To Investigate The Correlation Between DCE-MRI Quantitative Parameters And Tumor Microvascular Characteristics And The Value Of Combined Prediction Of Lymph Node Metastasis In Cervical Cancer

Objective The objective of this study is to investigate the correlation between dynamic contrast-enhanced magnetic resonance imaging(DCE-MRI)quantitative parameters(Ktrans,Kep and Ve)and tumor microvascular characteristics of primary cervical cancer and the value of combined prediction of lymph node metastasis(LNM).Materials and methods A retrospective analysis was performed on eighty-five patients with stage IB1-IIA2 cervical cancer confirmed by pathology in our hospital from January,2015 to June,2017.According to the postoperative pathological results of lymph node status,the patients were divided into LNM group(43 cases)and non-LNM group(42 cases).The clinicopathological characteristics of the patients were collected from the medical records,including age,the International Federation of Gynecology and Obstetrics(FIGO)2009 staging,maximum diameter of tumor on palpation,histologic types,degree of differentiation,lymph-vascular space invasion(LVSI),depth of stromal invasion,etc.Eighty-five patients with cervical cancer underwent conventional MRI and DCE-MRI before clinical treatment.Under the supervision of a chief radiologist,a radiology resident with 3 years of experience,blinded to the clinical and pathological information of the patients,used Tissue 4D software(SIEMENS Healthcare,Siemens Healthcare,China).China)post-processed DCE images to obtain quantitative DCE-MRI parameters of primary cervical cancer,including volume transfer constant Ktrans(min-1),interstitial to plasma rate constant Kep(min-1),extravascular to extracellular volume fraction Ve and their image pseudo-color images.CD34 was used to label the microvascular endothelial cells in the pathological sections of cervical cancer tissue,and α-smooth muscle actin(α-SMA)was used to label the pericytes in the microvessels.A senior pathologist with 10 years of work experience performed immunohistochemical double staining on the pathological sections of cervical cancer tissue according to the instructions.Microvessel density(MVD)and microvessel pericyte coverage index(MPI)were used to evaluate micro-angiogenesis and microvessel maturity in cervical cancer,respectively.Chi-square test and rank sum test were used to compare the differences in clinicopathological characteristics,DCE-MRI quantitative parameters,microvascular characteristic parameters(MVD and MPI)between the LNM group and the non-LNM group.Spearman correlation was used to analyze the relationship between DCE-MRI quantitative parameters(Ktrans,Kep,Ve)of primary tumor and MVD,MPI.Binary logistic regression analysis was used to screen out the risk factors of cervical cancer LNM,and a joint prediction model was established according to the analysis results.Receiver operating characteristic(ROC)curve analysis was used to compare the ability of each parameter and the combined model to diagnose LNM of cervical cancer.Results1、There were significant differences in 2009 FIGO stage(p =0.019),LVSI(p =0.003)and depth of stromal invasion(p =0.003)between the LNM group and the non-LNM group;and there was no significant difference in age(p =0.323),maximum diameter of the tumor on palpation(p =0.100),histologic types(p =0.527)and differentiation degree(p =0.840)between the two groups(p > 0.05).Compared with the non-LNM group,the Ktrans value and Kep value of the primary cervical cancer in the LNM group were higher,and the MPI was lower(P < 0.001).There was no significant difference in Ve value and MVD between the LNM group and the non-LNM group(p > 0.05).2、Spearman correlation analysis showed that all quantitative parameters of DCE-MRI were not correlated with MVD(p > 0.05).The quantitative parameters Ktrans and Kep of DCE-MRI were negatively correlated with MPI(r=-0.410,p < 0.001;r=-0.314,p =0.003),and there was no significant relationship between Ve and MPI(p > 0.05).3、Binary logistic regression analysis showed that LVSI,depth of stromal invasion,Ktrans,Kep and MPI were risk factors for LNM in cervical cancer.ROC curve analysis showed that the area under the curve(AUC)of 2009 FIGO stage,LVSI,depth of stromal invasion,Ktrans,Kep and MPI for predicting LNM were 0.614(0.494~0.734),0.660(0.543~0.777),0.645(0.526~0.763),0.779(0.680~0.877),0.731(0.625~0.837)and 0.781(0.681~0.881),respectively.The AUC of the combined model constructed by the above risk factors for predicting LNM was 0.894(0.828~0.960),the best diagnostic cut-off value was 0.497,the sensitivity was 86.0%,and the specificity was 81.0%,which was higher than that of any single parameter.Conclusion1、The maturity of microvessel in cervical cancer may be related to the LNM,and the pathological microvascular staining of cervical cancer tissue before surgery may be helpful to evaluate the status of LNM.2、DCE-MRI quantitative parameters of primary cervical cancer can noninvasively reflect the maturity of tumor microvessel to a certain extent,and pelvic DCE-MRI examination before surgery may help to predict LNM.3、Compared with DCE-MRI quantitative parameters or microvascular characteristics parameters alone,the combined model composed of clinicopathological features,DCEMRI quantitative parameters and microvessel maturity parameters has better diagnostic efficacy in predicting LNM in cervical cancer.