刘丹丹,段卡丹,李朝辉,等.胃肠道间质瘤术后复发转移危险因素及预测模型构建[J].安徽医药,2023,27(2):297-302. |
胃肠道间质瘤术后复发转移危险因素及预测模型构建 |
Risk factors of recurrence and metastasis after gastrointestinal stromal tumor surgery and establishment of prediction model |
|
DOI:10.3969/j.issn.1009-6469.2023.02.019 |
中文关键词: 胃肠道间质瘤 手术治疗 复发转移 危险因素 预测模型 |
英文关键词: Gastrointestinal stromal tumor Surgical treatment Recurrence and metastasis Risk factor Prediction model |
基金项目:河南省医学科技攻关计划(联合共建)项目( LHGJ20191219) |
|
摘要点击次数: 1402 |
全文下载次数: 644 |
中文摘要: |
目的探讨胃肠道间质瘤( GIST)术后复发转移的危险因素,建立复发转移预测模型,为 GIST病人术后复发转移的预防控制提供参考。方法选取 2015年 1月至 2018年 4月洛阳市中心医院 GIST手术病人 260例,随访资料完整者共 247例,对病人随访 2年,观察其复发和转移情况,采用 logistic回归分析分析病人术后 2年复发和转移的危险因素,基于危险因素的回归系数构建风险预测模型,采用受试者工作曲线( ROC)评估预测模型的区分度。结果临床与随访资料完整的 247例 GIST病人中共有 74例( 29.96%)出现复发或者转移,其中 37例( 50.00%)为单纯复发, 28例( 37.84%)为单纯转移,其余 9例( 12.16%)为复发并且转移;不同性别、年龄、浸润深度以及 CD117、Dog-1、CD34表达 GIST病人术后复发转移情况比较,差异无统计学意义(P>0.05); logisitic回归分析显示肿瘤原发部位在胃部以外的小肠或腹膜与直肠、手术 R1或 R2切除、肿瘤长径大、核分裂象数高、 GIST危险度高、 Ki67指数高是 GIST病人术后复发转移发生的独立危险因素,术后服用伊马替尼是术后未发生复发转移的保护因素; GIST病人术后两年复发转移概率值回归方程为: P=1/[1+e -(-4.753+0.849×肿瘤原发部位+1.620×手术根治情况+0.622×肿瘤大小+0.777×GIST危险度分级+1.007×核分裂象数+1.285×Ki67-0.647×术后服用伊马替尼)]采用 Hosmer-Lemeshow检验检测回归方程的拟合优度( P=0.210);对模型进行内部验证,其 ROC曲线下面积( AUC)为 0.81,95%,CI:(0.76,0.87),灵敏度为 73.05%,特异度为 75.70%;R语言建立的 Nomogram图,能迅速评估 GIST术后复发转移发生率。结论风险评分模型具有较好的判别效度,可用于识别 GIST术后复发转移高危人群,以期早期预防与控制。 |
英文摘要: |
Objective To explore the risk factors of recurrence and metastasis after gastrointestinal stromal tumor (GIST) surgery,and establish a prediction model for recurrence and metastasis, so as to provide reference for their prevention in GIST patients.Meth. ods A total of 260 patients who underwent GIST surgery in Luoyang Central Hospital from January 2015 to April 2018 were enrolled,including 247 cases with complete follow-up data. The patients were followed up for 2 years to observe their recurrence and metastasis.The risk factors of recurrence and metastasis at 2 years after surgery were analyzed by logistic regression analysis. The risk predictionmodel was established based on regression coefficients of risk factors. The discrimination of the prediction model was evaluated by re-ceiver operator characteristic (ROC) curves.Results Of the 247 GIST patients with complete clinical and follow-up data, there were74 cases (29.96%) with recurrence or metastasis, including 37 cases (50.00%) with simple recurrence, 28 cases (37.84%) with simplemetastasis and the remaining 9 cases (12.16%) with recurrence and metastasis. There were no significant differences in postoperativerecurrence and metastasis among GIST patients with different gender, age, invasion depth, CD117, Dog-1 and CD34 expression (P> 0.05). Logisitic regression analysis showed that the primary site of tumors at small intestine or peritoneum and rectum outside of thestomach, surgical R1 or R2 resection, long tumor diameter, high mitotic figures, high GIST risk and high Ki67 index were independentrisk factors of postoperative recurrence and metastasis in GIST patients, while taking imatinib after surgery was a protective factor of norecurrence or metastasis. The regression equation for recurrence and metastasis probability in GIST patients at 2 years after surgery was follow: P=1/[1+e -(-4.753+0.849×primary tumor site+1.620×radical surgery condition+0.622×tumor size+0.777×GIST risk grade+1.007×number of mitotic figures+1.285×Ki67-0.647×taking imatinib after surgery)]. The good‐as ness of fit of the regression equation was detected by Hosmer-Lemeshow test (P=0.210). Internal verification of the model was conduct‐ed.The area under the ROC curve (AUC), sensitivity and specificity were 0.81,95%CI:(0.76,0.87),73.05 % and 75.70 %, respectively.Nomogram graphics established by R language could quickly assess the incidence of recurrence and metastasis after GIST. Conclu. sion The established risk scoring model is of good discriminative validity, which can be applied to identify high-risk groups with re-currence and metastasis after GIST surgery so as to conduct early prevention and control. |
查看全文
查看/发表评论 下载PDF阅读器 |
关闭 |
|
|
|