ADC值直方图在子宫内膜癌病理组织学特征评估中的应用研究

发布时间：2018-03-21 01:59

本文选题：子宫内膜癌　切入点：MRI　出处：《吉林大学》2017年硕士论文　论文类型：学位论文

【摘要】：目的:探讨基于整个肿瘤体积的表观弥散系数(apparent diffusion coefficient,ADC)值直方图评估子宫内膜癌不同病理组织学特征及鉴别子宫内膜良恶性组织的价值。方法:收集2014年05月~2016年12月在吉林大学第一医院诊治并且经手术后病理确诊的51名子宫内膜癌患者入病例组;30名同一时期患有宫颈癌且术后病理确诊子宫内膜正常的患者入对照组。术前所有患者均行常规磁共振(Magnetic resonance imaging,MRI)序列、弥散加权成像(Diffusion weighted imaging,DWI)以及普通增强(MR contrast enhanced,CE-MR)扫描。将DWI扫描后自动生成的ADC图原始数据传输至后处理软件(Siemens Syngo),人工绘制ADC图每一层面感兴趣区(ROI)直至包括整个瘤体或正常子宫体内膜并由后处理软件自动生成相应层面的ADC值直方图,进而转换为频数分布表输入SPSS 20.0软件,重建出整个瘤体或正常子宫内膜的ADC值直方图进行数据分析。将病例组患者根据术后病理类型及组织学分级、肌层浸润的深度、有无宫颈间质浸润分组,比较各组间以及病例组与对照组的ADC值直方图参数,包括平均值(mean)、第5百分位数(P5)、第10百分位数(P10)、第25百分位数(P25)、第50百分位数(P50)、第75百分位数(P75)、第90百分位数(P90)及第95百分位数(P95)。两组间比较使用独立样本t检验,四组间比较使用单因素方差分析,P0.05被认为具有统计学意义。对以上参数进行受试者工作曲线分析(receiver operating characteristic curve,ROC)并计算曲线下面积(area under the ROC curve,AUC),获取最佳诊断截止值及相应的敏感度、特异度。结果:1.ADC值直方图参数鉴别病例组不同病理组织学特征的情况:(1)ADC P5、P10、P25值在子宫内膜样腺癌G1、G2、G3与非子宫内膜样癌中差异有统计学意义(P0.05);(2)子宫内膜样腺癌(G1+G2+G3)较非子宫内膜样癌的ADC P5值减低,差异有统计学意义(P0.05),其ROC曲线下面积AUC=0.654,取最佳截止值为719.00x10-6mm2/s,相应的敏感度、特异度分别为0.29、1.00;(3)深肌层浸润较浅肌层浸润的ADC mean、P5、P10、P25以及P50值均降低,差异有统计学意义(P0.05),其中ADC P5值鉴别肿瘤侵犯肌层深度的效能最高,其AUC=0.725,取最佳截止值为653.10x10-6mm2/s,相应的敏感度为0.51,特异度为1.00;(4)宫颈间质浸润阴性与阳性两组间的ADC值直方图各观察指标差异均无统计学意义(P0.05)。2.病例组与对照组的ADC mean、P_5、P_(10)、P_(25)、P_(50)、P_(75)、P_(90)及P_(95)值差异均有统计学意义(P0.05),其中P90的诊断效能最高,AUC=0.987,取最佳截止值为1438.90x10~(-6)mm~2/s,相应的敏感度、特异度分别为0.97、0.94。结论:1.基于整个肿瘤体积的ADC值直方图有助于预测子宫内膜癌高危因素患者,如子宫内膜样腺癌G3、非子宫内膜样癌以及深肌层浸润,但对于有无宫颈间质浸润则无明显差异。2.ADC值直方图有助于术前鉴别子宫内膜癌与正常子宫内膜,其中第90百分位数被认为是最有意义的参数。3.ADC值直方图可以对病变进行量化分析,为临床医生提供更精准的信息。
[Abstract]:Objective: To investigate the apparent diffusion coefficient of the tumor volume table (apparent diffusion coefficient ADC, based on histogram) assessment of endometrial carcinoma of different pathological features and differential diagnosis of primary endometrial benign and malignant tissue. Methods: from 2014 05 months ~2016 year in December in No.1 Hospital of Jilin University and by surgical pathology confirmed 51 endometrial cancer patients the case group; 30 the same period with cervical cancer and postoperative pathological diagnosis of uterine endometrial normal patients in the control group. All patients underwent conventional magnetic resonance (Magnetic resonance imaging, MRI) sequence and diffusion-weighted imaging (Diffusion weighted imaging DWI (MR contrast) and common enhanced enhanced, CE-MR scan) the automatic generation of DWI. After scanning ADC image of original data to the postprocessing software (Siemens Syngo), artificial drawing ADC of every level of region of interest (R OI) up to and including the whole tumor or normal uterus in vivo membrane and automatically generate the corresponding level ADC histogram by postprocessing software, and then converted into a frequency distribution table input SPSS 20 software to reconstruct the whole tumor or normal endometrium of ADC histogram is used to analyze the data. The patients according to the postoperative pathological type and histological grade, myometrial