Your browser doesn't support javascript.
loading
Predictive Value of Ultrasound Radiomics Nomogram of Lymph Node Metastasis in Papillary Thyroid Carcinoma / 中国医学影像学杂志
Article Dans Zh | WPRIM | ID: wpr-1026345
Responsable en Bibliothèque : WPRO
ABSTRACT
Purpose To investigate the value of ultrasound radiomics nomogram in predicting lymph node metastasis(LNM)of papillary thyroid carcinoma(PTC).Materials and Methods A retrospective analysis was conducted on 400 cases of PTC in the First Hospital of Shanxi Medical University from March 2021 to January 2022 confirmed by surgery and pathology,all of which underwent preoperative ultrasound examination,and were randomly divided into training cohort(n=280)and testing cohort(n=120)in a ratio of 7∶3.The relationship between ultrasound clinical features and LNM was evaluated via univariate analysis and a clinical model was established via multivariable Logistic regression.A total of 3 348 features were extracted from preoperative ultrasound images.Pearson correlation coefficient was used to screen the features,and Logistic regression was used to establish the radiomics model.Clinical risk factors and rad scores were combined to construct the nomogram,and the receiver operating characteristic curves and decision curve analysis were applied to evaluate the predictive efficacy and clinical benefit of each model for LNM of PTC.Results Age,primary lesion size,C-TIRADS and ultrasound-reported LNM were the independent risk factors for LNM(t/χ2=2.938,55.923,30.081,34.639,all P<0.05).The area under the curve of ultrasound radiomics nomogram to predict LNM of PTC in the training cohort and the testing cohort was 0.860 and 0.847,respectively;the combined model in 43%-85%had the highest clinical benefit.Conclusion Ultrasound radiomics nomogram has a certain value in predicting LNM of PTC.

Texte intégral: 1 Indice: WPRIM langue: Zh Texte intégral: Chinese Journal of Medical Imaging Année: 2024 Type: Article
Texte intégral: 1 Indice: WPRIM langue: Zh Texte intégral: Chinese Journal of Medical Imaging Année: 2024 Type: Article