点T>C基因型分布与食管癌淋巴结转移有统计学差异(P=0.037)。3.3.4 PES1基因rs5753220位点C>T基因型分布与临床分期有统计学差异(P=0.033),且与淋巴结转移也有差异(P=0.044)。3·4食管癌预后和基因多态性的分层分析为控制潜在混杂因素对基因多态性和预后的影响,我们分别以年龄、性别、食管癌发生部位、大体类型、临床分期及浸润程度和淋巴结转移等为分层变量进行SNPs和预后关系的分层分析,结果如下:3.4.1 AHDC1基因rs4908343A>G多态性改变分层后发现在男性患者组、家族史阴性组、食管癌晚期组、淋巴结转移阳性组中,与携带AA/GA基因型者相比,携带GG基因型者的死亡风险增加(HR=1.90; HR=1.85; HR=2.43; HR=2.66)3.4.2 PES1基因多态性rs5753220C>T多态性改变分层后发现在高分化组、淋巴结转移阴性组中,,与携带TC/CC基因型者相比,携带TT基因型者死亡风险增加(HR=131.96;HR=1.98)3.5食管癌预后和基因多态性的联合分析将三个和预后相关的基因AHDC1基因rs4908343A>G、PPP1R14A基因rs8102476T>C和PES1基因rs5753220C>T位点基因多态性联合分析,发现携带预后不良基因型个数的增加,其死亡风险也增大(Log-rank P=0.01),趋势性P=-0.012。3.6免疫组织化学结果C20orf54和食管癌易感有关。预后分析中发现C20orf54与食管癌的浸润深度有关联,所以进行了该基因在食管组织中表达的验证。C20orf54蛋白阳性表达部位主要在胞浆,食管癌组织阳性表达率为87%,但是阳性表达率和组织浸润程度、基因型及预后无关。4.结论4.1食管癌预后和临床分期、浸润程度、淋巴结转移、高/低发区、食管癌家族史、手术年代及手术方式有统计学相关。4.2 AHDC1基因rs4908343A>G、PPP1R14A基因rs8102476T>C位点的基因多态性和食管癌预后相关;食管癌分层后PES1基因rs5753220C>T位点的多态性在高分化或者淋巴结阴性组和食管癌预后风险相关。携带危险基因型数目越多,死亡风险也随之增大。4.3 C20orf54基因rs6140125位点、IRX1基因rs11134032位点、PPP1R14A基因rs8102476位点和PES1基因rs5753220位点基因多态性与食管癌临床表型相关。4.4 C20orf54基因蛋白表达在食管癌组织中较强,但是表达率和肿瘤组织浸润程度、基因型分布及预后无关。
【Abstract】 1. Background and ObjectiveEsophageal cancer (EC) is one of the six most common malignancies worldwide. The incidence of this disease is the eighth in the all carcinomas worldwide while its mortality rate is the sixth in the all cancers worldwide. Our country is the highest area of incidence and mortality rate of EC in the world while EC is the fourth most frequent cause of cancer-related deaths in China.The total incidence and deaths of EC in China takes up the half number of the world’s each year.The prognosis of EC is very poor. The natural survival time of advanced EC is 7-8 months approximately, the survival rate after treatment is also very unsatisfactory, and 5-year survival rate is 20% approximately. Although the therapy of EC has been improved greatly, the five-year survival rate of EC has not been significantly improved in recent decades.Three early" works has been attracted widespread attention. But in fact, the 80-90% of patients with EC is in the middle and late period when they get to hospital. To improve the survival of this large group of patients has become a decisive link of the actuality of prophylaxis and treatment in EC. The prognosis of patients with EC in middle and late period is very poor generally, but after the intervention by the treatment there are still 20% of patients who can survive in a long run. It presents a scientific problem. What causes a different prognosis in a similar population? In terms of pathogenesis, natural survival and clinical outcome, EC emerged significant individual characteristics. The effects of environmental factors or differences in congenital genetic? More consistent view is the result of genetic and environmental interaction, but the role of the two, the pros and cons as well as the proportion of both has not been understood clearly.The research for prognosis of disease has always been attracted by the researchers and doctors constantly. To search for the internal mechanism of prognosis and establish the impersonal evaluation system, it will proceed early intervention direct prognosis. It has realistic and important significance in the field to improve patient survival rates.Previous studies were mostly concentrated in the interrelated factors such as external environment, apparent symptoms, clinical pathological characteristics of prognosis. These factors include TNM stage, living habits, environment, treatment methods and so on. Some consensus has been achieved, however there are still have many controversies and doubts. There is still a great distance from the work requirements of clinical practice. So we can see the complexity of this problem obviously."All medicines are genetics." The occurrence, development, prognosis and clinical phenotype of diseases, all have the genetic material basis. With the development of molecular biology and genetics, molecular detection and genetic screening have become reality. People who pay attention to the studies for multiple phenomenon of clinical must pay more attention to the molecular mechanisms and genetic background underlying these phenomena. The relationships between genes and disease has become the focus of current research and have achieved encouraging results, such as oncogenes, tumor-suppressor genes, drug-sensitive genes, the prognosis genes, etc. The research