【作者】 赵鹏；

【导师】 吕卫国；

【作者基本信息】 浙江大学 ， 临床医学（专业学位）， 2017， 博士

【摘要】 研究背景:完全性葡萄胎(complete hydatidiformmole,CHM)是最常见的滋养细胞疾病,约占疾病谱的80%,亚洲人群发病率较高,是欧美人群的10倍。人类早在1895年就认识到葡萄胎是妊娠相关性疾病,从此之后,虽然人类对葡萄胎的病理生理的认识不断在深化,但其诊治策略却没有本质的提高:葡萄胎的诊断主要靠超声等影像学检查,治疗上及时清宫,对高龄人群可采取预防性全子宫切除术、高危人群可选择预防性化疗等治疗策略。葡萄胎患者即便及时得到上述诊治,仍有10%-17%的概率会恶变为妊娠滋养细胞肿瘤,3%发展为绒癌。葡萄胎清宫后的恶变率是其他妊娠形式如流产、足月产的2000倍,由此可见,葡萄胎的恶变是一个严重危害女性生殖健康的疾病。继发于葡萄胎妊娠的滋养细胞肿瘤(gestational trophoblastic neoplasia,GTN)的及时诊断和治疗是影响其预后的关键因素。早期诊断的病例几乎完全可以治愈,而延时发现和诊断的病例则有死亡风险。目前GTN的诊断标准是FIGO标准,主要依靠血清人绒毛膜促性腺激素(human chorionic gonadotropin,hCG)的动态监测,其诊断标准为:1)hCG测定4次呈平台期,即第1、7、14、21日;2)hCG测定3次升高,即第1、7、14日;3)hCG升高持续6个月;4)组织学诊断绒癌。这种随访方案高度依赖于患者的依从性,要求患者在清宫后必须严格按照每周一次的方案复查hCG,持续2年。这种长达2年的随诊方案给患者带来了巨大的不便和长期的心理负担,也大大增加了医疗资源的负担。而且,一旦患者未按照既定时间复查hCG,则势必会延误GTN的诊断。据统计,葡萄胎患者清宫后的依从性极差:在美国低收入患者中,能完全依从FIGO随诊方案的患者只占7%-36%,而这一比例在发展中国家及欠发达国家中更低,严重影响了 GTN的及时诊断和治疗。为了解决这个问题,国内外学者提出了很多应对措施,最为常见的方法是基于hCG下降曲线的预测模型。文献引用较多的预测模型有两种:每周hCG下降均值曲线,每周hCG下降率曲线。这些预测模型具有成本低廉、操作方便的优点,但由于模型仍然要求患者每周复查hCG,因此不能完全克服患者依从性差带来的难题:即患者如未能按照每周一次的方案复查hCG(比如间隔的时间不足一周或者超过一周不足两周),则上述预测模型不能应用。因此,本项研究的第一、第二部分的目标是尝试建立一种适合依从性差的患者的预测模型,即每日hCG下降均值曲线和每日hCG下降率曲线,并与其他学者建立的两种预测模型进行比较,以评价四种预测模型的临床应用价值。本项研究的第三、第四部分的目标是评估预防性全子宫切除术、预防性化疗对预防GTN的临床价值。考虑到GTN发生率与患者年龄的相关性,许多学者建议,对于高龄的葡萄胎患者,如果已完成生育计划,可以考虑预防性全子宫切除术,该术式可以降低葡萄胎清宫后GTN的发生率,也有学者建议对于存在高危因素的葡萄胎患者使用预防性化疗,通过全身化疗杀灭隐藏在脉管系统中的滋养细胞,阻止其转移,从而达到预防GTN的目的。但上述治疗策略仍存在较大争议:有学者认为预防性全子宫切除术不仅不能降低GTN的发生率,还可能因手术操作诱发滋养细胞的血行转移;预防性化疗则会导致80%的患者接受不必要的化疗,且会增加GTN患者的化疗耐药率,降低治愈率。综上所述,本文的研究目标,是建立一个能适用于所有葡萄胎妊娠患者尤其是依从性差的患者的预测模型(每日hCG下降率标准曲线),并与其他研究者提出的预测模型进行对比,对所有预测模型的临床价值通过前瞻性临床研究进行验证。同时,以研究者所在诊疗机构为中心,通过回顾性分析评估预防性全子宫切除术降低40岁以上葡萄胎患者恶变的作用;并通过筛选全球范围内的相关文献,进行系统评价和meta分析,评价预防性全子宫切除术预防40岁以上葡萄胎患者恶变的临床应用价值。第一部分基于血清hCG预测完全性葡萄胎清宫后恶变的新模型的建立和临床价值研究目的:建立一个适用于预测依从性差的葡萄胎妊娠患者恶变的预测模型,并与其他预测模型对比,评估新建预测模型的临床价值。材料与方法:回顾性筛选从2000年到2010年于我院确诊并接受治疗和随访的葡萄胎妊娠患者。入选标准:1)回顾性病理切片阅片,确诊为完全性葡萄胎;2)初始治疗为清宫;3)随访2年未失访。排除标准:1)回顾性病理切片阅片,确诊为部分性葡萄胎、侵蚀性葡萄胎、绒癌、胎盘部位滋养细胞肿瘤和上皮样滋养细胞肿瘤;2)初始治疗为预防性化疗或全子宫切除术;3)患者未能完成2年的随访。收集符合入选标准的患者的临床资料,包括一般情况,治疗策略,历次hCG随访的结果等。利用未恶变组患者的hCG数据(hCG数据共计1383人次)构建四种预测模型。用恶变组患者的hCG数据检验四种预测模型的预测效能。具体方法如下:根据其他学者报道的方法,构建清宫后未恶变患者的每周hCG下降均值标准曲线,每周hCG下降率标准曲线;根据本文研究者提出的理论,构建每日hCG下降均值标准曲线,每日hCG下降率标准曲线;对比新建预测模型和已有的预测模型,评估四种预测模型的临床应用价值。结果:共有222例患者符合入选标准。其中,195例患者的血清hCG在清宫后自然恢复到正常,定义为未恶变组;其余的27例患者进展为GTN,定义为恶变组。每日hCG下降率标准曲线的预测敏感度最高,为100%,与FIGO标准比较,能提前14日预测恶变,有显著性统计学差异(P=0.008);每日hCG下降均值标准曲线预测敏感度为92.6%,能提前约10日预测恶变,有显著性统计学差异(P=0.030);每周hCG下降率标准曲线预测敏感度为85.2%,能提前14日预测恶变,有显著性统计学差异(P=0.006);每周hCG下降均值标准曲线的预测能力与FIGO标准比较无显著性统计学差异。结论:与其他三种hCG预测模型比较,每日hCG下降率标准曲线的预测效果最佳,可更好的用于监测葡萄胎患者,尤其适用于那些依从性差不能严格按照FIGO标准随诊的患者。第二部分四种血清hCG预测模型预测完全性葡萄胎恶变的临床价值的前瞻性评价目的:前瞻性比较四种血清hCG预测模型预测完全性葡萄胎恶变的临床价值。材料与方法:筛选从2011年到2015年6月于我院接受治疗的葡萄胎妊娠患者。入选标准为:(1)病理阅片确诊为完全性葡萄胎;(2)初始治疗方案为清宫。所有入选患者在清宫后开始进行随访,为期两年。具有完整的hCG随访结果(包括清宫前hCG结果或随访期间任何一次hCG结果)的患者予以纳入本研究。以第一部分建立的四种预测模型为基础,用本项研究入选病例的hCG数据计算四种预测模型的临床诊断价值,评价指标是灵敏度,特异度,阳性预测值,阴性预测值,正确率,约登指数,优势比,阳性似然比和阴性似然比。结果:入选的117例完全性葡萄胎患者中,有16例(13.7%)患者发生恶变,101例(86.3%)患者未发生恶变。每周hCG下降均值标准曲线有7例GTN患者预测失败,有9例未恶变患者预测失败,预测完全性葡萄胎恶变的敏感度为56.3%,特异度为91.1%;每周hCG下降率标准曲线有3例GTN患者预测失败,有6例未恶变患者预测失败,敏感度为81.2%,特异度94.1%。每日hCG下降均值标准曲线有2例GTN患者预测失败,有20例未恶变患者预测失败,敏感度为87.5%,特异度80.2%。每日hCG下降率标准曲线全部GTN患者预测成功,有2例未恶变患者预测失败,敏感度为100%,特异度98.1%。每日hCG下降率标准曲线的敏感度、特异度和正确率均最高,并且,四种预测模型中,只有每日hCG下降率标准曲线的阳性似然比大于10且阴性似然比小于0.1。