A Clinical-radiomic Nomogram Based On Computed Tomography For Predicting The Risk Of Lymph Node Metastasis In CN0 Papillary Thyroid Carcinoma

Objective:To establish a clinical-radiomic nomogram Based on Computed Tomography(CT)based on preoperative CT images of c N0 papillary thyroid carcinoma(PTC),improving the accuracy of preoperative prediction of central lymph node metastases(CLNM)in patients with c N0 papillary thyroid carcinoma(PTC)and providing valuable information for clinical surgical decision making.Material and methods:Preoperative CT images and clinical data were collected from the first Affiliated Hospital of Gannan Medical College(center 1,n=120 cases)and the the Second Affiliated Hospital of Nanchang University(Center 2,n=58 cases)who were confirmed by surgery and pathology from November 2017 to August 2021.Patients in center 1were randomly divided 7:3 into training group(n=83),internal validation group(n=37),and patients in center 2 as external validation group(n=58).The imaging features were extracted from preoperative plain CT,arterial CT and venous CT.Single factor analysis and minimum absolute contraction selection operator(LASSO)were used to reduce and filter the omics features.The minimum absolute contraction selection operator(LASSO)was used to select CT semantic features and clinical factors.Logistic regression analysis was used to construct plain scan,arterial phase,venous phase radiomic models,CT semantic feature-clinical(clinical model for short),and clinical-radiomic combined classification models,and model validation was performed in internal validation group and external validation group,respectively.The area under curve(AUC),sensitivity,specificity,accuracy and Calibration curves were used to evaluated the model performance,and decision Curve analysis(DCA)evaluated the clinical effectiveness of the model.The optimal model was extracted in the form of clinical-radiomic nomogram.Result:The arterial clinical-radiomic constructed by Logistic regression analysis performed well in the training group in predicting lymph node metastasis in the cervical central group of c N0 papillary thyroid carcinoma,and its AUC value,sensitivity,specificity and accuracy were as follows: 0.870(0.795-0.944)、0.784、0.804、0.795.The analysis of the calibration curve of the arterial clinical-radiomic revealed that the predicted probability was in good agreement with the actual probability(pathological detection).Decision curve analysis showed that the nomogram had certain clinical application value.Conclusions:The clinical radiomic nomogram constructed by the arterial CT radiomic features can accurately predict the central lymph node metastasis in c N0 thyroid papillary carcinoma,which is helpful for clinical treatment decision.