for prognostic factors also entered into a new field of the cell and gone so far as to protein and genes levels.EC is a complex disease, also known as multi-genes disease or multi-factors disease. At present, the completion of the Human Genome Project and single nucleotide polymorphism haplotype map (HapMap plan) provide a large amount of data for the genetic study of tumor. At the same time, it becomes possibility that economic and efficient large-throughput genotyping technology can detect the tens of thousands or hundreds of thousands of SNPs within one response. It is known as the genome-wide association study (GWAS). Taking advantage of GWAS technology, our team found the 18 EC associative SNPs loci. On the basis, our study intends to observe the relationship between 18 SNPs loci and correlative genes and prognosis of EC, and then study the molecular mechanism of prognosis.Compared with the previous studies for prognosis related to molecular mechanisms of EC, this project has the following characteristics:(1). The number of detective SNPs loci is large.18 SNPs loci were selected and found from 506,666 SNPs loci of 1077 cases of patients with EC and 1733 controls in China by using GWAS technology. Compare with the previous studies for the relationships between limited genes and prognosis, this study have extensive coverage, intense representativeness, and more obvious racial characteristics and pertinence.(2). The sample volume is large and the clinicopathological data are completed. The numbers of observational items are large. First of all, selected 2290 cases patients which have clear survival time by followed up from 9703 cases, and then selected 13 common clinicopathological factors which to be analyzed from 21 prognosis relative items.(3). Combined the clinical study and basic medicine. Prognosis appears to be clinical problems, which has the complex molecular mechanisms. This study wants to know the interactions among the prognosis, clinicopathological factors and the single nucleotide polymorphisms.(4). Combined epidemiological survey work and laboratory validation. Detect the protein expression of EC prognosis related genes in cancer tissue which were found by epidemiological inquiry and statistical analysis, and then validate the influence of prognosis by gene at the molecular level. Conduct a preliminary study for gene function.(5). With a good foundation of preliminary work. The susceptibility genes of the tumor have consistency on the role of the occurrence, progress and prognosis, they have similar biology basis and research methods. This study is an extension of "the GWAS work for esophageal cancer susceptibility genes" by our project team, as well as GWAS results of preliminary work has been published in the journal of Nature Genetics.In summary, this paper will be set to three parts. The first part:the relationship between EC prognosis and clinical phenotype; the second part:the relationship between EC prognosis with correlative phenotype and single nucleotide polymorphism; the third part:the expression changes of prognosis related genes in the tissues of EC.2. Materials and methods2.1 The object of study2.1.1 The inclusion criteria of study objectThis study aims at the survival analysis of patients, inclusion criteria as follows:1) All subjects were Chinese Han population;2) All subjects were treated by surgery, excluded radiotherapy and chemotherapy disturbances;3) Relevant clinicopathological data (TNM stage, gross type, esophageal cancer location, the degree of differentiation, the degree of invasion, lymph node metastasis, blood type,etc);4) Home address or phone were registered clearly and can follow-up to the patients or their families;5) Medical record number and pathological number were detailed, as well as medical history informations were completed;6) All of the subjects had no genetic relationship each other.The study surveyed total of 9703 individuals with EC. These patients partly were hospital patients, partly from the new rural cooperative medical care system. After rigorous screening by the above criteria, followed up 6793 cases by telephone and home visits, get 2290 cases which have completed follow-up data and received surgical treatment. The follow-up deadline is March 1,2011. All the patients did not receive any preoperative radiotherapy and chemotherapy. In order to exclude duplicated investigation which