Long-term survival patients with advanced non-small cell lung cancer receiving thoracic radiotherapy: clinical characteristics and the construction of a nomogram prognostic model
10.3760/cma.j.cn112271-20221208-00478
- VernacularTitle:晚期非小细胞肺癌胸部放疗长期生存患者的临床特征及Nomogram预测模型构建
- Author:
Wei JIANG
1
;
Zhu MA
;
Qingsong LI
;
Yichao GENG
;
Daxian LUO
;
Wengang YANG
;
Xiaxia CHEN
;
Weiwei OUYANG
;
Yinxiang HU
;
Shengfa SU
;
Bing LU
Author Information
1. 贵州医科大学附属医院 贵州医科大学附属肿瘤医院胸部肿瘤科 贵州医科大学临床医学院肿瘤学教研室，贵阳 550000
- Keywords:
Advanced non-small cell lung cancer;
Radiotherapy;
Nomogram model
- From:
Chinese Journal of Radiological Medicine and Protection
2023;43(3):189-197
- CountryChina
- Language:Chinese
-
Abstract:
Objective:To analyze the clinical characteristics of long-term survival patients with advanced non-small cell lung cancer (NSCLC) treated with chemotherapy combined with primary tumor radiotherapy, and to establish a Nomogram prognostic model, aiming to provide a certain reference for making a decision about the treatment of advanced NSCLC.Methods:A retrospective analysis was made on the data of 260 NSCLC patients who participated in two prospective clinical studies from January 2003 to May 2012 and the data of 138 NSCLC patients admitted to the Affiliated Cancer Hospital of Guizhou Medical University from January 2014 to August 2020. The former 260 cases were used as a training set and the latter 138 cases were used as the validation set. The overall survival (OS) of ≥ 18 months was defined as long-term survival (LTS). The clinical characteristics of LTS patients were compared with those with OS less than 18 months. The clinical characteristics and treatment-related parameters between the two types of patients were compared using the χ2 test. A multivariate analysis was made using logistic regression, and a nomogram model was built using RStudio. Results:The median OS of the training set was 13.4 months (95% CI: 11.9-14.9), with 1-, 2-, and 3-year OS rates of 55.4%, 19.1%, and 11.9%, respectively. In the training set, 87 cases had LTS and were classified as the LTS group, while 173 cases had OS less than 18 months and were classified as the non-LTS group. The univariate analysis showed that the prognostic factors affecting LST included the KPS score, T status, the number of metastatic organs, the number of metastatic lesions, brain metastasis, bone metastasis, the number of chemotherapy cycles, the biologically effective dose (BED) to the primary tumor, hemoglobin level, platelet count, plasma D-dimer, fibrinogen level, lactate dehydrogenase, and lung immune prognostic index (LIPI; χ2=4.72-12.63, P < 0.05). The multivariable analysis showed that the independent prognostic factors of LTS included a number of chemotherapy cycles ≥ 4, BED ≥ 70 Gy, platelets ≤ 220×10 9/L, D-dimer ≤ 0.5 mg/L, and a good LIPI score ( P= 0.002, 0.036, 0.005, 0.008, and 0.002). A nomogram model was established using the meaningful parameters obtained in the multivariable analysis, determining that the training and validation sets had a consistency index (C-index) of 0.750 and 0.727, respectively. As shown by the analytical result of the corrected curves, for the advanced NSCLC patients treated with thoracic radiotherapy, their LTS probability predicted using the nomogram prognostic model was highly consistent with their actual LTS probability. Both the analytical result of the receiver operating characteristic (ROC) curves and the decision curve analysis (DCA) result showed that the composite prediction model was more beneficial than a single prediction model. Conclusions:For patients with advanced NSCLC treated with thoracic radiotherapy, the independent prognostic factors of LTS included the number of chemotherapy cycles, BED, platelet count, pre-chemotherapy D-dimer, and LIPI score. The Nomogram prognostic model built based on these prognostic factors is a convenient, intuitive, and personalized prediction model used to screen patients who can benefit from thoracic radiotherapy.
