基于APTw的影像组学术前预测宫颈癌淋巴血管间隙侵犯OA北大核心CSTPCD

Objective:To explore the value of amide proton transfer weighted imaging (APTw) radiomics in the preoperative assessment of lymphovascular space invasion (LVSI) in cervical cancer. Materials and Methods:Retrospective analysis of 66 cases of pathologically confirmed cervical cancer and their imaging data. All patients underwent pelvic 3.0 T MRI examination,including axial T2WI,sagittal T2WI,dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI),and 3D-APTw sequence scanning. Region of interest (ROI) within the tumor parenchyma were delineated on the APTw-T2WI fusion images,and APT values were recorded. Tumor lesions were segmented on the reconstructed APTw images,and radiomics features were extracted. Intra-class correlation coefficient (ICC) was employed to select radiomics features with good test-retest reliability both intra-and inter-observer assessments (ICC>0.900). Recursive feature elimination (RFE) and least absolute shrinkage and selection operator (LASSO) algorithms were employed for feature dimensionality reduction and selection. A clinical model,APTw radiomics model and combined model were constructed based on logistic regression classifier. The diagnostic performance and clinical utility of the models were evaluated using receiver operating characteristic (ROC) curves and decision curve analysis (DCA). The predictive performance of different models was compared using the DeLong test. Results:In the training set,the APTw radiomics model demonstrated higher efficacy in predicting cervical cancer LVSI compared to the clinical model (AUC=0.826 vs. 0.675),with statistically significant differences (DeLong test P<0.05). In the training set and the test set,the AUC values of the combined model were 0.838 and 0.825,respectively. DeLong test results showed that the combined model significantly outperformed the clinical model and APTw radiomics model in preoperative assessment of LVSI in the training set (all P<0.05). The decision curve demonstrated that the APTw radiomics model and the combined model exhibit higher clinical utility in both the train and test sets. Conclusions:The radiomics model based on the APTw shows great potential in preoperatively predicting the LVSI status of patients with cervical cancer. Integration with clinical factors further enhances predictive performance,holding prospects to provide crucial support for individualized treatment and prognosis assessment of cervical cancer patients.