Multimodal Magnetic Resonance Correlation Study Of Cervical Cancer Diagnosis

Part I Correlation study of multimodal magnetic resonance in the diagnosis of cervical cancerObjectiveTo explore the value of multimode MRI,including conventional MRI plain scan,intra-body incoherent motion diffusion weighted MRI(IVIM-DWI)and dynamic contrast enhanced MRI(DCE-MRI)in the diagnosis and clinical staging of cervical cancer.MethodsRetrospective analysis was performed on the patients who underwent MRI examination in our hospital from January 2017 to July 2018 and were pathologically confirmed as cervical cancer patients in our hospital,a total of 63 cases.The case group was divided into 1 group of cervical cancer patients and 2 groups of cervical cancer patients with stage IA as the boundary.Cervical cancer group 1 includes carcinoma in situ and stage IA,and cervical cancer group 2 is larger than stage IA.A total of 20 patients who underwent pelvic MRI for other diseases with normal cervix in our hospital during the same period were selected as the control group.All patients received routine pelvic MRI plain scan,IVIM-DWI and DCE-MRI scan.The corresponding statistical methods were selected to analyze and evaluate the diagnostic value of routine magnetic resonance imaging,ivim-dwi and dce-mri for cervical cancer group 1 and cervical cancer group 2,as well as the application value for clinical staging of cervical cancer.Results(1)The D value of cervical cancer group 1 was higher than that of the control group,and the difference was statistically significant.The ADC value,D* value,f value,Ktrans value,Kep value and Ve value were not significantly different from the control group.The area under the D value curve of cervical cancer group 1 was the largest.The values of Ktrans,Kep,Ve,D* and f in cervical cancer group were significantly higher than those in the control group.The ADC and D values were significantly lower than those in the control group(P<0.05).The area under the Ktrans curve of cervical cancer group 2 was the largest,f was positively correlated with Ktrans and Kep.(2)The values of Ktrans,Kep,Ve,D* and f in cervical cancer group 1 were lower than those in stage IIB group,and the values of Ktrans,Kep,Ve,D* and f in stage IIB group were all.The difference was statistically significant compared with the corresponding parameter values of the stage ? IIB group;the ROC curve showed that the area under the curve of f was the largest.(3)The area under the ROC curve of conventional cervical magnetic resonance imaging was 0.532.The area under the ROC curve of cervical cancer group 2 was 0.705.The area under the ROC curve of conventional magnetic resonance plain scan for cervical cancer staging was 0.714.The multi-modal magnetic resonance imaging with conventional magnetic resonance sequence,IVIM-DWI and DCE-MRI combined the area under the ROC curve of cervical cancer group 1 was 0.611,and the area under the ROC curve of cervical cancer group 2 was 0.993,and the stage of cervical cancer was staged.The area under the ROC curve is 0.898.Conclusions(1)The diagnostic efficacy of multimodal magnetic resonance imaging for cervical cancer is higher than that of single anatomical magnetic resonance and functional magnetic resonance.(2)Multimodal magnetic resonance is superior to single anatomical magnetic resonance or functional magnetic resonance in evaluating the staging of cervical cancer.Part? Preliminary discussion on the application of imaging omics in the diagnosis of cervical cancerObjectiveThe application of imaging omics in the diagnosis of cervical cancer was discussed.Image omics features were extracted from magnetic resonance images of cervical cancer to establish a basis for future prediction models.MethodsAll the information obtained by scanning the first part of cervical cancer patients is uploaded to the third-party image group cloud platform(Huiyi Huiying Cloud Platform),including the patient MRI image,clinical information and pathological information.In this experiment,DICOM images of axial T2 weighted imaging(T2WI),diffusion weighted imaging(DWI),incoherent incoherent motion diffusion weighted imaging(IVIM-DWI)and dynamic contrast enhanced magnetic resonance imaging(DCE-MRI)were used.The omics feature extraction.In the above four sequences,the solid part of the lesion was delineated as a region of interest by careful delineation,taking care to avoid liquefaction necrosis,calcification and edema.All images were not pre-processed or normalized,and the completed images were selected to calculate and extract the influence omics features.Using the Vari Variance,Selectbest,lasso,PCA,Covariance,and Clustering techniques to further reduce dimensionality analysis.ResultsThe number of eigenvalues for the conventional axial T2 WI sequence selection is: 6,the number of eigenvalues for the conventional axial DWI sequence is 16,the number of eigenvalues for the IVIM-DWI sequence is: 11,and the number of eigenvalues for the DCE-MRI sequence is : 15,the number of joint selection feature values of the above four sequences is: 47.ConclusionsImage omics can extract features from high-throughput images of magnetic resonance images,thus establishing the basis for the study of predictive models.