Augmenting endometriosis analysis from ultrasound data with deep learning

Adrian Balica; Jennifer Dai; Kayla Piiwaa; Xiao Qi; Ashlee N. Green; Nancy Phillips; Susan Egan; Ilker Hacihaliloglu

doi:https://doi.org/10.1117/12.2653940

Augmenting endometriosis analysis from ultrasound data with deep learning

Adrian Balica, Jennifer Dai, Kayla Piiwaa, Xiao Qi, Ashlee N. Green, Nancy Phillips, Susan Egan, Ilker Hacihaliloglu

Robert Wood Johnson Medical School (RWJMS), Obstetrics, Gynecology and Reproductive Sciences

Rutgers, The State University

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

Endometriosis is a non-malignant disorder that affects 176 million women globally. Diagnostic delays result in severe dysmenorrhea, dyspareunia, chronic pelvic pain, and infertility. Therefore, there is a significant need to diagnose patients at an early stage. Our objective in this work is to investigate the potential of deep learning methods to classify endometriosis from ultrasound data. Retrospective data from 100 subjects were collected at the Rutgers Robert Wood Johnson University Hospital (New Brunswick, NJ, USA). Endometriosis was diagnosed via laparoscopy or laparotomy. We designed and trained five different deep learning methods (Xception, Inception-V4, ResNet50, DenseNet, and EfficientNetB2) for the classification of endometriosis from ultrasound data. Using 5-fold cross-validation study we achieved an average area under the receiver operator curve (AUC) of 0.85 and 0.90 respectively for the two evaluation studies.

Original language	English (US)
Title of host publication	Medical Imaging 2023
Subtitle of host publication	Ultrasonic Imaging and Tomography
Editors	Christian Boehm, Nick Bottenus
Publisher	SPIE
ISBN (Electronic)	9781510660458
DOIs	https://doi.org/10.1117/12.2653940
State	Published - 2023
Event	Medical Imaging 2023: Ultrasonic Imaging and Tomography - San Diego, United States Duration: Feb 22 2023 → Feb 23 2023

Publication series

Name	Progress in Biomedical Optics and Imaging - Proceedings of SPIE
Volume	12470

Conference

Conference	Medical Imaging 2023: Ultrasonic Imaging and Tomography
Country/Territory	United States
City	San Diego
Period	2/22/23 → 2/23/23

ASJC Scopus subject areas

Electronic, Optical and Magnetic Materials
Atomic and Molecular Physics, and Optics
Biomaterials
Radiology Nuclear Medicine and imaging

Keywords

Artificial intelligence (AI)
convolutional neural networks
deep learning
endometriosis

Access to Document

https://doi.org/10.1117/12.2653940

Cite this

Balica, A., Dai, J., Piiwaa, K., Qi, X., Green, A. N., Phillips, N., Egan, S., & Hacihaliloglu, I. (2023). Augmenting endometriosis analysis from ultrasound data with deep learning. In C. Boehm, & N. Bottenus (Eds.), Medical Imaging 2023: Ultrasonic Imaging and Tomography Article 124700O (Progress in Biomedical Optics and Imaging - Proceedings of SPIE; Vol. 12470). SPIE. https://doi.org/10.1117/12.2653940

@inproceedings{188dfcfe6c3142b5a75eef863ce0f0c8,

title = "Augmenting endometriosis analysis from ultrasound data with deep learning",

abstract = "Endometriosis is a non-malignant disorder that affects 176 million women globally. Diagnostic delays result in severe dysmenorrhea, dyspareunia, chronic pelvic pain, and infertility. Therefore, there is a significant need to diagnose patients at an early stage. Our objective in this work is to investigate the potential of deep learning methods to classify endometriosis from ultrasound data. Retrospective data from 100 subjects were collected at the Rutgers Robert Wood Johnson University Hospital (New Brunswick, NJ, USA). Endometriosis was diagnosed via laparoscopy or laparotomy. We designed and trained five different deep learning methods (Xception, Inception-V4, ResNet50, DenseNet, and EfficientNetB2) for the classification of endometriosis from ultrasound data. Using 5-fold cross-validation study we achieved an average area under the receiver operator curve (AUC) of 0.85 and 0.90 respectively for the two evaluation studies.",

keywords = "Artificial intelligence (AI), convolutional neural networks, deep learning, endometriosis",

author = "Adrian Balica and Jennifer Dai and Kayla Piiwaa and Xiao Qi and Green, {Ashlee N.} and Nancy Phillips and Susan Egan and Ilker Hacihaliloglu",

