颅内动脉瘤血管内栓塞术后复发风险预测模型构建与验证

发布时间：2018-07-10 07:24

本文选题：动脉瘤介入栓塞 + 术后复发危险因素　；参考：《第二军医大学》2016年硕士论文

【摘要】：第一部分:颅内动脉瘤血管内栓塞术后复发的影响因素分析及复发风险评分表建立研究目的:本文旨在筛选颅内动脉瘤血管内栓塞术后复发的危险因素,并根据逻辑回归模型建立复发风险评分表,以简便直观地评价颅内动脉瘤患者经血管内栓塞治疗后的复发风险。研究方法:本文回顾性分析了自2012年5月至2014年5月间在第二军医大学长海医院脑血管病中心接受血管内栓塞治疗的颅内动脉瘤患者,共计441枚动脉瘤。根据随访时的复发情况将动脉瘤分为复发组和未复发组,收集与术后复发有关的患者相关因素、解剖学因素、技术及材料因素等,通过单因素分析比较两组之间上述相关因素是否具有差异。进一步采用logistic逐步回归分析对单因素分析结果中具有统计学意义的相关因素进行分析,筛选出与复发有关的危险因素并得到相应的复发风险回归模型,并根据回归模型的β系数对有统计学意义的影响因素进行赋值后建立术后复发风险评分表。采用ROC曲线下面积检验风险评分预测复发的有效性。研究结果:本研究共纳入动脉瘤441枚,其中复发组66枚,未复发组375枚。在单因素分析中,动脉瘤大小(P0.01)、破裂与否(P=0.04)、是否使用支架(P=0.002)、术后即刻栓塞结果(P=0.001)在两组之间具有统计学差异。年龄(P=0.521)、性别(P=0.377)、吸烟史(P=0.53)、高血压病史(P=0.781)、糖尿病病史(P=0.253)、瘤颈大小(P=0.216)、侧壁/分叉动脉瘤(P=0.661)、前循环/后循环动脉瘤(P=0.208)在两组之间无统计学差异。方差膨胀因子诊断法评价各因素间相关性结果显示:动脉瘤大小、破裂与否、是否使用支架、术后即刻栓塞结果的方差膨胀因子分别为1.041、1.323、1.296和1.146。Logistic逐步回归分析显示:动脉瘤大小(P0.001)、非支架辅助栓塞(P0.001)、术后即刻栓塞程度(P=0.012)具有统计学意义,其中动脉瘤25mm,动脉瘤10.1-25mm,非支架辅助栓塞,Raymond II和III的β系数分别为3.571,1.620,1.406,1.439,1.502.根据β系数与最小β系数的比值对危险因素进行赋值后得出复发风险评分表各因素得分为动脉瘤25mm-3分,动脉瘤10.1-25mm-1分,非支架辅助栓塞-1分,Raymond II和III-1分。利用ROC曲线下检验模型有效性,经检验复发风险logistic回归模型、复发风险评分表二者的ROC曲线下面积分别为:复发风险logistic回归模型-0.737、复发风险得分-0.735.二者无统计学差异。研究结论:颅内动脉瘤血管内栓塞术后复发的危险因素为动脉瘤大小、单纯栓塞、术后即刻栓塞程度,本研究建立的术后复发风险评分表能直观地反映患者术后复发风险大小。第二部分:颅内动脉瘤血管内栓塞术后复发风险评分表的验证和评价研究目的:本研究基于第一部分建立的颅内动脉瘤血管内栓塞术后复发风险评分表,为了进一步验证该评分表的预测复发风险的有效性,在第二部分中本研究通过建立验证队列,将验证队列动脉瘤进行评分后评价评分表预测颅内动脉瘤复发风险的效果有效性,同时与ARSS模型进行比较。研究方法:本部分回顾性分析了自2015年1月至2015年6月间在第二军医大学长海医院脑血管病中心接受血管内栓塞治疗的颅内动脉瘤患者,共计纳入109枚动脉瘤组成验证组病例用于本课题第二部分颅内动脉瘤血管内栓塞术后复发风险评分表的验证和评价。收集患者年龄、性别、吸烟史、高血压病史、糖尿病病史、动脉瘤大小、瘤颈大小、动脉瘤部位、破裂与否、是否使用支架、术后即刻栓塞结果等信息。将第一部分建模组与第二部分验证组基线资料进行比较,并将验证组109枚动脉瘤带入本研究第一部分所建立的复发风险评分表及Aneurysm Recanalization Stratification Scale进行评分后,利用ROC曲线下面积评价两种模型预测复发的有效性。利用评分表对本课题所有纳入的动脉瘤进行复发风险评分,将动脉瘤带入评分表评分后得出不同得分的复发率,比较两组不同得分间复发率,合并无统计学差异的得分将复发风险进行分层。将风险分层同真实复发结果进行比较,评价复发分层预测复发的敏感度和特异度。研究结果:将验证组患者基线资料与建模组比较结果显示,年龄、性别、吸烟、高血压病史、糖尿病病史、动脉瘤大小、破裂因素经检验P值分别为0.161、0.466、0.876、0.391、0.444、0.434、0.257,建模组和验证组患者基线资料经检验具有可比性。将109例动脉瘤带入本研究第一部分建立的复发风险评分表进行评分后结果显示,0分-20例,1分-48例,2分-37例,3分-2例,4分-2例。将109例动脉瘤带入ARSS评分后结果显示-1分-9例,0分11例,1分-28例,2分-13例,3分-31例,4分-10例,5分-6例,6分-1例。将复发风险评分及ARSS评分与实际复发情况比较,利用ROC曲线下面积检验其有效性,结果两种模型曲线下面积分别是:复发风险得分为73.8%;ARSS模型为69.0%.将本课题所有纳入动脉瘤进行评分后,利用卡方检验比较不同得分间复发率差异进行复发风险分层,不同得分的复发率分别为0分-1.59%,1分-8.90%,2分-28.79%,3分-55.55%,4分-83.33%。合并无统计学差异的得分后,将0-1分定义为低风险组,≥2分定义为高风险组。利用ROC曲线下面积检验复发风险分层预测有效性,结果显示曲线下面积为70.4%。结论:本研究建立的术后复发风险评分表可以有效地显示颅内动脉瘤术后复发风险大小。相较于现有的单因素结果预测,复发风险评分表的预测结果更为准确和稳定。
[Abstract]:The first part: analysis of the influencing factors of recurrence after intravascular embolization of intracranial aneurysm and the purpose of setting up a recurrence risk score table: This article aims to screen the risk factors of recurrence after intravascular embolization of intracranial aneurysms, and establish a recurrence risk score table according to the logistic regression model, so as to evaluate the patients with intracranial aneurysm easily and intuitively. The risk of recurrence after intravascular embolization. A retrospective analysis of 441 intracranial aneurysms from May 2012 to May 2014 at the cerebral vascular disease center, Changhai Hospital, Second Military Medical University. The aneurysms were divided into recurrent and non recurrent aneurysms according to the recurrence. The related factors, anatomical factors, technology and material factors related to the recurrence of postoperative