AUTHOR=Wu Ge-Ge , Lv Wen-Zhi , Yin Rui , Xu Jian-Wei , Yan Yu-Jing , Chen Rui-Xue , Wang Jia-Yu , Zhang Bo , Cui Xin-Wu , Dietrich Christoph F. 

TITLE=Deep Learning Based on ACR TI-RADS Can Improve the Differential Diagnosis of Thyroid Nodules

JOURNAL=Frontiers in Oncology

VOLUME=11

YEAR=2021

URL=https://www.frontiersin.org/journals/oncology/articles/10.3389/fonc.2021.575166

DOI=10.3389/fonc.2021.575166

ISSN=2234-943X

ABSTRACT=<sec><title>Objective</title><p>The purpose of this study was to improve the differentiation between malignant and benign thyroid nodules using deep learning (DL) in category 4 and 5 based on the Thyroid Imaging Reporting and Data System (TI-RADS, TR) from the American College of Radiology (ACR).</p></sec><sec><title>Design and Methods</title><p>From June 2, 2017 to April 23, 2019, 2082 thyroid ultrasound images from 1396 consecutive patients with confirmed pathology were retrospectively collected, of which 1289 nodules were category 4 (TR4) and 793 nodules were category 5 (TR5). Ninety percent of the B-mode ultrasound images were applied for training and validation, and the residual 10% and an independent external dataset for testing purpose by three different deep learning algorithms.</p></sec><sec><title>Results</title><p>In the independent test set, the DL algorithm of best performance got an AUC of 0.904, 0.845, 0.829 in TR4, TR5, and TR4&amp;5, respectively. The sensitivity and specificity of the optimal model was 0.829, 0.831 on TR4, 0.846, 0.778 on TR5, 0.790, 0.779 on TR4&amp;5, versus the radiologists of 0.686 (<italic>P</italic>=0.108), 0.766 (<italic>P</italic>=0.101), 0.677 (<italic>P</italic>=0.211), 0.750 (<italic>P</italic>=0.128), and 0.680 (<italic>P</italic>=0.023), 0.761 (<italic>P</italic>=0.530), respectively.</p></sec><sec><title>Conclusions</title><p>The study demonstrated that DL could improve the differentiation of malignant from benign thyroid nodules and had significant potential for clinical application on TR4 and TR5.</p></sec>