中国现代神经疾病杂志 ›› 2021, Vol. 21 ›› Issue (3): 147-155. doi: 10.3969/j.issn.1672-6731.2021.03.005

• 神经外科疾病大数据 • 上一篇    下一篇

2 基于美国SEER数据库的儿童胶质瘤临床特点和预后分析

石易鑫, 马文斌, 王裕   

  1. 100730 中国医学科学院 北京协和医学院 北京协和医院神经外科
  • 收稿日期:2021-03-08 出版日期:2021-03-25 发布日期:2021-04-02
  • 通讯作者: 王裕,Email:ywang@pumch.cn

Summary and analysis of clinical features and prognosis in pediatric gliomas based on the SEER database

SHI Yi-xin, MA Wen-bin, WANG Yu   

  1. Department of Neurosurgery, Peking Union Medical College Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing 100730, China
  • Received:2021-03-08 Online:2021-03-25 Published:2021-04-02

摘要:

目的 基于美国SEER数据库回顾总结儿童胶质瘤的临床特点,并探究不同组织学类型的生存预后相关影响因素。方法 纳入SEER数据库中2000-2015年经组织病理学证实的胶质瘤患儿共7759例,根据组织学类型分为毛细胞型星形胶质瘤、髓母细胞瘤、室管膜瘤、胶质母细胞瘤及其他类型胶质瘤,绘制Kaplan-Meier生存曲线,比较毛细胞型星形细胞瘤、髓母细胞瘤、室管膜瘤和胶质母细胞瘤患儿的5年生存率;采用单因素log-rank检验和多因素向前引入法Cox回归分析筛查上述4种组织学类型患儿生存预后相关影响因素。结果 本组7759例患儿根据组织学类型分为毛细胞型星形细胞瘤2585例(33.32%)、髓母细胞瘤2061例(26.56%)、室管膜瘤777例(10.01%)、胶质母细胞瘤443例(5.71%)、其他类型胶质瘤1893例(24.40%)。不同组织学类型之间性别(χ2=60.390,P=0.000)、年龄分布(χ2=600.318,P=0.000)、肿瘤大小(χ2=90.773,P=0.000)和原发部位(χ2=2117.948,P=0.000)、病理分级(χ2=1233.506,P=0.000)差异均有统计学意义,其中,髓母细胞瘤和胶质母细胞瘤好发于男性[60.89%(1255/2061)和57.56%(255/443)];学龄期为高峰发病年龄[35.48%(2753/7759)];毛细胞型星形细胞瘤和髓母细胞瘤以小肿瘤为主[31.30%(809/2585)和33.77%(696/2061)]且好发于小脑[42.24%(1092/2585)和59.58%(1228/2061)],室管膜瘤和胶质母细胞瘤则较大[32.43%(252/777)和29.57%(131/443)]且多发生于幕上及脑室[36.81%(286/777)和65.91%(292/443)],仅毛细胞型星形细胞瘤以低级别为主[24.06%(622/2585)]。生存分析显示,毛细胞型星形细胞瘤、髓母细胞瘤、室管膜瘤和胶质母细胞瘤患儿5年生存率分别为97.41%(2518/2585)、64.39%(1327/2061)、76.83%(597/777)和16.25%(72/443),不同组织学类型之间差异有统计学意义(χ2=2145.672,P=0.000)。单因素log-rank检验和多因素Cox回归分析显示,不同组织学类型胶质瘤预后相关影响因素不尽一致,主要包括确诊年龄、肿瘤大小和原发部位、病理分级、手术切除和放疗,除毛细胞型星形细胞瘤外,髓母细胞瘤、室管膜瘤和胶质母细胞瘤患儿接受手术切除均对生存预后有积极意义;放疗可降低髓母细胞瘤患儿死亡风险,但可能增加毛细胞型星形细胞瘤患儿的死亡风险。结论 毛细胞型星形细胞瘤比例最高,预后较好;胶质母细胞瘤发病率相对较低,预后较差。人口统计学信息、肿瘤特征和治疗方案均为儿童胶质瘤患者生存预后的预测因素,手术切除可降低除毛细胞型星形细胞瘤外的其他组织学类型胶质瘤患儿的死亡风险;放疗可降低髓母细胞瘤患儿的死亡风险,但有可能增加毛细胞型星形细胞瘤患儿的死亡风险。