结论:每日hCG下降率标准曲线预测完全性葡萄胎恶变的能力最佳,适用于依从性差的完全性葡萄胎患者的随诊。第三部分预防性全子宫切除术对预防40岁以上完全性葡萄胎患者恶变的临床价值的研究:回顾性队列研究目的:评估预防性全子宫切除术对预防40岁以上完全性葡萄胎患者恶变的临床价值。材料与方法:回顾性收集2004年至2013年于我院确诊和治疗的完全性葡萄胎患者。入选标准为:(1)病理阅片确诊为完全性葡萄胎;(2)年龄在40岁或40岁以上;(3)初始治疗为清宫术,清宫术后无妊娠物残留;(4)初始治疗前无局部浸润和远处转移的证据;(5)完成2年的随访未失访。收集符合入选标准的患者的临床资料,包括临床特征,传统高危因素,治疗策略,随访结局,GTN发生率,化疗耐药率,诊断GTN的时间间隔等指标,通过统计对比,评价预防性全子宫切除术降低GTN发病率的临床价值。结果:171名患者符合入选标准。患者年龄40-56岁;平均年龄为46.9岁。患者的平均产次为1.2(范围:0-4),而平均孕周为9.5周。根据治疗策略分为三组:单纯清宫组(124例)、预防性化疗组(12例)和预防性全子宫切除组(35例),其GTN发生率分别为37.1%,41.7%和11.4%。与单纯清宫相比,预防性全子宫切除术可明显降低GTN的发生率(χ2=8.342,P=0.004)。预防性化疗组的GTN发生率与单纯清宫组无显著性统计学差异(P=0.035,统计方法:Bonferroni校正,P<0.017(0.05/3)有统计学意义)。结论:预防性全子宫切除术是最有效的预防40岁以上完全性葡萄胎患者恶变的治疗策略。第四部分预防性全子宫切除术对40岁以上葡萄胎患者恶变的预防作用:系统评价和meta分析目的:通过系统评价和meta分析,对预防性全子宫切除术预防40岁以上葡萄胎患者恶变的临床价值进行全面评估材料与方法:本项研究检索了 5个数据库,包括美国国立医学图书馆,荷兰医学文摘数据库,Cochrane协作网系统评价数据库,中国知网,美国临床试验数据库。入选标准:1)人类研究;2)明确的说明干预措施是全子宫切除;3)明确地表示对照措施为清宫术;4)明确的说明入选对象是40岁以上的葡萄胎妊娠患者;5)统计数据含有GTN的发生率。排除标准:1)综述性文章,2)案例报告或案例系列,3)提供的研究数据无法计算风险比。根据GTN的发生率进行meta分析。结果:5篇文献符合入选标准,共纳入了 291名研究对象。2篇文献的目标人群是美国地区,3篇文献的目标人群是亚洲地区(中国2篇,泰国1篇)。研究对象的年龄,3篇文献是40岁以上,1篇是40到49岁,1篇是50岁以上。暴露组为接受预防性全子宫切除的葡萄胎患者,对照组为接受单纯清宫的葡萄胎患者,样本量最小22名,最大159名,1篇文献的随访时间是3-14年,1篇文献的随访时间是3-12年,另外3篇文献未报到随访时间。所有研究的结局指标均至少包含GTN发生率。Meta分析结果显示:与单纯清宫术相比,预防性全子宫切除术的总体效应OR值为0.14(95%可信区间,0.06-0.34,P<0.0004),总体的异质性较低(12=21%,P=0.28)。年龄亚组分析,以及地区亚组分析均与总体效应结果一致。发表偏倚结果(Egger’stest,P=0.149;Begg’stest,P=0.462)说明无显著性发表偏倚。结论:与单纯清宫术相比较,预防性全子宫切除术对于40岁以上葡萄胎患者的治疗效果更佳,因此推荐用于无生育要求的40岁以上葡萄胎患者以预防恶变。更多还原

【Abstract】 Background:Complete hydatidiform mole(CHM),the most common form of gestational trophoblastic disease,consists of 80%of the disease spectrum.The incidence of CHM in Asian population is ten times more than that of European population.The association between CHM and pregnancy was first noted in the year 1895,and since then,the diagnosis and treatment strategies of CHM have not much improved although its physiopathology has been deeply studied.The diagnosis is mainly dependent on ultrasonography and as soon as the diagnosis is established,uterine evacuation must be performed.For those older patients with CHM,total hysterectomy and prophylactic chemotherapy are considered as one of the treatment option.However,in spite of these above-mentioned therapeutic strategies,there are still 10%-17%of patients with CHM who progress to gestational trophoblastic neoplasia and 3%advance to choriocarcinoma.The incidence of GTN after hydatidiform mole is 2000 times greater than that after abortion and term pregnancy.Therefore,the malignancy of CHM after uterine evacuation is a severe life-threatening condition to reproductive women.A prompt diagnosis and treatment is the key to cure post-molar gestational trophoblastic neoplasia.Almost 100%of patients who are diagnosed and treated at an early stage can be eventually cured,while for those patients who delay the diagnosis or treatment,there will be an increased chance for metastases and mortality.The golden standard for the diagnosis of GTN is based on FIGO protocol,which requires weekly serum hCG measurements.Based on FIGO criteria,the diagnosis of post-molar GTN is made as follows;1)an elevated hCG plateau(days 1,7,14 and 21);2)rising hCG titers(days 1,7,14);3)when the hCG level remains elevated for 6 months or more;4)a histologic