北京大学医学部机构知识库
Advanced  
IR@PKUHSC  > 北京大学第九临床医学院  > 学位论文
学科主题: 内科学
题名:
甲状腺乳头状癌风险评估策略的研究
作者: 郑晓
答辩日期: 2016-05-24
导师: 姜涛
专业: 内科学
授予单位: 北京大学
授予地点: 北京大学第九临床医学院
学位: 硕士
关键词: 甲状腺乳头状癌风险评分量表 ; 甲状腺结节良恶性鉴别 ; Logistic回归分析 ; BRAFV600E基因突变 ; ROC曲线
其他题名: Research about the risk assessment of Papillary Thyroid Carcinoma
分类号: R736.1
摘要:

[1] 研究目的:

1.分析探讨人口学特征、临床特征、影像学特点等各指标在甲状腺结节良恶性鉴别中的应用价值;2.建立甲状腺乳头状癌风险评分量表;3.验证甲状腺乳头状癌风险评分量表;4.比较甲状腺乳头状癌风险评分量表与甲状腺穿刺细胞学检查的准确性;5.探讨BRAFV600E基因在甲状腺癌及其淋巴结转移中的预测意义。

 

[2] 研究方法:

1.收集北京世纪坛医院2010年1月—2014年12月,5年期间收治的1003例经手术病理证实良恶性的甲状腺结节患者的临床资料,回顾性分析甲状腺结节患者的人口学特征、临床特征、影像学表现等特点,在超声检测的基础上结合被公认为甲状腺癌危险因素的临床资料和具有良恶性鉴别参考价值的实验室指标,采用回顾性病例对照研究,从以上三个层次探讨甲状腺恶性结节的高危因素。甲状腺良性结节与甲状腺癌之间年龄、BMI使用 t 检验,病程使用秩和检验,其余两组间参数采用两独立样本比较的 χ2检验,比较两组的参数有无统计学差异。

2.将有统计学差异的参数作为自变量,以甲状腺结节良恶性作为因变量,分别进行赋值以建立Logistic回归分析,根据其结果以及临床上指标获得的难易程度,选取合适的指标进入甲状腺乳头状癌风险评分量表,计算每一自变量对疾病风险贡献的大小并得到回归系数β,分别进行标准赋值,构建出甲状腺乳头状癌风险评分量表。最终得到每个危险系数的相应分值,各项相加即得每一受试者的风险评分,建立ROC曲线,以AUC评价该评分量表的诊断效能,并根据其灵敏度、特异度得出鉴别甲状腺结节良恶性的最佳分界值,进一步得出甲状腺癌危险分层的分界值。

3.收集北京世纪坛医院2015年1月—2015年12月期间收治的200例经手术病理证实良恶性的甲状腺结节患者临床资料,应用建立的甲状腺乳头状癌风险评分量表对其进行评分。按ROC曲线所得的最佳分界值将其分为两组(恶性组与良性组),与术后病理进行比较,计算出其灵敏度、特异度、阳性预测值、阴性预测值及准确性。同时用验证组病例建立ROC曲线,再次验证该量表的诊断效能及最佳分界值。

4.收集北京世纪坛医院2010年1月—2015年12月期间收治的79例经手术病理证实良恶性,同时具备术前甲状腺穿刺病理学检查结果的甲状腺结节患者临床资料。应用建立的甲状腺乳头状癌风险评分量表对其进行评分,按ROC曲线所得的最佳分界值将其分为两组(恶性组与良性组),与术后病理进行比较,分别计算甲状腺穿刺病理学检查和甲状腺乳头状癌风险评分量表的灵敏度、特异度、阳性预测值、阴性预测值及准确性,并建立ROC曲线比较两者的诊断效能。

5.收集北京世纪坛医院2010年1月—2015年12月期间收治的159例经手术病理证实良恶性,并且具有BRAFV600E基因突变分析的甲状腺结节患者临床资料,以BRAFV600E基因突变阳性(突变型)为病例组,BRAFV600E基因突变阴性(野生型)为对照组,比较两者之间恶性率的高低。进一步回顾性分析甲状腺结节患者的超声征象、病理特征、免疫组化等特点。两组之间年龄、肿瘤最大径使用t检验,其余两组间参数采用两独立样本比较的 χ2检验,比较两组的参数有无统计学差异。