关键词: 神经胶质瘤, 儿童, 数据库, 神经外科手术, 放射疗法, 预后, 存活率分析, 回归分析

Abstract:

Objective Based on the SEER (Surveillance, Epidemidogy, and End Results) database, this study retrospectively summarized clinical characteristics, and explored prognostic factors of patients with different histological types of pediatric gliomas. Methods A total of 7759 pediatric patients with gliomas confirmed by histopathology from the SEER database from 2000 to 2015 were included in the study and were classified into pilocytic astrocytoma, medulloblastoma, ependymoma, glioblastoma and other gliomas according to histological classification. Kaplan-Meier survival analysis was utilized to compare 5-year survival rates. Univariate log-rank test and multivariate forward selection Cox regression analysis were used to screen the factors related to survival and prognosis. Results According to histological classification, 7759 patients were divided into 2585 cases of pilocytic astrocytoma (33.32%), 2061 cases of medulloblastoma (26.56%), 777 cases of ependymoma (10.01%), 443 cases of glioblastoma (5.71%) and 1893 cases (24.40%) of other types of gliomas. The differences of gender (χ2=60.390, P=0.000), age distribution (χ2=600.318, P=0.000), tumor size (χ2=90.773, P=0.000), location (χ2=2117.948, P=0.000) and pathological grade (χ2=1233.506, P=0.000) in different histological types were statistically significant respectively. Among them, medulloblastoma (60.89%, 1255/2061) and glioblastoma (57.56%, 255/443) were more likely to occur in male and school age was the peak age of onset (35.48%, 2753/7759); pilocytic astrocytoma and medulloblastoma were mainly small tumors (31.30%, 809/2585; 33.77%, 696/2061) and more likely occurred in cerebellum (42.24%, 1092/2585; 59.58%, 1228/2061); while ependymoma and glioblastoma were larger (32.43%, 252/777; 29.57%, 131/443) and were occurred in supratentorial and cerebral ventricles (36.81%, 286/777; 65.91%, 292/443); only pilocytic astrocytoma was mainly of low grade (24.06%, 622/2585). Survival analysis showed 5-year survival rates of patients with pilocytic astrocytoma, medulloblastoma, ependymoma and glioblastoma were 97.41% (2518/2585), 64.39% (1327/2061), 76.83% (597/777) and 16.25% (72/443) respectively, and difference was statistically significant (χ2=2145.672, P=0.000). Univariate log-rank test and multivariate Cox regression analysis showed prognostic factors of different histological types were not consistent, including age at diagnosis, tumor size and location, pathological grade, surgical resection and radiotherapy. Accept pilocytic astrocytoma, surgical resection of patients with medulloblastoma, ependymoma and glioblastoma had positive significance for survival and prognosis; radiotherapy can reduce risk of death in medulloblastoma, but it might increase risk of death in pilocytic astrocytoma. Conclusions Pilocytic astrocytoma is most common pediatric glioma with best prognosis, and glioblastoma has a relatively low incidence with poor prognosis. Demographic information, tumor characteristics and treatment options are all predictive factors for survival and prognosis of children with glioma. Surgical resection can reduce risk of death in children with gliomas of various histological types except pilocytic astrocytoma. Radiotherapy can reduce risk in medulloblastoma, however, it may increase risk of death in pilocytic astrocytoma.

Key words: Glioma, Child, Database, Neurosurgical procedures, Radiotherapy, Prognosis, Survival analysis, Regression analysis