diagnosis of choriocarcinoma.This follow-up regime demands patients’ fully compliance,which is to return for hCG monitoring every seven days for a minimum period of 2 years.This long-time-interval strategy is a huge burden and usually imposes a severe inconvenience as well as a potential waste of medical resources.Moreover,once the patients miss the appointment,the diagnosis of GTN will be delayed.According to relevant data,the compliance of patients with CHM was extremely poor:in low-income American people,the fully compliance rate was only 7%-36%.Furthermore,this phenomenon is believed to be even worse in developing countries and less-developed countries.To solve the problem,researchers have proposed several prediction models.The most popular method is the prediction model of hCG regression pattern.There are two promising prediction models:weekly hCG regression curve and weekly hCG regression rate curve,which are low-cost and convenient in clinical practice.However,these models still require weekly hCG monitoring thus cannot overcome the inherent disadvantage,that is,when the patient does not return for follow-up in the seven-day interval(for instance,less than seven days or more than seven days),the models will lose their ability to predict.Therefore,one of the objectives of the current study is to establish a novel daily-hCG-based prediction model which can be applied to patients with poor compliance.Comparisons were conducted to evaluate the clinical value of the newly established prediction models and the existing prediction models.The other objective of the current study is to evaluate the effectiveness of prophylactic hysterectomy and prophylactic chemotherapy in preventing post-molar GTN.Considering the age-incidence relationship involved,some researchers might choose to offer hysterectomy to reduce the risk of developing GTN in older patients who have completed their families.In addition,other researchers may recommend prophylactic chemotherapy as an alternative way to prevent post-molar GTN,as chemotherapeutic agents can prevent GTN via eliminating trophoblastic cells hidden in lympho-vascular space.However,these procedures are controversial.Some studies have found no significance in preventing post-molar GTN after undergoing prophylactic hysterectomy.Some author even argued that the manipulation in surgery could facilitate vascular metastasis of trophoblastic cells.A similar situation was noted concerning to prophylactic chemotherapy.It is reported that prophylactic chemotherapy exposed 80%of patients unnecessarily to toxic side effects which lead to incomplete protection against persistent tumor and lower the probability of full recovery.Taken all the above facts into consideration,we conducted the current study to1)establish a novel prediction model(daily hCG regression rate curve)that can be applied to poor-compliant patients and compare the new prediction model with existing models;2)compare the effectiveness of four prediction models in predicting post-molar GTN;3)compare the effectiveness of different therapeutic strategies for CHM in women who were at least 40 years old based on the data of our hospital;4)conduct a systematic review and meta-analysis to further study the value of prophylactic