Using Multimodal Magnetic Resonance Imaging To Identify Cervical Squamous/Adenocarcinoma And Predict Its Pathological Grading

Background:Uterus cervical cancer(UCC)is a common malignant tumor in the female reproductive system.UCC tissues with different pathological types and pathological grades have relatively specific biological behaviors.Therefore,the choice of treatment methods and survival and prognosis are also different.At present,the histopathological results are still clinically used as the gold standard for determining the classification and classification of UCC.However,due to the heterogeneity of the structure of the tumor itself,whether it is surgical resection of a gross specimen or a small-scale directional stereoscopic puncture,There is a potential risk due to sampling errors.Magnetic resonance imaging(MRI)is the preferred imaging method for cervical cancer exploration and evaluation,but its conventional sequence can only reflect the tumor volume,the length of the radial line and limited signal changes to reflect the lesion information,which is difficult to deepen.Evaluation of microscopic lesion information such as tissue typing and pathological grading.In recent years,with the continuous innovation of MR technology,various functional imaging,such as intravoxel incoherent motion(IVIM),diffusion kurtosis imaging(DKI),and amide proton transfer-weighted imaging(APTWI),etc.have been gradually applied to the evaluation of tumor information in various systems.These functional magnetic resonance imaging techniques not only provide tumor tissue microstructure information,such as cell density,intratumoral necrosis,cystic degree,extracellular water molecule transfer,and transmembrane movement but also detect tumor blood supply status and the vessel wall.Important parameters such as permeability provide a new direction for tumor detection,grading,efficacy monitoring,and prognosis evaluation.Part ?Using Amide Proton Transfer-weighted Imaging to Identify Cervical Squamous/Adenocarcinoma and Predict Its Pathological GradingObjective:To explore the possibility of using Amide proton transfer-weighted imaging(APTWI)for the identification and diagnosis of cervical squamous carcinoma(CSC),cervical adenocarcinoma(CA)and different levels of CSC.Materials and Methods:Seventy-six patients with newly diagnosed UCC were studied prior to treatment,including 20 with poorly differentiated CSC(G3),23 with moderately differentiated CSC(G2),17 with well-differentiated CSC(G1),and 16 with CA(13 with poorly differentiated CA and 3 with moderately differentiated CA).All patients underwent routine MRI(oblique axis-perpendicular to the long axis of the cervix,T1WI/T2WI,sagittal/coronal T2WI)scan,oblique axis DWI sequence(b=1000 s/mm~2)and oblique axis APTWI sequence scan.After the scan,the original image of the APTWI sequence was imported into the workstation for analysis and processing.Two high-grade imaging diagnosticians who are good at diagnosis of female pelvic disease select the region of interest(ROI)without knowing the pathological results.It is required to draw along the edge of the tumor layer by layer,including the tumor parenchyma as much as possible,avoiding areas such as liquefaction necrosis.Then,the asymmetric magnetization transfer ratio at 3.5ppm(MTRasym(3.5ppm))of each group of UCC tissues was measured.Using the relevant statistical software to compare and analyze the difference of MTRasym(3.5ppm)values between CSC group and CA group,CSC high-grade group/low-level group,and the receiver operating characteristic curve(ROC)was used to evaluate each.Parametric diagnostic performance,using the Spearman method to analyze the correlation between parameters and pathological grade.Results:The MTRasym(3.5ppm)in CA was higher than that in CSC(P=0.001).The MTRasym(3.5ppm)in high-level CSC was higher than that in low-level CSC(P=0.001).The MTRasym(3.5ppm)was positively correlated with the grade of CSC differentiation(r=0.498,P=0.001).The MTRasym(3.5ppm)in G3 CSC was higher than that in G2 and G1 CSC(P=0.02/0.01).There was no significant difference in the MTRasym(3.5ppm)between G2 CSC and G1 CSC(P=0.173).The area under the ROC curve(AUC)for the MTRasym(3.5ppm)in distinguishing CSC and CA was 0.779,with a cut-off,sensitivity,and specificity of 2.97%,60.0%,and 82.5%,respectively.The AUC for distinguishing high-/low-level CSC was 0.756,with a cut-off,sensitivity,and specificity of 3.29%,68.8%,and 83.3%,respectively.Conclusion:APTWI may be a useful technique for the identification and diagnosis of CSC,CA and different grades of CSC,which could have an important impact on clinical strategies for treating patients with UCC.Part ? Using Diffusion Kurtosis Imaging to Identify Cervical Squamous/Adenocarcinoma and Predict Its Pathological GradingObjective:To explore the possibility of using diffusion kurtosis imaging(DKI)for the identification and diagnosis of cervical squamous carcinoma(CSC),cervical adenocarcinoma(CA)and different levels of CSC.Materials and Methods:DKI sequence images of 60 patients with UCC who were examined by MRI in our hospital from July 2017 to February 2018 and confirmed by pathological findings were collected.Among them,38 patients in the CSC(17 in the highly-differentiated G1 group,23 in the moderately-differentiated G2 group,and 20 in the poorlydifferentiated G3 group),22 in the CA(5 in the highly-differentiated G1 group,8 in the moderately-differentiated G2 group and 9 in the highly-differentiated G1 group).All patients underwent routine MRI(oblique axis-perpendicular to the long axis of the cervix,T1WI/T2 WI,sagittal/coronal T2WI)scan,oblique axis DKI sequence(b = 0,500,1000,1500,2000 s/mm2).After the scan,the original image of the DKI sequence was imported into the workstation for analysis and processing.Two high-grade imaging diagnosticians who are good at diagnosis of female pelvic disease select the region of interest(ROI)without knowing the pathological results.It is required to draw along the edge of the tumor layer by layer,including the tumor parenchyma as much as possible,avoiding areas such as liquefaction necrosis.Then,the mean diffusion coefficient(MD)and mean kurtosis(MK)of each group of cervical cancer tissues were measured.After the consistency test,the average of the two observers was included in the final data.Using statistical software to compare and analyze the differences in MD and MK between the CSC and CA groups,CSC G1,G2,G3 groups and CA G1,G2,G3 groups.The receiver operating characteristic curve(ROC)was used to evaluate the diagnostic performance of each parameter.The Spearman method was used to analyze the correlation between each parameter and pathological grade.Results:The MK value of CSC group was higher than that of the CA group,and the MD value was lower than that of the CA group(P<0.05).MK values were positively correlated with a pathological grade in CSC and CA groups(r=0.741/0.819,P<0.05);MD values were negatively correlated with pathological grade(r=-0.647/-0.758,P<0.05).MK and MD values were statistically significant across groups of different grades.MK has high diagnostic efficiency in terms of the diagnostic value of various parameters for different pathological types and different grades of CA.The AUC is0.812(CSC/CA),0.904(CSC G1/G2),0.889(CSC G1/G2),0.923(CA G1/G2),0.965(CA G2/G3);MD value also has certain diagnostic performance : The AUC is 0.794(CSC/CA),0.861(CSC G1/G2),0.833(CSC G2/G3),0.897(CA G1/G2),0.918(CA G2/G3).Conclusion:DKI may be a useful technique for the identification and diagnosis of CSC,CA and different levels of CSC and CA,which could have an important impact on clinical strategies for treating patients with UCC.