A network-based prognostic prediction model for gastric signet ring cell carcinoma after laparoscopic surgery
10.3760/cma.j.cn113855-20240815-00537
- VernacularTitle:基于网络的胃印戒细胞癌腹腔镜术后预后预测模型的构建
- Author:
Yujuan JIANG
1
;
Xinxin SHAO
1
;
Haitao HU
1
;
Yiming LU
1
;
Haikuo WANG
1
;
Wangyao LI
1
;
Yantao TIAN
1
Author Information
1. 国家癌症中心 国家肿瘤临床医学研究中心 中国医学科学院北京协和医学院肿瘤医院胰胃外科，北京 100021
- Publication Type:Journal Article
- Keywords:
Stomach neoplasms;
Carcinoma, signet ring cell;
Disease-free survival;
Prognosis;
Forecasting
- From:
Chinese Journal of General Surgery
2025;40(10):806-810
- CountryChina
- Language:Chinese
-
Abstract:
Objective:The purpose of this study was to develop a dynamic prediction model for patients with gastric signet ring cell cancer (GSRCC)following laparoscopic radical gastrectomy in order to improve the precision and usefulness of prognoses prediction for overall survival and disease-free survival.Methods:From 2011 to 2018, 914 National Cancer Center patients participated in the study. To find independent prognostic indicators and create a prognostic nomogram model, univariate and multivariate regression analyses were performed. Calibration curves, receiver operating characteristic curves, and consistency indices were used to assess the model's performance. To make clinical application more convenient, two web-based prediction tools were created.Results:A training set of 639 cases and a validation set of 275 instances were randomly selected from among the patients. Important predictive variables such as age, tumor size, location, pN and pT staging, and postoperative chemotherapy were all incorporated in the model (all P<0.05). The model's consistency index and area under the receiver operating characteristic curves were both higher than 0.7, and the calibration curves demonstrated a good fit between the expected and actual values, indicating high accuracy and consistency in postoperative survival prediction for patients with gastric signet ring cell carcinoma. Conclusion:We successfully developed two dynamic prediction models in this study, which improved its clinical practicability using web-based tools and is anticipated to be crucial to clinical practice going forward.
