JoVE - Journal of Visualized Experiments
JoVE Visualize
Contact Us

Neural Network and Logistic Regression Models Based on Ultrasound Radiomics and Clinical-Pathological Features to Predict Occult Level II Lymph Node Metastasis in Papillary Thyroid Carcinoma

  • 0Department of General Surgery, Sir Run Run Shaw Hospital, School of Medicine, Zhejiang University, Hangzhou, Zhejiang, China (J-W.F., H.L.); Department of thyroid surgery, The Third Affiliated Hospital of Soochow University, Changzhou First People's Hospital, Changzhou, Jiangsu, China (J-W.F., Y-X.Y., J.Y., Y.J.).
Academic Radiology +

|

January 5, 2025

|

January 5, 2025

View abstract on PubMed

Summary

This summary is machine-generated.

This study developed predictive models to identify level II lymph node metastasis in papillary thyroid carcinoma (PTC). The logistic regression-radiomics model accurately predicts metastasis, guiding selective neck dissection and reducing unnecessary surgeries.

Area Of Science

  • Oncology
  • Radiology
  • Surgical Oncology

Background

  • Papillary thyroid carcinoma (PTC) frequently metastasizes to lateral cervical lymph nodes, particularly level II.
  • Accurate identification of level II lymph node metastasis (LNM) is crucial for appropriate surgical planning, such as selective neck dissection (SND).

Purpose Of The Study

  • To develop and validate predictive models for identifying level II LNM in PTC patients.
  • To guide personalized treatment strategies, optimizing surgical interventions and minimizing morbidity.

Main Methods

  • Retrospective analysis of 313 PTC patients who underwent modified radical neck dissection (MRND).
  • Development of five predictive models (neural networks and logistic regression) using ultrasound radiomic features and/or clinical-pathological data.
  • Evaluation of model performance using accuracy, AUC, sensitivity, and specificity; interpretation via SHapley Additive exPlanations and nomogram.

Main Results

  • Level II LNM occurred in 28% of patients.
  • The logistic regression-radiomics signature model achieved the highest performance with 96.8% accuracy and an AUC of 0.989 in the validation set.
  • This model demonstrated superior clinical utility compared to other developed models.

Conclusions

  • Predictive models integrating ultrasound radiomics and clinical data can accurately assess the risk of occult level II LNM in PTC.
  • The logistic regression-radiomics signature model is a valuable tool for personalized treatment, informing MRND for high-risk cases and supporting SND for low-risk cases.

Keywords: