基于SEER数据库构建子宫平滑肌肉瘤远处转移预测模型

Nomogram for prediction of distant metastasis in uterine leiomyosarcoma patients based on SEER database

目的:分析并筛选子宫平滑肌肉瘤(UL)患者发生远处转移的独立危险因素,并构建列线图预测远处转移的风险。方法:利用美国NCI监测、流行病学和结果数据库收集共计810例UL患者相关临床信息,按7∶3比例随机分成训练组(567例)及验证组(243例)。经单因素及多因素Logistic回归分析筛选变量,确定UL远处转移的独立危险因素,绘制列线图构建远处转移预测模型。通过一致性指数(C-index)、曲线下面积(AUC)和校准曲线评价预测模型的区分度和一致性。结果:单因素及多因素分析显示病理分化、肿瘤大小、手术及化疗为UL患者发生远处转移的独立危险因素(均P<0.05)。应用上述危险因素构建列线图预测模型,内部验证C-index为0.736,模型预测性能良好。结论:UL患者远处转移的列线图预测模型具有良好的预测精度,有助于临床医师评估患者远处转移风险,有针对性地实施个体化诊疗。

Objective:To analyze and screen the independent risk factors of distant metastasis in patients with uterine leiomyosarcoma,and construct nomogram to predict the risk of distant metastasis.Methods:A total of 810 patients with uterine leiomyosarcoma were collected from the National Cancer Institute's Surveillance,Epidemiology and End Results(SEER)database,and were randomly divided into training group(567 cases)and validation group(243 cases)according to the ratio of 7∶3.A nomogram was constructed by the dependent risk factors of distant metastasis of uterine leiomyosarcoma,which were determined by univariate and multivariate logistic regression analysis.The discrimination and ability of the prediction model were evaluated by C-index,AUC and calibration curve.Results:Univariate and multivariate logistic regression analysis showed that pathological differentiation,tumor size,surgery and chemotherapy were independent risk factors for distant metastasis of uterine leiomyosarcoma patients.A nomogram prediction model was established successfully and the C-index for internal validation was 0.736,which showed excellent prediction accuracy of the nomogram.Conclusion:This study had a high prediction accuracy,which was helpful for clinicians to assess the risk of distant metastasis of patients and implement individualized diagnosis and treatment.