Multi-modal Magnetic Resonance Quantitative Assessment Of Histological Risk Factors And Recurrence Risk Stratification In Early-stage Cervical Cancer

BackgroundTreatment for early-stage cervical cancer includes surgical resection or radiotherapy alone,both of which have similar efficacy.Surgical excision is currently the preferred treatment,but postoperative risk stratification for recurrence based on histological risk factors is required to determine the need for postoperative adjuvant therapy.The multi-mode treatment of surgery combined with postoperative adjuvant therapy superimposed toxic and side effects,and the survival time was not significantly improved compared with radiotherapy alone.Therefore,it is suggested that radiotherapy should be the first choice for patients who may need adjuvant therapy after surgery.However,there is a lack of effective preoperative assessment of histological risk factors and risk stratification of recurrence.Multi-modal magnetic resonance can detect tumor tissue microstructure and microenvironment information noninvasively,but traditional image analysis methods have inherent defects in quantifying tumor spatial heterogeneity.The emerging magnetic resonance"habitat imaging"has realized the visualization of the functional heterogeneity of the key components of the microenvironment in tumor evolution and diagnosis and treatment.Driven by the theory of tumor ecology,landscape ecological analysis technology is expected to become a new method to quantitatively analyze the characteristics of large-scale patches in habitat images and reveal the ecological mechanism of the difference in tumor prognosis.This provides a new way to accurately evaluate the histological risk factors and recurrence risk of early cervical cancer before surgery.PurposePart1:To evaluate the value of dynamic contrast-enhanced magnetic resonance imaging,diffusion-weighted imaging intravoxel incoherent motion model and extracellular volume fraction in preoperative evaluation of Lymphovascular space invasion in early-stage cervical cancer.Part 2:Blood perfusion habitat imaging was constructed based on DCE-MRI to identify spatial subregions with similar perfusion characteristics within tumors,and to explore the value of spatial heterogeneity of perfusion subregions in the evaluation of lymph node metastasis and pathological grading of early-stage cervical squamous cell carcinoma.Part 3:To explore the value of interpretable spatial features of blood perfusion habitat imaging based on DCE-MRI in assessing the recurrence risk stratification of early-stage cervical squamous cell carcinoma.Materials and methodsPart1:A total of 79 patients with early-stage cervical cancer were included in the retrospective analysis based on prospective studies.All patients underwent IVIM,DCE-MRI,noncontrast and contrast-enhanced T1 mapping before surgery.The patients were classified into LVSI group（n=29）and without LVSI group（n=50）according to postoperative pathology.The IVIM-derived parameters（ADC,D,f,D*）,DCE-derived pharmacokinetic parameters(K^trans,K_ep,v_e)and ECV fraction based on T1 mapping between groups with and without LVSI were compared using Student’s t-test or Mann-Whitney U test.The receiver operating characteristic（ROC）curve was used to evaluate the diagnostic accuracy of the above indicators in differentiating LVSI.Part 2:A retrospective analysis based on a prospective study was conducted to collect the data of IVIM,DCE-MRI,noncontrast and contrast-enhanced T1 mapping in88 patients with early-stage cervical squamous cell carcinoma.The quantitative pharmacokinetic parameter map(K^trans,K_ep,v_e)was obtained through post-processing analysis of DCE-MRI.Based on the two-stage cluster analysis at the individual and population levels,each tumor was divided into several spatial subregions with the same hemodynamic phenotype,and the blood perfusion habitat image was constructed.Landscape ecological index was introduced to extract the spatial features of each subregion and the overall landscape of tumor.Student’s t-test or Mann-Whitney U test and ROC curve analysis were used to select the most representative spatial features of each subregion and the overall landscape of habitat image.Multivariate logistic regression analysis was used to further screen spatial features,and establish diagnosis models of lymph node metastasis and pathological grading for early-stage cervical squamous cell carcinoma based on single habitat subregion,overall landscape and multi-scale combined habitat image.In addition,ECV fraction and IVIM-derived parameters were calculated and evaluated using the traditional method of drawing area of interest on slices depicting the maximum cross-section of the tumor.Part 3:On the basis of the perfusion habitat imaging of early cervical squamous cell carcinoma constructed in Part 2,the recurrence risk was stratified into a low recurrence risk group and a positive recurrence risk group according to three high-risk pathological factors and the combination of three intermediate-risk pathological factors（Sedlis criteria）.The differences in the spatial features of the habitat subregions and the overall landscape of the tumor were compared between different recurrence risk groups.Multivariate logistic regression analysis was used to further screen the interpretable spatial features,and establish diagnosis models of recurrence risk of early cervical squamous cell carcinoma based on single habitat subregion,overall landscape and multi-scale combined habitat image.The results were compared with ECV fraction and IVIM-derived parameters obtained from slices depicting the maximum cross-section of the tumor.ResultsPart1:The ECV fractions in the group with LVSI was significantly higher than that of without LVSI group（52.86%vs 36.77%,P<0.001）.In IVIM-derived and quantitative DCE-derived parameters,the K^trans of LVSI group was significantly higher than that of non-LVSI group（0.239 vs 0.176,P=0.034）,and no differences in ADC,D,f,D*,K_ep,v_e values.ROC curve analysis showed that the AUC values of ECV and K^transin differentiating LVSI were 0.874 and 0.655,respectively.De Long test showed that the difference was statistically significant（P=0.0059）.Part2:After two-stage cluster analysis,blood perfusion habitat image of early cervical squamous cell carcinoma was constructed,and three spatial subregions（subregion1,subregion2,subregion3）with different characteristics of blood perfusion were found at the population level.The multi-scale combined diagnosis model based on spatial features from landscape and subregion2 of habitat images had the highest diagnostic efficiency in the identification of lymph node metastasis,with an AUC of0.898.In differentiating the pathological grade of early cervical squamous cell carcinoma,the multi-scale combined diagnosis model based on spatial features from subregion1 and landscape of habitat images had the highest diagnostic efficiency,with an AUC of 0.866.In addition,among the average quantitative parameters of the ECV fraction and IVIM model obtained on the maximum tumor slice,only ADC and D values had significant differences in different pathological grades（P<0.05）.Part 3:The multi-scale combined diagnosis model based on spatial features from subregion3 and landscape of habitat images had the highest diagnostic efficacy in the recurrence risk stratification of early cervical squamous cell carcinoma before surgery.The AUC was 0.813.The average quantitative parameters of ECV fraction and IVIM model had no diagnostic value in the recurrence risk stratification of early cervical squamous cell carcinoma.ConclusionECV fraction is a valuable imaging marker for preoperative identification of lymphovascular space invasion of early-stage cervical cancer.Quantitative spatial features extracted by blood perfusion habitat imaging based on DCE-MRI of early-stage cervical squamous cell carcinoma are helpful for preoperative evaluation of lymph node metastasis and pathological grading.The interpretable quantitative spatial features obtained from the blood perfusion habitat imaging of early-stage cervical squamous cell carcinoma can be used to assess the recurrence risk stratification and provide a basis for clinical treatment decisions.