invasion depth, there is no cervical stromal invasion group were compared between groups, the case group and the control group of ADC histogram parameters, including the average value (mean), the fifth percentile (P5), the tenth percentile number (P10), the twenty-fifth percentile (P25) and the fiftieth percentile (P50), the seventy-fifth percentile (P75), the ninetieth percentile (P90) and 9 thousand and 500 percentile (P95). The two groups were compared using independent samples t test, the four groups were compared using single factor analysis of variance, P0.05 was considered to be statistically significant. On the above parameters. Analysis of curve line (receiver operating characteristic curve, by ROC) and calculate the area under the curve (area under the ROC curve, AUC), to obtain the best diagnostic cut-off value and the corresponding sensitivity and specificity. Results: the 1.ADC value of parameter identification cases with histogram of histopathological features: (1) ADC P5, P10, P25 in endometrial adenocarcinoma G1, G2, there are statistically significant differences between G3 and non endometrioid carcinoma (P0.05); (2) endometrial adenocarcinoma (G1+G2+G3) compared with non endometrioid carcinoma of the ADC P5 value decreased, the difference was statistically significant (P0.05). The area under the ROC curve of AUC=0.654, the optimal cut-off value of 719.00x10-6mm2/s, the sensitivity and specificity were respectively 0.29,1.00; (3) the depth of myometrial invasion shallow muscle layer infiltration of ADC mean, P5, P10, P25 and P50 were decreased, the difference was statistically significant (P0.05), ADC P5 the differential value of tumor invasion Make the muscle layer depth of AUC=0.725, the highest efficiency, the optimal cut-off value of 653.10x10-6mm2/s, the sensitivity was 0.51, specificity was 1; (4) cervical stromal invasion with positive and negative two group ADC values difference between each observation index histogram were not statistically significant (P0.05) cases of.2. group and control group ADC mean, P_5, P_ (10), P_ (25), P_ (50), P_ (75), P_ (90) and P_ (95) values of the differences were statistically significant (P0.05), the diagnostic efficiency of P90 is the highest, AUC=0.987, the optimal cutoff value for 1438.90x10~ (-6) mm~2/s. The corresponding sensitivity and specificity for 0.97,0.94. conclusion: 1. the tumor volume were ADC based on histogram is helpful to predict the patients with high risk factors of uterine endometrial carcinoma, such as endometrial adenocarcinoma G3, non uterine endometrioid carcinoma and deep myometrial invasion, but there is no cervical stromal invasion of.2.ADC was no significant difference the histogram is helpful to preoperative diagnosis Endometrial cancer and normal endometrium, of which ninetieth percentile is considered to be the most significant parameter..3.ADC histogram can make quantitative analysis of lesions and provide more accurate information for clinicians.

【学位授予单位】：吉林大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：R737.33;R445.2

【相似文献】