Related Experiment Video
Updated: May 7, 2026
03:55
Computer-Aided Three-Dimensional Visualization in the Treatment of Locally Advanced Thyroid Cancer
Published on: June 9, 2023
1.2K
Malignancy risk assessment in patients with thyroid nodules using classification and regression trees
Shokouh Taghipour Zahir1, Fariba Binesh, Mehrdad Mirouliaei
1Department of Clinical Pathology, Sadoughi Hospital, Shahid Sadoughi University of Medical Sciences, Yazd, Iran.
Journal of Thyroid Research
|October 9, 2013
Summary
Classification and regression trees (CART) reliably identify patients with low risk of thyroid malignancy, potentially avoiding unnecessary surgery. This AI approach enhances diagnostic accuracy for thyroid nodules.
Area of Science:
- Oncology
- Medical Informatics
Background:
- Thyroid nodules are common, with a small percentage being malignant.
- Accurate differentiation between benign and malignant nodules is crucial for appropriate patient management.
- Distinguishing malignant from benign thyroid nodules often requires invasive procedures.
Purpose of the Study:
- To evaluate the effectiveness of classification and regression trees (CART) in distinguishing benign from malignant thyroid nodules.
- To assess the utility of CART in identifying patients at low risk for thyroid malignancy who may avoid surgery.
Main Methods:
- A two-step CART classification approach was applied to clinical, demographic, and ultrasonographic data of 271 patients.
- The first step prioritized minimizing false negatives for screening.
- The second step refined risk stratification within a high-risk group.
Main Results:
- Age, sex, and nodule size were significant in the initial screening.
- Hypoechogenicity and microcalcifications on ultrasound were key discriminators in the second step.
- The combined CART model achieved 80.0% sensitivity and 94.1% specificity, with a 97.0% positive predictive value.
Conclusions:
- CART analysis provides a reliable method for classifying thyroid nodules.
- This computational approach can effectively identify patients with a low risk of malignancy, enabling avoidance of unnecessary surgical interventions.
