Department of Neurosurgery, First Affiliated Hospital of Nanchang University, Nanchang, China.
Department of Neurosurgery, First Affiliated Hospital of Nanchang University, Nanchang, China; Department of Gastroenterology, First Affiliated Hospital of Nanchang University, Nanchang, China.
World Neurosurg. 2019 Oct;130:e605-e612. doi: 10.1016/j.wneu.2019.06.169. Epub 2019 Jul 15.
The present study aimed to develop and evaluate a nomogram for predicting the overall survival (OS) of patients with low-grade glioma (LGG).
Patients with LGG diagnosed from 1973 to 2013 were identified using the Surveillance, Epidemiology, and End Results (SEER) database. A total of 3732 patients were randomly divided into a training set (n = 2612) and a validation set (n = 1120). Univariate and multivariate Cox regression analyses of the clinical variables were performed to screen for significant prognostic factors. Next, a nomogram that included significant prognostic variables was formulated to predict for LGG. Harrell's concordance index (C-index) and calibration plots were formulated to evaluate the reliability and accuracy of the nomogram using bootstrapping according to the internal (training set) and external (validation set) validity.
A nomogram was developed to predict the 5- and 9-year OS rates using 7 variables in the training set: age, tumor site, sex, marital status, histological type, tumor size, and surgery (P < 0.05). The C-index for internal validation, which the nomogram used to predict OS according to the training set, was 0.777 (range, 0.763-0.791), and the C-index for external validation (validation set) was 0.776 (range, 0.754-0.797). The results of the calibration plots showed that the actual observation and prediction values obtained by the nomogram had good consistency between the 2 sets.
We have developed a ready-to-use nomogram model that includes clinical characteristics to predict OS. The nomogram might provide consultation and risk assessments for subsequent treatment of patients with LGG.
本研究旨在开发并评估用于预测低级别胶质瘤（LGG）患者总生存期（OS）的列线图。
使用监测、流行病学和最终结果（SEER）数据库确定了 1973 年至 2013 年间诊断为 LGG 的患者。共有 3732 名患者被随机分为训练集（n=2612）和验证集（n=1120）。对临床变量进行单因素和多因素 Cox 回归分析，以筛选出显著的预后因素。然后，构建了一个包含显著预后变量的列线图，以预测 LGG。根据内部（训练集）和外部（验证集）验证，通过自举法计算 Harrell 一致性指数（C-index）和校准图，以评估列线图的可靠性和准确性。
在训练集中，使用 7 个变量开发了一个预测 5 年和 9 年 OS 率的列线图：年龄、肿瘤部位、性别、婚姻状况、组织学类型、肿瘤大小和手术（P<0.05）。列线图用于根据训练集预测 OS 的内部验证的 C-index 为 0.777（范围为 0.763-0.791），外部验证（验证集）的 C-index 为 0.776（范围为 0.754-0.797）。校准图的结果表明，2 组之间列线图的实际观察值和预测值具有良好的一致性。
我们已经开发了一个包含临床特征的即用型列线图模型，用于预测 OS。该列线图可以为 LGG 患者的后续治疗提供咨询和风险评估。
