Deep convolutional neural network for the diagnosis of thyroid nodules on ultrasound

Su Yeon Ko; Ji Hye Lee; Jung Hyun Yoon; Hyesun Na; Eunhye Hong; Kyunghwa Han; Inkyung Jung; Eun Kyung Kim; Hee Jung Moon; Vivian Y. Park; Eunjung Lee; Jin Young Kwak

doi:10.1002/hed.25415

Deep convolutional neural network for the diagnosis of thyroid nodules on ultrasound

Su Yeon Ko, Ji Hye Lee, Jung Hyun Yoon, Hyesun Na, Eunhye Hong, Kyunghwa Han, Inkyung Jung, Eun Kyung Kim, Hee Jung Moon, Vivian Y. Park, Eunjung Lee, Jin Young Kwak

Research output: Contribution to journal › Article › peer-review

57 Citations (Scopus)

Abstract

Background: We designed a deep convolutional neural network (CNN) to diagnose thyroid malignancy on ultrasound (US) and compared the diagnostic performance of CNN with that of experienced radiologists. Methods: Between May 2012 and February 2015, 589 thyroid nodules in 519 patients were diagnosed as benign or malignant by surgical excision. Experienced radiologists retrospectively reviewed the US of the thyroid nodules in a test set. CNNs were trained and tested using retrospective data of 439 and 150 US images, respectively. Diagnostic performances were compared between the two groups. Results: Of the 589 thyroid nodules, 396 were malignant and 193 were benign. The area under the curve (AUC) for diagnosing thyroid malignancy was 0.805-0.860 for radiologists. The AUCs for diagnosing thyroid malignancy for the three CNNs were 0.845, 0.835, and 0.850. There was no significant difference in AUC between radiologists and CNNs. Conclusions: CNNs showed comparable diagnostic performance compared to experienced radiologists in differentiating thyroid malignancy on US.

Original language	English
Pages (from-to)	885-891
Number of pages	7
Journal	Head and Neck
Volume	41
Issue number	4
DOIs	https://doi.org/10.1002/hed.25415
Publication status	Published - 2019 Apr

Bibliographical note

Funding Information:
This study was supported by the Basic Science Research Program through the National Research Foundation of Korea (NRF) by the Ministry of Education (2016R1D1A1B03930375). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Publisher Copyright:
© 2019 Wiley Periodicals, Inc.

All Science Journal Classification (ASJC) codes

Otorhinolaryngology

Access to Document

10.1002/hed.25415

Cite this

@article{8bd8a3af7bfb43c49291f257a384cd27,

title = "Deep convolutional neural network for the diagnosis of thyroid nodules on ultrasound",

abstract = "Background: We designed a deep convolutional neural network (CNN) to diagnose thyroid malignancy on ultrasound (US) and compared the diagnostic performance of CNN with that of experienced radiologists. Methods: Between May 2012 and February 2015, 589 thyroid nodules in 519 patients were diagnosed as benign or malignant by surgical excision. Experienced radiologists retrospectively reviewed the US of the thyroid nodules in a test set. CNNs were trained and tested using retrospective data of 439 and 150 US images, respectively. Diagnostic performances were compared between the two groups. Results: Of the 589 thyroid nodules, 396 were malignant and 193 were benign. The area under the curve (AUC) for diagnosing thyroid malignancy was 0.805-0.860 for radiologists. The AUCs for diagnosing thyroid malignancy for the three CNNs were 0.845, 0.835, and 0.850. There was no significant difference in AUC between radiologists and CNNs. Conclusions: CNNs showed comparable diagnostic performance compared to experienced radiologists in differentiating thyroid malignancy on US.",

author = "Ko, {Su Yeon} and Lee, {Ji Hye} and Yoon, {Jung Hyun} and Hyesun Na and Eunhye Hong and Kyunghwa Han and Inkyung Jung and Kim, {Eun Kyung} and Moon, {Hee Jung} and Park, {Vivian Y.} and Eunjung Lee and Kwak, {Jin Young}",