 

    [3] 研究结果:

1.一般情况分析:1003例甲状腺结节患者中年龄11~82岁之间,甲状腺恶性肿瘤455例(45.4%),均为PCT,平均年龄为46.35±12.3岁,恶性肿瘤高峰年龄为40-50岁;甲状腺良性结节548例(54.6%),平均年龄为50.39±11.9岁,良性结节高峰年龄为50-60岁(t=-5.260,P=0.000)。所有患者中男性254例,占25.2%,女性749例,占74.8%。男女比例约为1:3,但性别在甲状腺结节良恶性鉴别中无明显统计学差异(P=0.638)。甲状腺恶性肿瘤患者BMI均数±标准差为24.0621±3.75 Kg/m2,良性结节患者均数±标准差为24.1810±3.20 Kg/m2,两者间无统计学差异(P=0.742)。

2.病理类型分析:1003例甲状腺结节患者中,甲状腺乳头状癌455例(45.4%),肿瘤最大径为11.14±8.782mm,其中多灶病变有179例(39.2%),伴淋巴结转移者有160例(35.2%),气管食管沟处最常见部位(70.6%),微小癌316例(69.6%),包膜侵犯者155例(34.2%),癌周伴结节性甲状腺肿有280例(61.5%),癌周伴桥本甲状腺炎有60例(13.1%);甲状腺良性结节 548例(54.6%),肿瘤最大径为26.99±14.759mm,以结节性甲状腺肿为主(48.7%),单发占281例(57.4%),多发占208例(42.6%)。

3.甲状腺乳头状癌风险评分量表的建立:对纳入研究的1003例患者按结节良恶性分组,进行临床资料分析,差异有统计学意义的指标有年龄、病程等23项,进一步使用Logistic回归分析,筛选出可用于甲状腺癌风险评估的最优筛查指标共9项:年龄、临床指标(结节触诊大小、结节形态、结节边界)、超声指标(结节性质、形态、边界、钙化形态、血流信号),构建成甲状腺乳头状癌风险评分量表,并得出了各项目相应分值,总分值为19分。建立ROC曲线,得出该量表的AUC=0.854 > 0.8,故认为该量表诊断甲状腺癌的效能较高。将7分作为甲状腺乳头状癌风险评分量表鉴别甲状腺结节良恶性的诊断界点时,灵敏度为67.3%,特异度为87.1%,≥7分者恶性率高达83.3%,<7分者恶性率仅为29.2%。进一步进行甲状腺癌危险分层,将评分结果分为4类:评分≥ 8分者为极高危(恶性率> 90%),评分6~7分者为高危(恶性率50~90%),评分3~5分者为中危(恶性率10%~50%),评分≤ 2分者为低危(恶性率< 10%)。

4.甲状腺乳头状癌风险评分量表的验证:以ROC曲线所得出的良恶性分界值7分为诊断界点,将200例验证组患者分为恶性组(评分≥7分)和良性组(评分<7分),与术后病理对比,计算出其灵敏度、特异度、阳性预测值、阴性预测值及准确性分别为:80%、89.3%、91.1%、76.5%、83.5%,用验证组病例建立的ROC曲线下AUC=0.906 > 0.8,故再次验证该量表诊断甲状腺癌的效能较高。此表中选7分为诊断界点时,灵敏度为84.3%,特异度为84.5%。用验证组进行的危险分层情况同样符合上述危险分层:评分≥ 8分者,恶性率在90%以上;评分6~7分者,恶性率为50~90%;评分3~5分者,恶性率为10%~50%;评分≤ 2分者,恶性率在10%以下。

