Department of Respiratory Diseases, The Second Affiliated Hospital, Jiangxi Medical College, Nanchang University, Nanchang, China.
Department of Pathology, The Second Affiliated Hospital, Jiangxi Medical College, Nanchang University, Nanchang, China.
Medicine (Baltimore). 2024 Sep 13;103(37):e39294. doi: 10.1097/MD.0000000000039294.
Nearly half of lung large cell neuroendocrine carcinoma (LCNEC) patients are diagnosed at an advanced stage and face a high early death risk. Our objective was to develop models for assessing early death risk in stage IV LCNEC patients. We used surveillance, epidemiology, and end results (SEER) databases to gather data on patients with stage IV LCNEC to construct models and conduct internal validation. Additionally, we collected a dataset from the Second Affiliated Hospital of Nanchang University for external validation. We used the Pearson correlation coefficient and variance inflation factor to identify collinearity among variables. Logistic regression analysis and least absolute shrinkage and selection operator analysis were employed to identify important independent prognostic factors. Prediction nomograms and network-based probability calculators were developed. The accuracy of the nomograms was evaluated using receiver operating characteristic curves. The goodness of fit of the nomograms was evaluated using the Hosmer-Lemeshow test and calibration curves. The clinical value of the models was assessed through decision curve analysis. We enrolled 816 patients from the surveillance, epidemiology, and end results database and randomly assigned them to a training group and a validation group at a 7:3 ratio. In the training group, we identified 9 factors closely associated with early death and included them in the prediction nomograms. The overall early death model achieved an area under the curve of 0.850 for the training group and 0.780 for the validation group. Regarding the cancer-specific early death model, the area under the curve was 0.853 for the training group and 0.769 for the validation group. The calibration curve and Hosmer-Lemeshow test both demonstrated a high level of consistency for the constructed nomograms. Additionally, decision curve analysis further confirmed the substantial clinical utility of the nomograms. We developed a reliable nomogram to predict the early mortality risk in stage IV LCNEC patients that can be a helpful tool for health care professionals to identify high-risk patients and create personalized treatment plans.
大约有一半的肺大细胞神经内分泌癌（LCNEC）患者被诊断为晚期，面临着较高的早期死亡风险。我们的目标是为 IV 期 LCNEC 患者建立评估早期死亡风险的模型。我们使用监测、流行病学和最终结果（SEER）数据库收集了 IV 期 LCNEC 患者的数据，用于构建模型和进行内部验证。此外，我们还从南昌大学第二附属医院收集了一组数据集进行外部验证。我们使用 Pearson 相关系数和方差膨胀因子来识别变量之间的共线性。使用逻辑回归分析和最小绝对收缩和选择算子分析来识别重要的独立预后因素。建立预测列线图和基于网络的概率计算器。使用受试者工作特征曲线评估列线图的准确性。使用 Hosmer-Lemeshow 检验和校准曲线评估列线图的拟合优度。通过决策曲线分析评估模型的临床价值。我们从监测、流行病学和最终结果数据库中纳入了 816 名患者，并将他们随机分为训练组和验证组，比例为 7:3。在训练组中，我们确定了 9 个与早期死亡密切相关的因素，并将其纳入预测列线图。总体早期死亡模型在训练组中的曲线下面积为 0.850，在验证组中的曲线下面积为 0.780。对于癌症特异性早期死亡模型，训练组的曲线下面积为 0.853，验证组的曲线下面积为 0.769。校准曲线和 Hosmer-Lemeshow 检验均表明所构建的列线图具有较高的一致性。此外，决策曲线分析进一步证实了列线图的重要临床实用性。我们开发了一个可靠的列线图来预测 IV 期 LCNEC 患者的早期死亡率，这可以成为医疗保健专业人员识别高风险患者和制定个性化治疗计划的有用工具。