hysterectomy in preventing post-molar GTN in patients with HM being at least 40 years old.Part ⅠA novel hCG-based prediction model for post-molar gestational trophoblastic neoplasia in women with complete hydatidiform mole and its clinical valueObjective:To establish a novel prediction model applicable to poor compliant patients and compare with existing models in the aspect of clinical value.Materials and Methods:We retrospectively reviewed the medical records of all patients who were treated and were confirmed with a diagnosis of complete hydatidiform moles during the period from January 2000 to December 2010.Inclusion criteria were:1)histologically confirmed diagnosis of complete hydatidiform mole,2)initial therapy was D&C;3)complete follow-up period for a minimum of two years.Exclusion criteria were:1)histologically confirmed diagnosis of partial mole,invasive mole,choriocarcinoma,placental site trophoblastic tumor or epithelial trophoblastic tumor;2)initial therapy was prophylactic chemotherapy or hysterectomy;3)a follow-up period of less than two years.The medical records were reviewed and the clinical data such as general characteristics,therapeutic strategies,hCG results were collected.Four prediction models were constructed based on the hCG data of the spontaneous regression group(a sample of 1383 hCG values).The hCG data from post-molar GTN group was used to evaluate the clinical performance of each prediction model in predicting GTN.The weekly hCG regression standard curve and weekly hCG regression rate standard curve were constructed according to previous studies;the daily hCG regression standard curve and daily hCG regression rate standard curve were plotted based on the theory pioneered by us.Comparisons were conducted between the novel models and the existing models to evaluate the clinical value.Results:From a total of 222 patients,a natural regression was noted in 195 patients(spontaneous regression group)while the remaining 27 patients entered post-molar GTN(post-molar GTN group).All the patients with post-molar GTN were pre-identified by daily hCG regression rate standard curve,with an average of 15.3 days gained compared to FIGO criteria(P=0.008).The daily hCG regression standard curve had a sensitivity of 92.6%,with an average of 10 days gained(P=0.030).The weekly hCG regression rate standard curve had a sensitivity of 85.2%,with an average of 14 days gained(P=0.006).No significant differences were found between the weekly hCG regression standard curve and FIGO criteria.Conclusion:In comparison with the other three prediction models,our data showed that the daily hCG regression rate standard curve gives the best prediction of post-molar GTN,and might potentially enhance the monitoring of patients with molar pregnancy,especially those who could not adhere to FIGO protocols.Part IIThe clinical value of four hCG models in predicting post-molar gestational trophoblastic neoplasia:A prospective studyObjective:To prospectively study the clinical value of four hCG prediction models in predicting post-molar GTN.Materials and Methods:From 2011 to June 2015,Patients who were treated in our hospital and met the following criteria were prospectively shortlisted in the study:1)histologically confirmed diagnosis of