5.甲状腺乳头状癌风险评分量表与甲状腺穿刺细胞学检查的准确性的比较:与术后病理对比,甲状腺穿刺病理学检查的灵敏度、特异度、准确性分别为85%、92.9%、86.5%,甲状腺乳头状癌风险评分量表的灵敏度、特异度、准确性分别为78.7%、70.6%、76.9%。甲状腺穿刺细胞学检查诊断的AUC=0.880 > 0.8,甲状腺乳头状癌风险评分量表诊断的AUC= 0.807 > 0.8。前者选取1.50作为诊断界点时,灵敏度为85.0%,特异度为92.9%,后者选取7分作为诊断界点时,灵敏度为82.0%,特异度为70.6%。

6.BRAF基因在甲状腺癌及其淋巴结转移中的预测意义:BRAFV600E基因突变型患者恶性肿瘤预测率为100%,BRAFV600E基因野生型患者恶性肿瘤预测率为51.5%,两者对于甲状腺癌的预测价值具有显著统计学差异(χ2=59.876,P=0.000 < 0.05)。BRAFV600E基因的不同类型对于预测甲状腺癌是否伴有淋巴结转移有统计学意义(χ2=59.876,P=0.000 < 0.05)。BRAFV600E的不同类型与患者包膜侵犯、淋巴结转移、超声下低回声有关(P < 0. 05),与患者年龄、性别、病灶大小、数目,以及超声下结节大小、性质、单纯微钙化、血流信号无关(P均 > 0. 05)。免疫组化结果显示BRAFV600E突变型患者中CK19、Galectin3的阳性率均为100%(83/83)(72/72),KI-67的阳性率为41.7%,均高于野生型(P均 < 0.05);反之,CD56在BRAFV600E野生型患者中的阳性率为59.1%(26/44),显著高于突变型(P=0.000 < 0.05)。

 

[4] 结论:

1.甲状腺结节患者以女性居多,但性别对于甲状腺良恶性肿瘤的鉴别无统计学意义。甲状腺恶性结节的高峰年龄为40-50岁,良性结节的高峰年龄为50-60岁。

2.甲状腺恶性肿瘤以PTC为主,淋巴结转移率为35.2%,主要以气管食管沟为主。

3.甲状腺乳头状癌风险评分量表:总分值为19分,根据ROC曲线得出:该量表的良恶性诊断界点为7分,评分≥7分者,甲状腺结节恶性率高达90%,需接受进一步的治疗干预。危险分层可分为极高危(评分≥ 8分者,恶性率>90%),高危(评分6~7分者,恶性率50%~90%),中危(评分3~5分者,恶性率10%~50%),低危(评分≤ 2分者,恶性率<10%)。

4.甲状腺危险评分量表的灵敏度、特异度、准确性均高于75%,AUC=0.807 > 0.8,与甲状腺穿刺病理学检查差异不大,故在临床上具有十分重要的应用前景,评分<7分者可以定期观察,评分≥7分者则先行甲状腺穿刺细胞学检查,结果提示恶性或恶性可能性大时,再进一步行手术治疗。因此甲状腺危险评分量表的广泛应用能有效地减少临床上对甲状腺结节患者进行的有创操作及不必要的手术治疗。

5.BRAFV600E基因突变型患者恶性肿瘤发生率明显高于野生型患者,两组之间的不同特征有包膜侵犯、淋巴结转移、超声下低回声。CK19、Galectin3、KI-67与BRAFV600E突变类型呈正相关,CD56与BRAFV600E突变类型呈负相关。

英文摘要:

 

Objective:1. Analyze various indicators of demography, clinical features, imaging findings in the identification bewteen benign and malignant thyroid nodules; 2. Establish a risk rating scale about Papillary Thyroid Carcinoma; 3. Validate the risk rating scale about Papillary Thyroid Carcinoma; 4. Compare the accuracy of risk rating scale about thyroid carcinoma and  thyroid fine needle aspiration cytology(FNAC) examination; 5. Explore BRAF gene mutation and its significance in prediction of thyroid carcinoma and lymph node metastasis.

