To Study The Correlation Between Lymph Node Metastasis And Prognosis Of Early Cervical Squamous Cell Carcinoma By ¹⁸F-FDG PET Radiomic Signature

Part Ⅰ ¹⁸F-FDG PET radiomic signature as a predictive factor for lymph node metastasis in early-stage cervical squamous cancerPurpose Lymph node metastasis is the most crucial independent risk factor for recurrence and death in patients with early-stage cervical cancer.The goal of the present study was to develop a ¹⁸F-FDG PET radiomic signature of LN involvement based on¹⁸F-FDG PET/CT imaging.Materials and Methods:We enrolled 389 pathologically confirmed early-stage cervical squamous cancer.All underwent ¹⁸F-FDG PET/CT scanning before surgery treatment.The CT criteria for the diagnosis of pelvic lymph nodes were short diameter≥10mm,and the PET criteria were SUVmax value≥2.5.Patients were randomly divided into 2cohorts.251 patients were allocated to a trainning cohort,117 patients were allocated to an independent validation cohort.We used ¹⁸F-FDG PET DICOM original images archived in the PACS.The segmentation of a region of interest（ROI）is essential for the extraction of quantitative radiomic features.A radiologist with 10 years of experience used the A.K.software for 3D manual segmentation.A uclear medicine doctor with10-years of experience validated all segmentations.A total of 396 radiomic features and relative clinical characteristics were extracted from each patient.The features were divided into 3 groups:（I）Histogram Features,（II）High Order Texture Features,（III）Form Features characteristics.MRMR and LASSO were applied to select features and construct a radiomic signature.The performance of the radiomic signature was assessed in training and validation cohorts.Statistical methods were based on the R analysis platform（version 3.5.1）.In univariate analysis,the T test or Mann-Whistney U test and the chi-square test were used for continuous and categorical variables,respectively,to test the performance of clinical characteristics and potential prognostic outcomes.P values<0.05 were considered to indicate statistical significance.The area under the receiver operating characteristics curve（ROC AUC）,classification accuracy,sensitivity,and specificity were used to assess the quantitative discrimination performance.The point on the ROC curve with the maximum positive likelihood ratio was considered the optimal cutoff threshold value.These prediction measures were computed using the R package p ROC,and MRMR feature selection was done using the glm package.Results:1.In the training cohort and the verification cohort,the relevant clinicopathological characteristics（age,SCC level,FIGO stage,differentiation degree and lymph node metastasis,etc.）of all patients were analyzed by univariate analysis,and there was no significant difference between the lymph node metastasis group and the non-lymph node metastasis group（P>0.05）.2.Eight metabolic imaging omics tags were finally extracted and a model was established,which showed a good distinguishing effect for whether lymph node metastasis was present or not.The accuracy,sensitivity and specificity were 0.803,0.754 and 0.852,respectively,in the training group,and 0.8150.926 and 0.704,respectively,in the verification group.3.The diagnostic efficacy of PET/CT in pelvic lymph node metastasis of cervical cancer was moderate.4.Further study was found that the AUC（95%CI）of the diagnostic model in the training group and the test group were 0.831（0.757-0.906）and 0.830（0.714-0.946）,and the AUC of the diagnostic model was the highest and there was a significant statistical difference（P<0.05）when compared with the diagnostic criteria of PET,CT and PET/CT.Conclusions:1.¹⁸F-FDG PET radiomic signature can be used as a noninvasive biomarker for preoperative assessment of lymph node status and potentially influence the therapeutic decision making in early-stage cervical squamous cancer patients.2.FDG metabolic imaging has better sensitivity than traditional PET/CT diagnosis,and the two can complement each other in the diagnosis of lymph node metastasis in early cervical cancer.Part Ⅱ ¹⁸F-FDG PET radiomic signature as a predictive factor for prognosis in early-stage cervical squamous cancer Objective:To explore the feasibility of establishing a prognostic nomogram predictionmodel based on preoperative PET radiomics,so as to provide new ideas for precise treatment of early-stage cervical squamous cell cancer.Materials and Methods:Three hundred and sixty patients with newly diagnosed early-stage cervical squamous cell cancer who treated with radical resection were retrospectively reviewed.All patients underwent pretreatment whole-body ¹⁸F-FDG PET/CT.All the patients were randomly divided into the training（n=251）and independent validation（n=109）cohorts.FDG PET images data of the cervical cancer were collected for extraction of radiomic features by A.K.software.Using the least absolute shrinkage and selection operator algorithm,the training set was processed for reducing data dimension,feature selection,and construction of a radiomics signature.Then,a prediction model including radiomics signature,pathological features and SCC level as presented in a nomogram was analyzed.Multivariable regression analysis was then used to construct a nomogram.The performance of the nomogram was then estimated by discrimination,calibration mid clinical usefulness.The internal cohort was validated.The consistency between the model prediction and the actual situation was evaluated by using the correction curve.Results:A total of 3 features were chosen from 396 radiomics features,and radiomic signature was established.Combining the radiomics signature,lymph node metastasis,neural tube invasion,vascular infiltration and SCC level,we developed the individualized radiomics nomogram.The model showed good discrimination,with a C-index of 0.787 and good calibration.The nomogram also showed good discrimination,with a C-index of 0.716 and good calibration in the validation cohort.Conclusion:We first used the radiomics to predict the occurrence of FDG PET imaging and the prognosis after radical hysterectomy.As a new preoperative prediction method,the nomogram model of the imaging group showed good prediction accuracy of DFS on the individualization,non-invasive assessment,which can help us to identify more benefit in surgical treatment of cervical cancers in the future.