CHM,2)and initial therapy was D&C.All patients were then followed for a minimum of two years,after which only those patients who had an intact data of pre-evacuation hCG level and hCG levels during the period of follow-up were finally included in this study.Based on the four hCG prediction models of Part I,the sensitivity,specificity,positive predictive value,negative predictive value,accuracy,Youden index,odds ratio,positive likelihood ratio and negative likelihood ratio,all of which were calculated and compared.Results:A total of 117 patients were selected.Of whom 16 patients(13.7%)developed post-molar GTN while 101 patients(86.3%)regressed naturally after initial treatment.For weekly hCG regression standard curve,7 patients with GTN were failed to predict and 9 patients who regressed naturally were failed to predict,with a sensitivity of 56.3%and a specificity of 91.1%.For weekly hCG regression rate standard curve,3 patients with GTN was failed to predict and 6 patients who regressed naturally were failed to predict,with a sensitivity of 81.2%and a specificity of 94.1%.For daily hCG regression standard curve,2 patients with GTN was failed to predict and 20 patients who regressed naturally were failed to predict,with a sensitivity of 87.5%and a specificity of 80.2%.For daily hCG regression rate standard curve,no patients with GTN were failed to predict and 2 patients who regressed naturally was failed to predict,with a sensitivity of 100%and a specificity of 98.1%.In general,daily hCG regression rate curve had the highest sensitivity,specificity and accuracy.Moreover,daily hCG regression rate curve was the only prediction model which met the criteria that positive likelihood ratio was over 10 and negative likelihood was ratio less than 0.1.Conclusion:Our data showed that daily hCG regression rate curve had the best performance in predicting post-molar GTN,and might be applied to monitor poor-compliant patients with CHM.Part ⅢThe clinical value of prophylactic hysterectomy for preventing malignant transformation of complete hydatidiform mole in patients at least 40 years old:a retrospective cohort studyObjective:To evaluate the clinical value of prophylactic hysterectomy in preventing post-molar gestational trophoblastic neoplasia(GTN)pertaining to patients with complete hydatidiform mole(CHM)who were at least 40 years old.Materials and Methods:Patients from our hospital who had histologically been diagnosed of CHM and underwent treatment from January 2004 to December 2013 were included.Inclusion criteria were:1)histologically confirmed diagnosis of CHM,2)being at least 40 years old of age,3)initial therapy was D&C and no evidence of residual tissue in the uterus,4)no evidence of local invasion or metastasis before initial therapy,5)follow-up for a minimum of two years.Relevant clinical data,such as clinical characteristics,traditional high risk factors,therapeutic methods,outcome of treatment,malignant rate of GTN,chemotherapy resistance rate and the time interval to GTN diagnosis were extracted and analyzed.Results:A total of 171 patients were included.The age ranged from 40 to 56 with a mean of 46.9.The mean parity was 1.2(range:0-4).The mean of gestational age was 9.5 weeks.All patients were divided into three groups