Development and optimisation strategies for a nomogram-based predictive model of malignancy risk in thyroid nodules.

Area of Science:

• Radiology
• Medical Imaging
• Oncology

Background:

• Thyroid nodules require accurate classification for optimal patient management.
• The Chinese Thyroid Imaging Reporting and Data System (C-TIRADS) provides a standardized framework for thyroid nodule assessment.
• Improving the diagnostic accuracy and clinical utility of C-TIRADS is crucial for effective thyroid cancer screening and diagnosis.

Purpose of the Study:

• To develop and validate a clinical prediction model for optimizing C-TIRADS classification.
• To enhance the diagnostic efficiency and clinical utility of the TIRADS classification system.

Main Methods:

• A binary logistic regression model was constructed using data from 1659 patients across two hospitals.
• The study employed a derivation cohort (909 patients) for model development and internal validation, and an external validation cohort (750 patients).
• Model performance was assessed using receiver operating characteristic (ROC) curves, nomograms, and calibration curves.

Main Results:

• Significant predictors for C-TIRADS optimization included original C-TIRADS category, abnormal cervical lymph node sonographic findings, and thyroid nodule size changes.
• The optimized nomogram achieved an area under the ROC curve (AUC) of 0.730 in the derivation set and 0.865 in the external validation set.
• The model demonstrated good calibration and favorable net clinical benefit, with specific probability thresholds for upgrading or downgrading C-TIRADS categories.

Conclusions:

• An optimized C-TIRADS model integrating imaging features and clinical risk factors can significantly aid radiologists.
• This enhanced model improves the diagnostic efficiency and clinical utility of the TIRADS classification for thyroid nodules.
• The validated model offers a valuable tool for more precise thyroid nodule assessment and management.