recurrence were collected, and the differences between the two groups were compared by single factor analysis. The logistic stepwise regression analysis was used to analyze the relevant factors of statistical significance in the results of single factor analysis. The recurrence risk regression model was obtained with the recurrence related risk factors and the recurrence risk score table was set up after the evaluation of the statistically significant influencing factors according to the beta coefficient of the regression model. The effectiveness of the recurrence was predicted by the area test risk score under the ROC curve. The results of the study were included in the aneurysm 44. 1, of which 66 were relapsed and 375 were not recurred. In single factor analysis, aneurysm size (P0.01), rupture or failure (P=0.04), stent (P=0.002) and immediate postoperative embolization (P=0.001) were statistically different among the two groups. Age (P=0.521), sex (P=0.377), smoking history (P=0.53), hypertension history (P=0.781), diabetes history (P=0.253) the size of the tumor neck (P=0.216), the lateral wall / branched aneurysm (P=0.661), the anterior circulation / posterior circulation aneurysm (P=0.208) between the two groups. The variance expansion factor diagnostic method evaluated the correlation between the size of the aneurysm, the rupture or not, the use of the stent, and the variance expansion factor of the postoperative immediate embolization results. 1.041,1.323,1.296 and 1.146.Logistic stepwise regression analysis showed that aneurysm size (P0.001), non stent assisted embolization (P0.001) and postoperative immediate embolization (P=0.012) were statistically significant, of which aneurysm 25mm, aneurysm 10.1-25mm, non stent embolic plug, and Raymond II and III beta coefficients were 3.571,1.620,1.406,1.439,1.50. 2. according to the ratio of the beta coefficient and the minimum beta coefficient to the risk factors, the scores of the factors of the recurrence risk score were 25mm-3 score of aneurysm, 10.1-25mm-1 score of aneurysm, non stent assisted embolization -1, Raymond II and III-1. The validity of the model under the ROC curve was tested, and the recurrence risk logistic regression model was tested, and the recurrence was recurred. The area under the ROC curve of the risk score table two were the recurrence risk logistic regression model -0.737 and the recurrence risk score of -0.735. two. The study concluded that the risk factors of recurrence after intravascular embolization for intracranial aneurysms were aneurysm size, simple embolism, immediate postoperative embolism, and the postoperative recurrence was established in this study. The risk score table can directly reflect the recurrence risk of patients after operation. Second part: the purpose of verification and evaluation of the recurrence risk score of intracranial aneurysm after intravascular embolization: This study was based on the first part of the recurrence risk score of intracranial aneurysm after intravascular embolization in order to further verify the preview of the score The effectiveness of the risk of recurrence was measured in the second part. In this study, the effectiveness of predicting the recurrence risk of intracranial aneurysms in a cohort of aneurysms was verified by establishing a verification queue. The results were compared with the ARSS model. The methods of research