note = "Funding Information: This study was supported by the Basic Science Research Program through the National Research Foundation of Korea (NRF) by the Ministry of Education (2016R1D1A1B03930375). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Publisher Copyright: {\textcopyright} 2019 Wiley Periodicals, Inc.",

year = "2019",

month = apr,

doi = "10.1002/hed.25415",

language = "English",

volume = "41",

pages = "885--891",

journal = "Head and Neck Surgery",

issn = "1043-3074",

publisher = "John Wiley and Sons Inc.",

number = "4",

}

TY - JOUR

T1 - Deep convolutional neural network for the diagnosis of thyroid nodules on ultrasound

AU - Ko, Su Yeon

AU - Lee, Ji Hye

AU - Yoon, Jung Hyun

AU - Na, Hyesun

AU - Hong, Eunhye

AU - Han, Kyunghwa

AU - Jung, Inkyung

AU - Kim, Eun Kyung

AU - Moon, Hee Jung

AU - Park, Vivian Y.

AU - Lee, Eunjung

AU - Kwak, Jin Young

N1 - Funding Information: This study was supported by the Basic Science Research Program through the National Research Foundation of Korea (NRF) by the Ministry of Education (2016R1D1A1B03930375). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Publisher Copyright: © 2019 Wiley Periodicals, Inc.

PY - 2019/4

Y1 - 2019/4

N2 - Background: We designed a deep convolutional neural network (CNN) to diagnose thyroid malignancy on ultrasound (US) and compared the diagnostic performance of CNN with that of experienced radiologists. Methods: Between May 2012 and February 2015, 589 thyroid nodules in 519 patients were diagnosed as benign or malignant by surgical excision. Experienced radiologists retrospectively reviewed the US of the thyroid nodules in a test set. CNNs were trained and tested using retrospective data of 439 and 150 US images, respectively. Diagnostic performances were compared between the two groups. Results: Of the 589 thyroid nodules, 396 were malignant and 193 were benign. The area under the curve (AUC) for diagnosing thyroid malignancy was 0.805-0.860 for radiologists. The AUCs for diagnosing thyroid malignancy for the three CNNs were 0.845, 0.835, and 0.850. There was no significant difference in AUC between radiologists and CNNs. Conclusions: CNNs showed comparable diagnostic performance compared to experienced radiologists in differentiating thyroid malignancy on US.

AB - Background: We designed a deep convolutional neural network (CNN) to diagnose thyroid malignancy on ultrasound (US) and compared the diagnostic performance of CNN with that of experienced radiologists. Methods: Between May 2012 and February 2015, 589 thyroid nodules in 519 patients were diagnosed as benign or malignant by surgical excision. Experienced radiologists retrospectively reviewed the US of the thyroid nodules in a test set. CNNs were trained and tested using retrospective data of 439 and 150 US images, respectively. Diagnostic performances were compared between the two groups. Results: Of the 589 thyroid nodules, 396 were malignant and 193 were benign. The area under the curve (AUC) for diagnosing thyroid malignancy was 0.805-0.860 for radiologists. The AUCs for diagnosing thyroid malignancy for the three CNNs were 0.845, 0.835, and 0.850. There was no significant difference in AUC between radiologists and CNNs. Conclusions: CNNs showed comparable diagnostic performance compared to experienced radiologists in differentiating thyroid malignancy on US.

UR - http://www.scopus.com/inward/record.url?scp=85061032120&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85061032120&partnerID=8YFLogxK

U2 - 10.1002/hed.25415

DO - 10.1002/hed.25415

M3 - Article

C2 - 30715773

AN - SCOPUS:85061032120

SN - 1043-3074

VL - 41

SP - 885

EP - 891

JO - Head and Neck Surgery

JF - Head and Neck Surgery

IS - 4

ER -

Deep convolutional neural network for the diagnosis of thyroid nodules on ultrasound

Abstract

Bibliographical note

All Science Journal Classification (ASJC) codes

Access to Document

Other files and links

Fingerprint

Cite this