note = "Publisher Copyright: {\textcopyright} 2023 SPIE.; Medical Imaging 2023: Ultrasonic Imaging and Tomography ; Conference date: 22-02-2023 Through 23-02-2023",

year = "2023",

doi = "https://doi.org/10.1117/12.2653940",

language = "English (US)",

series = "Progress in Biomedical Optics and Imaging - Proceedings of SPIE",

publisher = "SPIE",

editor = "Christian Boehm and Nick Bottenus",

booktitle = "Medical Imaging 2023",

address = "United States",

}

Balica, A, Dai, J, Piiwaa, K, Qi, X, Green, AN, Phillips, N, Egan, S & Hacihaliloglu, I 2023, Augmenting endometriosis analysis from ultrasound data with deep learning. in C Boehm & N Bottenus (eds), Medical Imaging 2023: Ultrasonic Imaging and Tomography., 124700O, Progress in Biomedical Optics and Imaging - Proceedings of SPIE, vol. 12470, SPIE, Medical Imaging 2023: Ultrasonic Imaging and Tomography, San Diego, United States, 2/22/23. https://doi.org/10.1117/12.2653940

Augmenting endometriosis analysis from ultrasound data with deep learning. / Balica, Adrian; Dai, Jennifer; Piiwaa, Kayla et al.
Medical Imaging 2023: Ultrasonic Imaging and Tomography. ed. / Christian Boehm; Nick Bottenus. SPIE, 2023. 124700O (Progress in Biomedical Optics and Imaging - Proceedings of SPIE; Vol. 12470).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

TY - GEN

T1 - Augmenting endometriosis analysis from ultrasound data with deep learning

AU - Balica, Adrian

AU - Dai, Jennifer

AU - Piiwaa, Kayla

AU - Qi, Xiao

AU - Green, Ashlee N.

AU - Phillips, Nancy

AU - Egan, Susan

AU - Hacihaliloglu, Ilker

PY - 2023

Y1 - 2023

N2 - Endometriosis is a non-malignant disorder that affects 176 million women globally. Diagnostic delays result in severe dysmenorrhea, dyspareunia, chronic pelvic pain, and infertility. Therefore, there is a significant need to diagnose patients at an early stage. Our objective in this work is to investigate the potential of deep learning methods to classify endometriosis from ultrasound data. Retrospective data from 100 subjects were collected at the Rutgers Robert Wood Johnson University Hospital (New Brunswick, NJ, USA). Endometriosis was diagnosed via laparoscopy or laparotomy. We designed and trained five different deep learning methods (Xception, Inception-V4, ResNet50, DenseNet, and EfficientNetB2) for the classification of endometriosis from ultrasound data. Using 5-fold cross-validation study we achieved an average area under the receiver operator curve (AUC) of 0.85 and 0.90 respectively for the two evaluation studies.

AB - Endometriosis is a non-malignant disorder that affects 176 million women globally. Diagnostic delays result in severe dysmenorrhea, dyspareunia, chronic pelvic pain, and infertility. Therefore, there is a significant need to diagnose patients at an early stage. Our objective in this work is to investigate the potential of deep learning methods to classify endometriosis from ultrasound data. Retrospective data from 100 subjects were collected at the Rutgers Robert Wood Johnson University Hospital (New Brunswick, NJ, USA). Endometriosis was diagnosed via laparoscopy or laparotomy. We designed and trained five different deep learning methods (Xception, Inception-V4, ResNet50, DenseNet, and EfficientNetB2) for the classification of endometriosis from ultrasound data. Using 5-fold cross-validation study we achieved an average area under the receiver operator curve (AUC) of 0.85 and 0.90 respectively for the two evaluation studies.

KW - Artificial intelligence (AI)

KW - convolutional neural networks

KW - deep learning

KW - endometriosis

UR - http://www.scopus.com/inward/record.url?scp=85160803945&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85160803945&partnerID=8YFLogxK

U2 - https://doi.org/10.1117/12.2653940

DO - https://doi.org/10.1117/12.2653940

M3 - Conference contribution

T3 - Progress in Biomedical Optics and Imaging - Proceedings of SPIE

BT - Medical Imaging 2023

A2 - Boehm, Christian

A2 - Bottenus, Nick

PB - SPIE

T2 - Medical Imaging 2023: Ultrasonic Imaging and Tomography

Y2 - 22 February 2023 through 23 February 2023

ER -

Augmenting endometriosis analysis from ultrasound data with deep learning

Abstract

Publication series

Conference

ASJC Scopus subject areas

Keywords

Access to Document

Other files and links

Fingerprint

Cite this