Application of Ultrasound Artificial Intelligence and Elastography in Differential Diagnosis of Thyroid Nodules

Overview

The application of computer-aided diagnosis (CAD) technology "S-DetectTM" enables qualitative and quantitative automated analysis of ultrasound images to obtain objective, repeatable and more accurate diagnostic results. The Elastic Contrast Factor (ECI) technique, unlike conventional strain-elastic imaging techniques, can evaluate the elastic distribution in the region of interest. The purpose of the study was to evaluate the differential diagnosis value of ultrasound S-DetectTM technology for benign and malignant thyroid nodules and evaluate the consistency of ultrasound S-DetectTM technology and sonographer in the differential diagnosis of thyroid nodules and explore the differential diagnosis value of Samsung ultrasound ECI technology for benign and malignant thyroid nodules.

Thyroid nodules are one of the most common nodular lesions in adults, and their incidence is increasing. Clinical studies have found that about 10%-15% of adults have thyroid nodules, most of which are benign nodules, only 7% of thyroid nodules tend to be malignant, but if not treated promptly, 5% of patients will still from benign to malignant. The incidence of thyroid cancer increases with age. According to the National Cancer Research Center, about 64,300 new cases of thyroid cancer occurred in the world in 2016, and about 1,980 died of thyroid cancer. Timely and accurate detection of thyroid nodules and differentiation of benign and malignant are important for improving clinical treatment and improving patient prognosis. S-Detect technology is a computer-aided (CAD) system recently developed by Samsung Medical Center for thyroid ultrasound to assist in morphological analysis based on the Thyroid Imaging Reporting and Data System (TI-RADS) description and final assessment.This provides a new way to identify the benign and malignant thyroid nodules. The ECI technique, unlike conventional strain-elastic imaging technology, performs an elastic analysis of the entire two-dimensional image. Moreover, when measuring the elastic ratio, it is only necessary to place a region of interest (ROI) at the nodule. Compared with the average elasticity of the surrounding area, it is more reflective of the elastic ratio of the mass to the surrounding tissue.