Multiparametric Analysis Based On Diffusion-weighted Imaging For Assessing Cervical Cancer

Part 1 Whole-tumor Texture Model based on Diffusion Kurtosis Imaging for Assessing Cervical Cancer:A preliminary studyPurpose:To evaluate the diagnostic potential of diffusion kurtosis imaging（DKI）functional maps with whole-tumor texture analysis in differentiating cervical cancer（CC）subtype and grade.Materials and Methods:Seventy-six patients with CC were enrolled.First-order texture features of the whole tumor were extracted from DKI and DWI functional maps,including apparent kurtosis coefficient averaged over all directions（MK）,kurtosis along the axial direction（Ka）,kurtosis along the radial direction（Kr）,mean diffusivity（MD）,fractional anisotropy（FA）,and ADC maps,respectively.The Mann-Whitney U test and ROC curve were used to select the most representative texture features.Models based on each individual and combined functional maps were established using multivariate logistic regression analysis.Conventional parameters-the average values of ADC and DKI parameters derived from the conventional ROI method were also evaluated.Results:The combined model based on Ka,Kr,MD,and FA maps yielded the best diagnostic performance in discrimination cervical squamous cell cancer（SCC）and cervical adenocarcinoma（CAC）with the highest AUC（0.932）.Among individual functional map derived models.Kr map derived model showed the best performance when differentiating tumor subtypes（AUC=0.828）.MK₉0th percentile was useful for distinguishing high-grade and low-grade in SCC tumors with an AUC of 0.701.The average values of MD,FA,and ADC were significantly different between SCC and CAC,but no conventional parameters were useful for tumor grading.Conclusions:The whole-tumor texture analysis applied to DKI functional maps allows for accurate differential diagnosis of cervical cancer subtypes and is helpful in grading SCC.Part 2 Predicting Tumor Recurrence in Cervical Cancer Treated with Chemoradiotherapy-Additional Value of Intravoxel Incoherent Motion Diffusion-weighted ImagingPurpose:To determine whether intravoxel incoherent motion（IVIM）DWI derived parameters may improve the prediction of tumor recurrence for advanced cervical cancer patients following concurrent chemoradiotherapy（CCRT）,especially after integrating into clinical variables.Materials and Methods:Between March 2014 and November 2019,86 consecutive advanced cervical cancer patients were prospectively enrolled and underwent pelvic MRI examinations with IVIM sequence before treatment.The primary outcome was disease-free survival（DFS）.IVIM-derived parameters,conventional imaging features（tumor maximum size,lymph node metastasis[LNM],primary tumor infiltration）,and clinical prognosis variables（age,clinical International Federation of Gynecological[FIGO]）2018 staging system,tumor subtype,serum levels of squamous cell carcinoma antigen[SCC-Ag]and hemoglobin)were collected.Univariate and multivariate Cox hazard regression analyses were performed to evaluate these parameters in the prediction of DFS.The independent and prognostic interested variables were combined to build a prediction model as compared with the clinical FIGO 2018 staging system.Survive curves were generated using the Kaplan-Meier method,and compared with the log-rank test.Results:The median follow-up was 700.5 days for all patients.f,LNM and SCC-Ag were independently associated with tumor recurrence（HR:<0.001,4.553,1.036 P=0.018,<0.001,0.046）.The prediction model based on f,tumor maximum size,LNM and SCC-Ag demonstrated a moderate predictive capability in identifying cervical cancer patients with a high risk of tumor recurrence;the model was more accurate than FIGO staging system alone（c-index:0.753 vs.0.618）and the combination of the FIGO staging system,SCC-Ag and f（c-index:0.753 vs.0.712）.Moreover,patients were grouped into low-,medial-and high-risk levels based on tumor maximum size,positive LNM,SCC-Ag>6ng/ml and f<0.115,with which the 2-year DFS was significantly stratified（P<0.001）.Conclusion:IVIM derived perfusion parameter-f value could be used as a useful and accurate biomarker to predict DFS in CCRT treated advanced cervical cancer patients as an assistant of clinical prognosis variables.Part 3 Feasibility of Intravoxel Incoherent Motion Diffusion-weighted Imaging in Distinguishing Adenocarcinoma Originated from Uterine Corpus or CervixPurpose:To prospectively assess the incremental value of intravoxel incoherent modtion（IVIM）DWI in determining whether the adenocarcinoma originated from the uterine corpus or cervix.Materials and Methods:Eighty consecutive uterine adenocarcinomas from the cervix or endometrium confirmed by histopathology underwent multi-b value DWI acquisition on a 3.0T MR scanner before treatment.Five morphologic features were analyzed using Fisher exact test;Based on mono-exponential and IVIM model,multi-b value DWI derived parameters,including standard apparent diffusion coefficient（ADCstandard）,true coefficient diffusivity（D）,perfusion-related diffusivity（D*）,and perfusion fraction（f）were compared using two-sample independent t-test or Mann-Whitney U test.Logistic regression analysis was used to develop different diagnosis model.The ROCs of these variables and diagnostic models were compared to evaluate the diagnostic efficiency.Results:Among single morphologic features,tumor location yielded the highest AUC of 0.891 in distinguishing endometrial adenocarcinoma from cervical adenocarcinoma.Among single multi-b value DWI derived parameters,f values showed the best diagnostic performance（AUC:0.837）at the optimal cut-off value of 0.261.Additionally,the combined diagnostic model,which consisted of tumor location,ADCstandard and f showed the largest AUC of 0.967 with the highest sensitivity of 88.14%,highest specificity of 100.00%,and highest accuracy of 91.25%.Conclusion:Multi-b value DWI derived parameters add additional diagnostic value to conventional morphologic features.A combined diagnosis model is a promising imaging tool for predicting the origin of uterine adenocarcinoma,further contributing to therapeutic decision-making.