 

Methods:1. Collect the clinical data of 1003 thyroid nodule cases in Beijing Shijitan hospital during the period of 5 years from January 2010 to December 2014, whose pathological result confirmed by surgery pathology. Analyzed retrospectively demographic acteristics, clinical features and imaging findings of those patients. On the basis of thyroid ultrasound, combine with the recognized thyroid carcinoma risk factors and laboratory indexes of benign and malignant thyroid nodule identification, we conduct this retrospective case-control study to investigate the risk of malignant thyroid nodule. We acquire the statistically significant by using t test in age, BMI between two groups, rank-sum test in course, chi-square test about two independent sample comparison in the rest parameters.

Use parameters with statistical significant as the independent variable, benign and malignant thyroid nodule as the dependent variable to establish Logistic regression analysis. According to the results and difficulty to get the clinical indicators, choose appropriate indicators into the risk rating scale about Papillary Thyroid Carcinoma, then get and assign the regression coefficient of beta on based of calculating contribution of each variable, which use to build up the risk rating scale about Papillary Thyroid Carcinoma. Eventually get corresponding points of each risk factor and tatal points. Establish the ROC curve to judge its diagnostic effectiveness in accordance with AUC. According to its best sensitivity and specificity to choose the optimum boundary to identificate the benign and malignant thyroid nodule and risk stratification.

Collect the clinical data of 200 thyroid nodule cases in Beijing Shijitan hospital during January to December 2015, whose pathological result confirmed by surgery pathology. Calculate the score with the risk rating scale about Papillary Thyroid Carcinoma. According to the optimum boundary of ROC curve to divide the cases into two groups (malignant group and benign group), and calculate sensitivity, specificity, positive predictive value, negative predictive value and accuracy. Setting up ROC curve with the validation group cases to validate the risk rating scale about Papillary Thyroid Carcinoma.

Collect the clinical data of 79 thyroid nodule cases in Beijing Shijitan hospital during January 2010 to December 2015, whose pathological result confirmed by surgery pathology and FNAC. Calculate the score with the risk rating scale about Papillary Thyroid Carcinoma. According to the optimum boundary of ROC curve to divide the cases into two groups (malignant group and benign group), and calculate and compare with the sensitivity, specificity and accuracy of FNAC and the risk rating scale respectively. Finally we establish the ROC curve to compare their diagnostic performance.

5. Collect the clinical data of 159 thyroid nodule cases with BRAFV600E mutation analysis in Beijing Shijitan hospital during January 2010 to December 2015, whose pathological result confirmed by surgery pathology. Patients with BRAFV600E mutation positive (mutant) as the case group, the rest with BRAFV600E mutation negative (wild type) as the control group. Compare with the malignant rates and analyze the ultrasound manifestations,  pathological features and immunohistochemical acteristics retrospectively between two groups. We acquire the statistically significant by using t test in age and tumor diameter between two groups, chi-square test about two independent sample comparison in the rest parameters.

Results:1. General situation analysis:1003 patients with thyroid nodules range from 11 to 82 years old. There are 455 thyroid malignant tumors (45.4%), their mean age are 46.35±12.3y, the peak age of malignant tumor is 40 to 50 years old, the rest 548 cases are thyroid benign nodules(54.6%), their mean age are 50.39±11.9y, the peak age of benign tumors is 50 to 60 years old (t=5.260, P=5.260).There are 254 male cases(25.2%) and 749 female cases(74.8%), their ratio is 1:3, but gender has no statistical significance to identificate the  benign and malignant thyroid nodules(P=0.638). Mean BMI of malignant thyroid tumor patients is 24.0621±3.75 Kg/m2, benign thyroid tumor patients is 24.1810±3.20 Kg/m2(P = 0.329), there has no statistical significance in BMI.

< > analysis: 455/1003 cases(45.4%) are thyroid papillary carcinoma, tumor mean diameter is 11.14±8.782 mm. Among those cases,179 cases(39.2%) are multifocal lesions, 160 cases(35.2%) have lymph node metastasis, the most common metastasis site is the tracheoesophageal groove(70.6%), 316 cases(69.6%) are papillary thyroid microcarcinoma, 155 cases(34.2%) have capsular invasion. Which is thyroid carcinoma with thyroid nodular goiter are 280 patients(61.5%), and thyroid carcinoma with hashimoto thyroiditis are 60 cases(13.1%); 548/1003 cases(54.6%) are thyroid benign tumor, tumor mean diameter is 26.99 ±14.759 mm (P < 0.05), mainly consisting of thyroid nodular goiter (48.7%), there are 281(57.4%) single lesions and 208 multifocal lesions(42.6%).

