Prediction of lymphovascular space invasion and lymph node metastasis in early-stage cervical cancer based on MRI radiomics combined with clinical features
10.3969/j.issn.1002-1671.2025.06.023
- VernacularTitle:基于MRI影像组学联合临床特征预测早期宫颈癌淋巴脉管间隙浸润及淋巴结转移
- Author:
Liuxia LI
1
;
Baoyi QIN
;
Xinguan YANG
Author Information
1. 桂林市人民医院放射科,广西 桂林 541002
- Publication Type:Journal Article
- Keywords:
cervical cancer;
lymphovascular space invasion;
lymph node metastasis;
magnetic resonance imaging;
radiomics
- From:
Journal of Practical Radiology
2025;41(6):1003-1007
- CountryChina
- Language:Chinese
-
Abstract:
Objective To explore the predictive value of a model established by combining MRI radiomics with clinical features for lymphovascular space invasion(LVSI)and lymph node metastasis(LNM)in early-stage cervical cancer.Methods A retrospective analysis was conducted on relevant data from 123 patients with early-stage cervical cancer confirmed by surgical pathology.They were randomly divided into training set(n=74)and testing set(n=49)in a 6∶4 ratio.The tumor regions in the T2WI-fat suppression(FS),diffusion weighted imaging(DWI),and T1WI with enhancement sequences were segmented and radiomics features were extracted.After screening and dimensionality reduction,the most relevant features were retained.The effectiveness of the model was evaluated using the area under the curve(AUC)of the receiver operating characteristic(ROC)curve.Results The combined model of MRI radiomics-clinical-MRI features had the best predictive performance for early-stage cervical cancer LVSI,with AUC of 0.848 and 0.821 in the training and testing sets,respectively.The predictive performance of various MRI radiomics models for early-stage cervical cancer LNM was superior to the clinical model.Conclusion The MRI radiomics model and clinical model have certain predictive value for early-stage cervical cancer LVSI and LNM,and the combined model performs the best.Radiomics will provide a basis for developing individualized treatment plans in clinical practice.
