Intratumoral and peritumoral magnetic resonance imaging radiomics combined with clinical characteristics to predict lymphovascular space invasion in cervical cancer
Objective To investigate the value of a nomogram model constructed from intratumoral and peritumoral magnetic resonance imaging radiomics combined with clinical characteristics in predicting the status of lymphovascular space invasion(LVSI)in cervical cancer.Methods A retrospective analysis was conducted on 178 cervical cancer patients confirmed by postoperative pathology,with 70 cases of LVSI(+)and 108 cases of LVSI(-).The patients were divided into a training set[142 cases,including 54 cases of LVSI(+)and 88 cases of LVSI(-)]and a test set[36 cases,including 16 cases of LVSI(+)and 20 cases of LVSI(-)]at a ratio of 8:2.All underwent magnetic resonance imaging before surgery,and regions of interest were manually delineated layer by layer on the T2WI sequence,with the peritumoral region being uniformly expanded outward.Univariate logistic analysis was performed on clinical factors to select independent factors for cervical cancer LVSI(+).Radiomic features were extracted separately from the intratumoral region,the peritumoral region,and the intratumoral-peritumoral region to construct radiomics models,and the differences between the peritumoral and the intratumoral-peritumoral models were compared.A combined model was established based on the radiomics scores of the optimal intratumoral-peritumoral model and clinical independent predictive factors,and a nomogram was plotted.Receiver operating characteristic curves were used to evaluate the diagnostic performance of each model,and decision curve analysis was used to assess the clinical value of the models.Results The combined model demonstrated the best performance among the various models,with AUC of 0.970 in the training set and 0.803 in the test set.Conclusion Intratumoral and peritumoral magnetic resonance imaging radiomics combined with clinical characteristics can effectively predict LVSI in cervical cancer.
cervical cancerradiomicsmagnetic resonance imaginglymphovascular space invasion
NETL
NSTL
万方数据