< >stablishing the risk rating scale about Papillary Thyroid Carcinoma: we divide the 1003 patients into two groups (malignant group and benign group), and do statistical analysis to get 23 indicators that have statistical significance such as age, course. Further using Logistic regression analysis to choose 9 optimal indicators for the risk rating scale about Papillary Thyroid Carcinoma: age, clinical indicators(nodule palpate size, nodular palpate form, nodules palpate boundary), ultrasonic properties(property, form, boundary, calcified form, blood flow signals), and each various projects is obtained corresponding score, there are 19 points in total. On based of the ROC curve, we know that AUC = 0.854 > 0.8, so that the efficiency of the risk rating scale to diagnose thyroid carcinoma is high, and we get the result that 7 point is the best boundary point to identify benign and malignant thyroid nodules, whose sensitivity is 67.3% and specificity is 87.1%, the malignant rate of patients, whose score is higher than 6 point, is as high as 83.3%, and those less than 7 point is only 29.2%.Further we divide the scores of the risk rating scale into three categories: 1.very high risk group: their score is higher than 7 points and their malignant rate reach up to 90%; 2.high risk group: their score are 6 ~ 7 points and their malignant rate range from 50% ~ 90%; 3.moderate risk group: their score are 3 ~ 5 points and their malignant rate range from 10% ~ 50%; 4.low risk group: their score is less than 3 points and their malignant rate are below 10%.

Validate the risk rating scale about Papillary Thyroid Carcinoma: Using 7 point as the boundary to divide the 200 validation group cases into two groups (malignant group ≥ 7 point and benign group < 7 point). Compared with postoperative pathology, we calculate the sensitivity, specificity, positive predictive value, negative predictive value and accuracy are 80%, 89.3%, 91.1%, 76.5% and 91.1%. On based of the ROC curve, we know that AUC = 0.906 > 0.8, which verify the efficiency of the scale in the diagnosis of thyroid carcinoma is high as well. And we also turn out that 7 point is the best boundary point to identify benign and malignant thyroid nodules, the sensitivity and specificity are 84.3% and 84.5%. Risk stratification of validation cases is the same with the above risk stratification: 1.very high risk group: score > 7 points and malignant rate > 90%; 2.high risk group: 6 ~ 7 points and malignant rate range from 50% ~ 90%; 3.moderate risk group: 3 ~ 5 points and malignant rate range from 10% ~ 50%; 4.low risk group: < 3 points and malignant rate < 10%.

5. Comparing the accuracy of the risk rating scale about Papillary Thyroid Carcinoma and thyroid fine needle aspiration cytology(FNAC): Compared with postoperative pathology, we calculate the sensitivity, specificity and accuracy of FNAC are 85%, 92.9% and 85% respectively, while the sensitivity, specificity and accuracy of the risk rating scale are 78.7%, 70.6% and 76.9% respectively. AUC of the FNAC = 0.880 > 0.8, and AUC of the risk score scale = 0.807 > 0.8. When choosing 1.50 as the diagnostic point of FNAC, the sensitivity and specificity are 85.0% and 92.9%, choosing 7 point as the diagnostic point of the risk rating scale, the sensitivity and specificity are 82.0% and 70.6%.