based on the therapeutic strategy they received.Group 1 consisted of 124 patients,treated by dilation and curettage only,and the incidence of post-molar GTN was 37.1%.Group 2 included 12 patients who received prophylactic chemotherapy,with an incidence of 41.7%.The remaining 35 patients,Group 3,underwent total hysterectomy,with the lowest incidence of 11.4%.A significantly lower incidence was noted in group 3 as compared to group 1(χ2=8.342,P=0.004).No significant differences were noted in group 2 compared to group 1(P=0.035,P<0.017(0.05/3)were considered significant,Bonferroni correction).Conclusion:Prophylactic hysterectomy was the most effective therapeutic strategy for preventing malignancy of CHM in women being at least 40 years old of age.Part ⅣManagement of hydatidiform mole with prophylactic hysterectomy in women who were at least 40 years old:a systematic reviewand meta-analysisObjective:To conduct a systematic review and meta-analysis to evaluate the clinical value of total hysterectomy in preventing post-molar gestational trophoblastic neoplasia(GTN)pertaining to patients with HM who were at least 40 years old.Materials and Methods:Five electronic databases,including PubMed,Embase,the Cochrane database of systematic review,China National Knowledge Infrastructure and ClinicalTrials.gov,were searched.Studies were included if they:1)were human studies,2)explicitly indicated exposure to hysterectomy,3)explicitly indicated control to uterine evacuation,4)explicitly indicated the participants were older patients with hydatidiform mole being at least 40 years in age,5)compared the outcome of interest as the incidence of post-molar GTN.Studies were excluded if they were1)review articles,2)case report or case series,3)studies with data that could not be extracted to calculate a risk ratio.Meta-analysis was conducted based on the incidence of post-molar GTN.Results:Five studies with a total of 291 participants were included in the current meta-analysis.Two studies were conducted in USA and three studies were conducted in Asia(one in Thailand and two in China).The age of participants was at least 40 years old in three studies,40 to 49 in one study and 50 or older in one study.The case group included the group of patients who received total hysterectomy.The control group comprised patients who received only uterine evacuation.The sample size ranged from 22 to 159.The follow-up time was 3-14 years in one study and 3-12 years in another study,while for the other three studies,such data were not reported.The outcome of each study was predominantly focused on incidence of post-molar GTN.The overall pooled effect size of hysterectomy group had a significant advantage over uterine evacuation group with an OR of 0.14(95%CI,0.06-0.34;P<0.0004)and a low heterogeneity(I2=21%,P=0.28).Similar results were noted in the subgroup analysis based on the age and locale.No publication bias was found(Egger’s test,P=0.149;Begg’s test,P=0.462).Conclusion:Hysterectomy,comparing to uterine evacuation,is a better therapeutic method for treating patients with HM who are at least 40 years old.Therefore,total hysterectomy can be recommended for this age-specific group of patients in preventing post-molar GTN unless fertility is still desired.更多还原