Objective:

To develop an online nomogram model for precise preoperative assessment of malignancy risk in indeterminate thyroid nodules (ITNs) following fine-needle aspiration.

Approach:

• Patient Recruitment: Patients with thyroid nodules were recruited from five centers and divided into training, testing, and validation cohorts.
• Feature Extraction: Radiomics and deep learning features were extracted from B-mode ultrasound (BMUS) and strain elastography ultrasound (SEUS) images.
• Model Development: Malignancy-associated features were selected to construct signature scores, followed by multivariate regression analysis to develop a comprehensive diagnostic model.
• Model Evaluation: The model was evaluated for discrimination, calibration, and clinical usefulness.

Key Findings:

• Multimodal imaging features, genetic testing, and elastography levels were identified as key biological markers for diagnosing ITNs.
• The nomogram model demonstrated strong performance with AUC values of 0.907 (95% CI: 0.877-0.931), 0.885 (95% CI: 0.821-0.932), and 0.860 (95% CI: 0.762-0.929) in training, external test, and prospective validation sets, respectively.
• The nomogram outperformed clinical and individual scoring models in terms of performance and calibration.

Interpretation:

The proposed nomogram accurately diagnoses ITNs.

Limitations:

• The study is limited to specific patient cohorts from five centers.
• Further validation in diverse populations may be necessary to confirm generalizability.

Conclusion:

The online nomogram may improve preoperative evaluation of ITNs.