Development and validation of a multimodal predictive model based on clinical, biochemical, and quantitative dual-energy CT parameters: for predicting the benignity and malignancy of thyroid nodules - Report - MDSpire

Development and validation of a multimodal predictive model based on clinical, biochemical, and quantitative dual-energy CT parameters: for predicting the benignity and malignancy of thyroid nodules

  • By

  • Yafei Zhang

  • Congyan Yin

  • Ranran Huang

  • Guowei Zhang

  • Xuhong Pan

  • June 24, 2026

  • 0 min

Share

Clinical Report: Predictive Model for Distinguishing Thyroid Nodules

Overview

This study developed a predictive model combining clinical, biochemical, and DECT parameters to differentiate between benign and malignant thyroid nodules.

Background

Thyroid cancer is the most prevalent endocrine malignancy, with a rising incidence, particularly among women. Accurate differentiation between benign and malignant thyroid nodules is crucial, as current methods like fine-needle aspiration biopsy often yield indeterminate results. Advanced imaging techniques, including dual-energy CT, provide valuable quantitative data that can enhance preoperative assessments.

Data Highlights

PredictorOdds Ratio (OR)p-value
Age0.93<0.001
TSH1.650.037
Thyroid Nodule Volume (TV)0.910.017
AUC (Training Cohort)0.866N/A
AUC (Validation Cohort)0.852N/A

Key Findings

  • The predictive model integrates clinical, biochemical, and DECT parameters.
  • Age, TSH, and thyroid nodule volume were identified as independent predictors of malignancy.
  • The model achieved an AUC of 0.866 in the training cohort and 0.852 in the validation cohort.
  • Goodness-of-fit tests indicated an acceptable model fit.
  • Calibration curves showed good agreement between predicted and observed outcomes.

Clinical Implications

The developed predictive model may assist clinicians in accurately stratifying the risk of malignancy in thyroid nodules, potentially reducing unnecessary surgical interventions. This tool offers a non-invasive approach to enhance patient management.

Conclusion

The study successfully created a multimodal predictive model for thyroid nodules.

Related Resources & Content

  1. Frontiers in Oncology, 2026 -- A Multicenter, Clinically Interpretable Prediction Model for Malignancy Risk in C-TIRADS 3–4 Thyroid Nodules
  2. The ASCO Post, 2022 -- AI Model May Aid in Screening, Staging, and Treatment Planning for Thyroid Cancer
  3. Utilizing Artificial Intelligence for the Pre-operative Assessment of Malignant Thyroid Nodules Through Sonographic Characteristics and Cytological Classification
  4. Evaluation and Management of Thyroid Nodules: A Joint Consensus Statement From the British Thyroid Association (BTA), British Association of Endocrine and Thyroid Surgeons (BAETS) and Collaborating Bodies - PubMed
  5. Frontiers in Endocrinology — Development and validation of an intra-tumoral and peri-tumoral radiomics model based on dynamic contrast-enhanced ultrasound for predicting lymph node metastasis in type 2 diabetic patients with thyroid cancer
  6. Evaluation and Management of Thyroid Nodules: A Joint Consensus Statement From the British Thyroid Association (BTA), British Association of Endocrine and Thyroid Surgeons (BAETS) and Collaborating Bodies - PubMed
  7. Distinguishing benign from malignant thyroid nodules via virtual biopsy: a study on using quantitative parameters and classical radiomics features from dual-energy CT imaging - PMC
  8. American Thyroid Association 2026 Guidelines for Thyroid Disease in Preconception, Pregnancy, and Postpartum - Tim I.M. Korevaar, Angela M. Leung, Erik K. Alexander, Sofie Bliddal, Kristien Boelaert, Gabriela Brenta, Roger Chou, Rima Dhillon-Smith, Chrysoula Dosiou, Jennifer L. Eaton, Haixia Guan, Sarah J. Kilpatrick, Bente J. Lasserre, Sun Y. Lee, Spyridoula Maraka, Kara D. Meister, Lilah F. Morris-Wiseman, Caroline T. Nguyen, Elizabeth N. Pearce, Zhongyan Shan, 2026

Original Source(s)

Related Content