results from the patients going to the different hospitals and regions, each family only has one proband at investigation.2.1.2 The general situation of study objectThe study observed survival time and related factors, so the statistics data were all from the 2290 cases of patients which had straightforward survival time. Therein,1447 male cases that the youngest was 22 years old, maximum age was 81 years and the mean ages were 54.69±8.746; and 843 female cases that the youngest was 29 years old, maximum age was 86 years and the mean ages were 54.31±8.478 years.2.1.3 The pathology diagnosis of study objectAll these patients were confirmed by at least two pathologists, meanwhile accord with diagnostic criteria for esophageal squamous cell carcinoma. And access medical records by going to the hospital, to track complete data included clinical TNM stage, lymph node metastasis, the degree of invasion, the degree of differentiation, tumor type, lesions location of EC, blood type, the type of surgical operation and a clear date of surgical operation.2.2 The investigation of epidemiologyThis study meanly uses the way of questionnaire survey. The contents of survey include:general condition (age, sex, native place etc.), family history of EC, the date of initial diagnosis and treatment methods. We conducted retrospective study of patients’records, added clinical pathological data of patients with EC (including location of EC, general types, the degree of differentiation, the degree of invasion, lymph node metastasis, etc). And then follow up survival time by telephone manner and home visits manner in the patients of 80% and 20% respectively. The follow-up period was 3 months. The deadline of followed up is March 2011. All investigators were trained strictly while the standard of questionnaire is unified.2.3 Genotyping of 18 SNPs lociGenomic DNA was extracted by using Qiagen Gene DNA kits and the concentration was measured, then electrophoresis, and normalized to 15ng-20ng/μl. According to our research team in the early study that made full scan of 506,666 SNP loci for 1077 cases of EC patients of endoscopic and pathological diagnosis and 1733 controls by using GWAS technology, as well as detected 18 SNP loci related closely with EC, and then primer design. By using Sequenom MassArray technology conducted genotyping for 18 SNP loci on 472 cases of patients which had complete follow-up data. Consistent with Hardy-Weinberg equilibrium, observed and expected value had people representative through the crowd goodness of fit tests. Then inquire the chromosome and corresponding genes which 18 SNPs loci located in through the official HapMap website or NCBI website, such as rs12263737 and rs2274223 loci corresponding to the PLCE1 gene, rs6140125 and rs13042395 sites corresponding to the C20orf54 gene, etc.2.4 Immunohistochemistry69 specimens of operation excises of EC from several hospital that include 25 cases were infiltrated to the lamina propria, submucosa and superficial muscle layer and 44 cases were infiltrated to the fiber membrane or the adjacent organs. All patients did not receive any radiotherapy and chemotherapy before operation while all specimens were confirmed esophageal squamous cell carcinoma by histology. Among them,28 patients had complete follow-up data, and the tissues of 69 corresponded cases had genotyping of 18 SNPs loci. Immunohistochemistry used the method of avidin-biotin-horseradish peroxide enzyme complex (ABC) that our laboratory had been established to conduct qualitative and quantitative detection.2.5 Statistical analysisUsing SPSS 17.0 statistical software, Cox-regression analysis determine clinical factors that influence the prognosis of patients with EC, which include age, sex, family history, high and low incidence areas, location of EC, general types, the differentiation degree of pathology, lymph node metastasis, TNM staging,etc. Use Hardy-Weinberg equilibrium to test the crowd coincide degrees of 18 SNPs loci. Use Cox-regression, Kaplan-Meier and Log-rank test to analyse the relationship between the 18 SNPs loci and survival time. Compare the genetic differences of the 18 different SNPs loci in different Clinical phenotypes and examine immunohistochemical results by chi-square test (a=0.05 was considered as the test standard).3. Results3.1 The relationship between clinical characters of patients with EC and prognosis3.1.1 There