Deep learning to assist composition classification and thyroid solid nodule diagnosis: a multicenter diagnostic study

Chen, Chen1,2,3; Jiang, Yitao4; Yao, Jincao1,5,6; Lai, Min7; Liu, Yuanzhen1,2,3; Jiang, Xianping8; Ou, Di1,5,6; Feng, Bojian1,2,3; Zhou, Lingyan1,5,6; Xu, Jinfeng9; Wu, Linghu9; Zhou, Yuli9; Yue, Wenwen10; Dong, Fajin9; Xu, Dong1,2,3

2023-10-01

10.1007/s00330-023-10269-z

EUROPEAN RADIOLOGY   Impact Factor & Quartile

0938-7994

1432-1084

Objectives This study aimed to propose a deep learning (DL)-based framework for identifying the composition of thyroid nodules and assessing their malignancy risk.MethodsWe conducted a retrospective multicenter study using ultrasound images from four hospitals. Convolutional neural network (CNN) models were constructed to classify ultrasound images of thyroid nodules into solid and non-solid, as well as benign and malignant. A total of 11,201 images of 6784 nodules were used for training, validation, and testing. The area under the receiver-operating characteristic curve (AUC) was employed as the primary evaluation index.Results The models had AUCs higher than 0.91 in the benign and malignant grading of solid thyroid nodules, with the Inception-ResNet AUC being the highest at 0.94. In the test set, the best algorithm for identifying benign and malignant thyroid nodules had a sensitivity of 0.88, and a specificity of 0.86. In the human vs. DL test set, the best algorithm had a sensitivity of 0.93, and a specificity of 0.86. The Inception-ResNet model performed better than the senior physicians (p < 0.001). The sensitivity and specificity of the optimal model based on the external test set were 0.90 and 0.75, respectively.Conclusions This research demonstrates that CNNs can assist thyroid nodule diagnosis and reduce the rate of unnecessary fine-needle aspiration (FNA).Clinical relevance statement High-resolution ultrasound has led to increased detection of thyroid nodules. This results in unnecessary fine-needle aspiration and anxiety for patients whose nodules are benign. Deep learning can solve these problems to some extent.

Ultrasonography Thyroid nodule Artificial intelligence Deep learning

SCI

English

Corresponding

Radiology, Nuclear Medicine & Medical Imaging

Radiology, Nuclear Medicine & Medical Imaging

WOS:001083922000001

SPRINGER

CLINICAL MEDICINE

Web of Science

1.Chinese Acad Sci, Zhejiang Canc Hosp, Hangzhou Inst Med HIM, Dept Diagnost Ultrasound Imaging & Intervent Thera, Hangzhou 310022, Peoples R China
2.Wenling Big Data & Artificial Intelligence Inst Me, Taizhou 317502, Zhejiang, Peoples R China
3.Zhejiang Canc Hosp, Taizhou Canc Hosp, Taizhou Key Lab Minimally Invas Intervent Therapy, Taizhou Branch, Taizhou Campus, Taizhou 317502, Peoples R China
4.Illuminate LLC, Shenzhen 518000, Guangdong, Peoples R China
5.Translat Res Zhejiang Prov, Key Lab Head & Neck Canc, Hangzhou 310022, Peoples R China
6.Zhejiang Prov Res Ctr Canc Intelligent Diag & Mol, Hangzhou 310022, Peoples R China
7.Zhejiang Chinese Med Univ, Clin Med Coll 2, Hangzhou 310053, Peoples R China
8.Zhejiang Univ, Shengzhou Peoples Hosp, Dept Ultrasound, Affiliated Hosp 1,Shengzhou Branch, , Shengzhou, Shengzhou 312400, Peoples R China
9.Southern Univ Sci & Technol, Jinan Univ, Affiliated Hosp 1, Shenzhen Peoples Hosp,Dept Ultrasound,Clin Med Col, Shenzhen 518020, Peoples R China
10.Tongji Univ, Shanghai Peoples Hosp 10, Ctr Minimally Invas Treatment Tumor, Dept Med Ultrasound,Sch Med, Shanghai 200072, Peoples R China

Chen, Chen,Jiang, Yitao,Yao, Jincao,et al. Deep learning to assist composition classification and thyroid solid nodule diagnosis: a multicenter diagnostic study[J]. EUROPEAN RADIOLOGY,2023.

Chen, Chen.,Jiang, Yitao.,Yao, Jincao.,Lai, Min.,Liu, Yuanzhen.,...&Xu, Dong.(2023).Deep learning to assist composition classification and thyroid solid nodule diagnosis: a multicenter diagnostic study.EUROPEAN RADIOLOGY.

Chen, Chen,et al."Deep learning to assist composition classification and thyroid solid nodule diagnosis: a multicenter diagnostic study".EUROPEAN RADIOLOGY (2023).