Related Experiment Video
Updated: May 22, 2026
A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images
Published on: April 21, 2023
Predicting lateral cervical lymph node involvement in papillary thyroid carcinoma patients: development of a nomogram
Jinfang Fan1, Weiwei Li1, Lingling Tao1
1Department of Ultrasound, Ruijin Hospital Luwan Branch, School of Medicine, Shanghai Jiao Tong University, Shanghai, China.
Background:
Preoperative prediction of lateral cervical lymph node metastases has a major impact on prognosis and recurrence for patients with papillary thyroid carcinoma (PTC). This study aimed to explore the value of the qualitative and quantitative contrast-enhanced ultrasound (CEUS) characteristics of the interior and periphery of lateral cervical lymph nodes (LNs) in PTC patients, construct a nomogram prediction model, and verify its clinical practicality.
Methods:
The qualitative and quantitative CEUS characteristics of lateral cervical LNs in 200 PTC cases were retrospectively analyzed. The data were chronologically divided into a training set and a validation set, and the risk factors for lymph node metastasis (LNM) were obtained. A nomogram prediction model was constructed, and a receiver operating characteristic curve was drawn to evaluate the diagnostic efficacy of the model. Calibration curves were drawn to evaluate the accuracy of the model. Clinical decision curves were drawn to analyze and calculate the clinical practicality of the model.
Results:
The univariate analysis revealed that statistically significant differences were observed in LN size after CEUS, perfusion mode, perfusion uniformity, perfusion defect, LN internal peak, internal sharpness, peripheral peak, peripheral time to peak (TP), and peripheral sharpness between benign and malignant LNs. The multivariate analysis revealed that LN size, perfusion mode, perfusion defect, peripheral TP, and peripheral sharpness were independent risk factors for cervical LNs. These risk factors were used to construct the corresponding nomogram with area under the curve (AUC) of 0.907 (95% confidence interval: 0.85-0.96, P<0.001), accuracy of 85%, sensitivity of 88.2%, and specificity of 80.0%. The concordance index of the training and validation sets was 0.907 and 0.847, respectively. The calibration curves in both the training and validation sets were close to the standard curve.
Conclusions:
The nomogram was constructed based on the qualitative and quantitative CEUS features of the interior and periphery of lateral cervical LNs, and it showed good predictive ability and clinical practicality for lateral cervical LNs in PTC, and it might provide useful support for clinicians.
