Construction and validation of predictive model for postoperative recurrence in early non-small cell lung cancer patients
10.3760/cma.j.cn115455-20230906-00212
- VernacularTitle:早期非小细胞肺癌患者术后复发预测模型的构建及验证
- Author:
Songbai WANG
1
;
Shirong ZHANG
1
;
Qiang LIU
1
;
Chunna GUO
1
;
Jiaping XU
1
;
Shijia PU
1
;
Huan JIE
1
Author Information
1. 解放军联勤保障部队第九二六医院血液肿瘤科，开远 661606
- Publication Type:Journal Article
- Keywords:
Carcinoma, non-small-cell lung;
Early diagnosis;
Recidivism;
Forecasting;
Models, statistical
- From:
Chinese Journal of Postgraduates of Medicine
2025;48(4):357-360
- CountryChina
- Language:Chinese
-
Abstract:
Objective:To construct and validate a predictive model for postoperative recurrence in early non-small cell lung cancer patients.Methods:The clinical data of 252 patients with early non-small cell lung cancer admitted to the 926th Hospital of Joint Logistic Support Force of PLA from January 2016 to January 2018were retrospectively collected. All of the patients underwent surgical treatment and they were followed up for 5 years after surgery, according the recurrence after surgery, they were divided into the recurrence group (103 cases) and non- recurrence group (149 cases). The risk factors for postoperative recurrence in early non-small cell lung cancer patients were analyzed. A predictive model for postoperative recurrence in early non-small cell lung cancer patients was constructed and validated.Results:The results of Logistic regression analysis showed that tumor long diameter≥ 3 cm, lymph node metastasis, low differentiation, spicules and pleural traction were independent risk factors for postoperative recurrence in early non-small cell lung cancer patients ( P<0.05). Using R4.0.3 statistical software, the dataset was randomly divided into a training set and a validation set, with a sample size of 176 cases in the training set and 76 cases in the validation set. A prediction model was constructed, with thearea under the curve (AUC) of the receiver operating characteristic (ROC) curve of 0.754 (95% CI 0.679 - 0.828) in the training set and AUC of 0.749 (95% CI 0.634 - 0.864) in the validation set. The model was subjected to a Hosmer-Lemeshow Goodness-of-Fit Test in the validation set, χ2 = 11.31, P = 0.185. Conclusions:The predictive model base on tumor long diameter ≥ 3 cm, lymph node metastasis, low differentiation, spicules and pleural traction can identify patients at high risk of postoperative recurrence in early non-small cell lung cancer effectively.