were no significant difference between the distribution of age and survival time (Log-rank P=0.061), however, Cox-regression analysis shown that >65 years age group of patients lowered the risk of death by 45.7% than≤35 years age group (HR=0.643), and the difference reached statistical significance (P=0.023). Male and female survival time differ unsignificantly (Log-rank P=0.104), and male patients had increased risk of death than female patients (HR=1.078), but the difference did not reach statistical significance (P=0.137).3.1.2 The difference of overall survival time among three frequently occurred location of EC did not reach statistically significance (Log-rank P=0.342). The middle part of EC had higher cancer risk of death than the lower part (HR=2.691), but the difference did not reach statistical significance (P=0.098); The upper part of EC have higher cancer risk of death compared with the lower part (HR=2.008), and the difference was statistically significant (P=0.045).3.1.3 The difference of overall survival time among four sub-types was no statistically significant (Log-rank P=0.114), mushroom type, ulcerative type, and medulla type had decreased risk of death than the narrow type (HR=0.588; HR=0.65; HR=0.838), but only the difference of ulcer type was statistically significant (P=0.05).3.1.4 The difference of survival time among patients with various degree of differentiation was no significant (Log-rank P=0.542). The patients with moderately differentiated and poorly differentiated had higher risk of death than well differentiated (HR=1.762; HR=1.413), but the difference did not reach statistical significance (P=0.277; P=0.204).3.1.5 The survival time among patients with early stage, metaphase and later period had significant difference (Log-rank P=0.000), The patients of metaphase and later period had increased risk of death than patients with early stage (HR=1.702; HR=2.154), The difference of risk of death between later period patients and early stage had statistical significance (P=0.004).3.1.6 The patients with invasion to muscle and fiber layer had shorter survival time than that of invasion to mucosa (log-rank P=0.028). At the same time the patients with invasion to muscle and fiber layer have an increased risk of death than that of invasion to mucous layer (HR=1.676; HR=1.752), and the difference of the risk of death between patients with invasion to fiber layer and mucosa layer had statistical significance (P=0.016).3.1.7 Patients without lymph node metastasis had longer survival time than patients with lymph node metastasis (Log-rank P=0.000);Cox regression analysis showed that patients with lymph node metastasis increased risk of death than patients without lymph node metastasis (HR=2.159), the difference had statistical significance (P=0.000).3.1.8 The patients in the high incidence areas had longer survival time than that of middle-high incidence, as well as patients in the middle-high incidence had longer survival time than that of low-incidence areas (Log-rank P=0.003); Cox regression showed that the patients in high incidence had lower risk of death than patients in low-incidence areas (HR=0.647), and the difference had noteworthy statistical significance (P=0.001).3.1.9 The patients with positive family history of EC had longer survival time than that with negative family history (Log-rank P=0.18E-06); Cox regression analysis showed that patients with positive family history of EC had lower risk of death compared with patients with negative family history of EC (HR=0.647), the difference was statistically significant (P=2.53E-06).3.1.10 There was significantly difference between the distribution of surgical age and the survival time of patients (Log-rank P=3.49E-52), as well as the distribution of surgical methods (Log-rank.P=2.31E-21). The Nineties, after 2000 have the lower risk of death compared with surgery patients of the Seventies (HR=0.666; HR=0.432), the differences were statistically significant (P= 6.52E-1; P=2.85E-2); The patients with under the arch anastomosis and neck anastomosis had decreased risk of death than that of over the arch anastomosis (HR=0.943; HR=0.814), and difference of the risk of death between patients with neck anastomosis and over the bow anastomosis was statistically significant (P=0.046)3.1.11 The distribution of different blood type and survival time of patients had no significant