6. BRAFV600E gene mutation and its significance in prediction of thyroid carcinoma and lymph node metastasis: The prediction rate of malignant tumor patients with BRAFV600E mutation type is 100%, and the prediction rate of malignant tumor patients with BRAFV600E wild type is 51.5%. There is statistical significance between the two types to predict the thyroid carcinoma( χ2 = 59.876, P = 59.876 < 0.05). And there is also statistical significance between the different types of BRAFV600E genes to predict whether thyroid carcinoma with lymph node metastasis( χ2 = 59.876, P = 0.000 < 0.05). The different types of BRAFV600E is connected with capsular invasion, lymph node metastasis and ultrasonic echo(P < 0. 05), and it is unconnected with patients' age, gender, lesion size, lesion number, and nodule diameter, nature, microcalcification and blood flow signal under ultrasound(all P > 0. 05). Immunohistochemical results show that the positive rate of CK19 and Galectin3 in patients with BRAFV600E mutation type are 100% (83/83) and (72/72), the positive rate of KI - 67 is 41.7%, which is higher than patients with BRAFV600E wild type (P < 0.05); Conversely, the positive rate of CD56 in patients with BRAFV600E wild type is 59.1% (26/44), which is higher than BRAFV600E mutation type(P = 0.000 < 0.05).

Conclusions:1. Thyroid nodules used to appear mostly in female, but gender has no statistical significance in identification of benign and malignant thyroid nodules.The peak age of malignant tumors is 40 to 50 years old, the peak age of benign tumors is 50 to 60 years old.

2. Papillary thyroid carcinoma is the most common type in thyroid malignant tumor, its lymph node metastasis rate is 35.2%, the mainly metastasis site is tracheoesophageal groove.

3. The risk rating scale about Papillary Thyroid Carcinoma: There are 19 points in total. According to the ROC curve, we obtain the fact that 7 point is the best boundary point to identify benign and malignant thyroid nodules. When the score is higher than 7 points, the malignant rate of thyroid nodule is as high as 90%, the patient need to accept further treatment intervention. Risk stratification: 1.very high risk group: score > 7 points and malignant rate > 90%; 2.high risk group: 6 ~ 7 points and malignant rate range from 50% ~ 90%; 3.moderate risk group: 3 ~ 5 points and malignant rate range from 10% ~ 50%; 4.low risk group: < 3 points and malignant rate < 10%.

4. The sensitivity, specificity and accuracy of the risk rating scale about Papillary Thyroid Carcinoma are higher than 75%, AUC = 0.807 > 0.8, and the diffirence between the risk rating scale and FNAC is small. Therefore, the risk rating scale has very important application prospect in clinic, we consider that when the patient’s score < 7 points, maybe observating regularly is OK for him; When his score ≥ 7 points, he had better to do the FNAC, if the result of FNAC is malignant or suspected malignancies, he maybe need a further surgery. So the wide application of the risk rating scale about Papillary Thyroid Carcinoma can reduce the invasive operation and unnecessary surgery for patients with thyroid nodules effectively in clinically.

5. The incidence of malignant tumors in patients with BRAFV600E mutation types is significantly higher than BRAFV600E wild types, the different acteristics between the two groups have capsular invasion, lymph node metastasis and ultrasonic echo.CK19,  Galectin3 and KI - 67 are positively correlated with the BRAFV600E mutation types, while CD56 are negatively correlated with BRAFV600E mutation types.

语种: 中文
相关网址: 查看原文
内容类型: 学位论文
URI标识: http://ir.bjmu.edu.cn/handle/400002259/124808
Appears in Collections:北京大学第九临床医学院_学位论文

Files in This Item:

There are no files associated with this item.


作者单位: 北京大学第九临床医学院

Recommended Citation:
郑晓. 甲状腺乳头状癌风险评估策略的研究[D]. 北京大学第九临床医学院. 北京大学. 2016.
Service
Recommend this item
Sava as my favorate item
Show this item's statistics
Export Endnote File
Google Scholar
Similar articles in Google Scholar
[郑晓]'s Articles
CSDL cross search
Similar articles in CSDL Cross Search
[郑晓]‘s Articles
Related Copyright Policies
Null
Social Bookmarking
Add to CiteULike Add to Connotea Add to Del.icio.us Add to Digg Add to Reddit

Items in IR are protected by copyright, with all rights reserved, unless otherwise indicated.

 

 

Valid XHTML 1.0!
Copyright © 2007-2017  北京大学医学部 - Feedback
Powered by CSpace