were retrospectively analyzed from January 2015 to June 2015 in second. The patients with intracranial aneurysm treated with intravascular embolization in the center for cerebrovascular disease in Changhai Hospital, Military Medical University, were included in a total of 109 aneurysms to verify and evaluate the recurrence risk score table after intravascular embolization of intracranial aneurysms in the second part of the subject. The history of diabetes, the size of the aneurysm, the size of the aneurysm, the location of the aneurysm, the rupture or not, whether the stent was used, and the outcome of the immediate embolization. The first part of the modeling group was compared with the baseline data of the second part of the validation group, and 109 aneurysms in the verifying group were brought into the recurrence risk score table and Aneury established in the first part of this study. After the score of SM Recanalization Stratification Scale, the two models of area evaluation under the ROC curve were used to predict the recurrence. The score table was used to score the recurrence risk of all the aneurysm included in the subject, and the recurrence rate of the different scores was obtained after the score of the aneurysm, and the recurrence rate was compared between the two groups. Rate, stratified the risk of recurrence without statistical difference. Compare the risk stratification with the real recurrence results and evaluate the sensitivity and specificity of recurrent stratified prediction. Results of the comparison of the baseline data of the patients with the model group showed that the age, sex, smoking, history of hypertension, diabetes history, and movement were compared with the modeling group. The size of the aneurysm, the factor of rupture, the P value was 0.161,0.466,0.876,0.391,0.444,0.434,0.257, the baseline data of the model group and the verification group were comparable. 109 cases of aneurysm were taken into the first part of this study and the results of the recurrence risk score showed that 0 -20 cases, 1 -48 cases, 2 -37 cases, 3 -2 cases, 4 -2 cases, 109 cases of aneurysm were brought into ARSS score, and the results showed -1 -9 cases, 0 points, 1 -28, 2 -13 cases, 3 -31 cases, 4 -10 cases, 5 -6 cases, 6 -1 cases. The recurrence risk score and ARSS score were compared with the actual recurrence, and the area under the ROC curve was the recurrence of the recurrent area under the two model curve area respectively: recrudescent The risk score was 73.8%. After the ARSS model was used to score all the aneurysms in 69.0%., the recurrence rates of different scores were compared with the chi square test. The recurrence rates of different scores were 0 -1.59%, 1 -8.90%, 2 -28.79%, 3 -55.55%, and 4 -83.33%. with no statistical difference. The 0-1 point is defined as a low risk group, which is defined as a high risk group with more than 2 points. The area under the area of ROC curve is used to predict the effectiveness of the recurrence risk stratification. The results show that the area under the curve is 70.4%. conclusion: the postoperative recurrence risk score table established in this study can effectively show the risk of recurrence of intracranial aneurysms after operation. Results the prediction of recurrence risk score was more accurate and stable.
【学位授予单位】：第二军医大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：R651.12

【相似文献】