differences (Log-rank P=0.329);Cox regression analysis showed that the risk of patients of B type, O type and A type is almost similar and the difference had no significances (P=0.882;P=0.692)3.2 The relationship between the genotype of 18 SNPs loci and prognosis of EC3.2.1 In the gene distribution of AHDC1 gene rs4908343A>G genotype, compared with carriers of mutation homozygous GG genotype, the carriers of wild type homozygous AA genotype have reduced risk of death (HR=0.54,95% CI=0.30-0.99). Compared with carriers of AA/GA genotype, carriers of GG genotype had shorter survival time (Log-rank P=0.048).3.2.2 Compared with the three genotype carriers of gene distribution of PPP1R14A gene rs8102476T>C genotype, patients carrying TT and TC genotype had shorter survival time (Log-rank P=0.043).3.3 The relationship between 18 SNPs loci polymorphism and clinical phenotypes of EC3.3.1 Genotype distribution diversities of C20orf54 gene rs6140125C>A polymorphism among different infiltrating degree of EC (mucosa, muscle, fiber coating) had statistically significance (P=0.007).3.3.2 There was significantly different between IRX1 gene locus rs11134032 T>C genotype and esophageal cancer sites (P=0.014).3.3.3 Genotype distribution of PPP1R14A gene rs8102476T>C polymorphism and lymph node metastasis had significantly statistical difference (P=0.037).3.3.4 Genotype distribution of PES1 gene rs5753220C>T polymorphism and clinical stage had significantly statistical difference (P=0.033), as well as lymph node metastases (P=0.044).3.4 The stratified analysis of gene polymorphism and prognosis of ECTo control the influence of potential confounding factors on gene polymorphism and prognosis, we conducted stratified analysis between SNPs and prognosis by stratified variables such as age, sex, location of EC, macroscopic type, clinical stage, the degree of invasion and lymph node metastases etc, the results as follows:3.4.1 The stratified analysis of AHDC1 gene rs4908343A>G polymorphisms had found that carrier of GG genotype had increased risk of death compared with carrier AA/GA genotype in male patients stratified group, family history negative group, the later period group of EC, positive lymph node metastases group (HR=1.90; HR=1.85; HR=2.43; HR=2.66).3.4.2 The stratified analysis of PES1 gene rs5753220C>T polymorphism found that carrier of TT genotype had increased the risk of death compared with carrier TC/CC genotype in well differentiated group, negative lymph node metastases group (HR=131.96; HR=1.98).3.5 The joint analysis of gene polymorphism and prognosis of ECUse joint analysis for the polymorphism the three genes, AHDC1 genes rs4908343A>G, PPP1R14A genes rs8102476T>C and PES1 genes rs5753220C>T related to prognosis, and show that patients with more genotypes of poor prognosis, the risk of death also increased (log-rank P=0.01), trend P=0.012.3.6 The results of immunohistochemistryC20orf54 and esophageal cancer susceptibility were related. In our study, C20orf54 was found associated with the invasion, so this gene was verificated in the expression of esophageal tissues.The positive expressions of protein mainly took place in endochylema. The positive expression rate of C20orf54 in tissues of EC was 87%, but the positive expression rate had no correlation with the degree of tissues infiltration, genotype and prognosis.4 Conclusions 4.1 The prognosis of EC were statistically correlated with clinical stage, depth of invasion, lymph node metastasis, high/low-incidence areas, family history of EC, age of surgery and surgical methods.4.2 The polymorphism of AHDC1 gene rs4908343A>G, PPP1R14A gene rs8102476T>C sites were correlated with prognosis of EC.In the stratified analysis, PES1 gene rs5753220C>T polymorphism was related with the prognoses in well differentiated group, negative lymph node metastases group.Moreover, the more quantity of carried risk genotypes, the more increases the risk of death.4.3 C20orf54 gene rs6140125 locus, IRX1 gene rs11134032 locus, PPP1R14A gene rs8102476 locus and PES1 gene rs5753220 locus are associated with the clinical phenotype of EC.4.4 The protein expressions of C20orf54 gene in tissues of EC are intense. But its expression rates are not associated with the degree of tissues depth of invasion, the distribution of genotype and prognosis.
【Key words】 esophageal cancer; survival time; phenotype of clinic; single-nucleotide polymorphism;