Breast Cancer Detection on Dual-View Sonography via Data-Centric Deep Learning

<italic>Goal:</italic> This study aims to enhance AI-assisted breast cancer diagnosis through dual-view sonography using a data-centric approach. <italic>Methods:</italic> We customize a DenseNet-based model on our exclusive dual-view breast ultrasound dataset to enhance the...

Full description

Saved in:
Bibliographic Details
Main Authors: Ting-Ruen Wei, Michele Hell, Aren Vierra, Ran Pang, Young Kang, Mahesh Patel, Yuling Yan
Format: Article
Language:English
Published: IEEE 2025-01-01
Series:IEEE Open Journal of Engineering in Medicine and Biology
Subjects:
Online Access:https://ieeexplore.ieee.org/document/10666269/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1846163763100647424
author Ting-Ruen Wei
Michele Hell
Aren Vierra
Ran Pang
Young Kang
Mahesh Patel
Yuling Yan
author_facet Ting-Ruen Wei
Michele Hell
Aren Vierra
Ran Pang
Young Kang
Mahesh Patel
Yuling Yan
author_sort Ting-Ruen Wei
collection DOAJ
description <italic>Goal:</italic> This study aims to enhance AI-assisted breast cancer diagnosis through dual-view sonography using a data-centric approach. <italic>Methods:</italic> We customize a DenseNet-based model on our exclusive dual-view breast ultrasound dataset to enhance the model&#x0027;s ability to differentiate between malignant and benign masses. Various assembly strategies are designed to integrate the dual views into the model input, contrasting with the use of single views alone, with a goal to maximize performance. Subsequently, we compare the model against the radiologist and quantify the improvement in key performance metrics. We further assess how the radiologist&#x0027;s diagnostic accuracy is enhanced with the assistance of the model. <italic>Results:</italic> Our experiments consistently found that optimal outcomes were achieved by using a channel-wise stacking approach incorporating both views, with one duplicated as the third channel. This configuration resulted in remarkable model performance with an area underthe receiver operating characteristic curve (AUC) of 0.9754, specificity of 0.96, and sensitivity of 0.9263, outperforming the radiologist by 50&#x0025; in specificity. With the model&#x0027;s guidance, the radiologist&#x0027;s performance improved across key metrics: accuracy by 17&#x0025;, precision by 26&#x0025;, and specificity by 29&#x0025;. <italic>Conclusions:</italic> Our customized model, withan optimal configuration for dual-view image input, surpassed both radiologists and existing model results in the literature. Integrating the model as a standalone tool or assistive aid for radiologists can greatly enhance specificity, reduce false positives, thereby minimizing unnecessary biopsies and alleviating radiologists&#x0027; workload.
format Article
id doaj-art-62e8de5f2cd04e038be3ce3d0be448eb
institution Kabale University
issn 2644-1276
language English
publishDate 2025-01-01
publisher IEEE
record_format Article
series IEEE Open Journal of Engineering in Medicine and Biology
spelling doaj-art-62e8de5f2cd04e038be3ce3d0be448eb2024-11-19T00:03:48ZengIEEEIEEE Open Journal of Engineering in Medicine and Biology2644-12762025-01-01610010610.1109/OJEMB.2024.345495810666269Breast Cancer Detection on Dual-View Sonography via Data-Centric Deep LearningTing-Ruen Wei0https://orcid.org/0009-0008-2649-4543Michele Hell1Aren Vierra2Ran Pang3Young Kang4Mahesh Patel5https://orcid.org/0000-0001-7385-3483Yuling Yan6https://orcid.org/0000-0002-8104-4624Santa Clara University, Santa Clara, CA, USASanta Clara University, Santa Clara, CA, USASanta Clara Valley Medical Center, San Jose, CA, USASanta Clara Valley Medical Center, San Jose, CA, USASanta Clara Valley Medical Center, San Jose, CA, USASanta Clara Valley Medical Center, San Jose, CA, USASanta Clara University, Santa Clara, CA, USA<italic>Goal:</italic> This study aims to enhance AI-assisted breast cancer diagnosis through dual-view sonography using a data-centric approach. <italic>Methods:</italic> We customize a DenseNet-based model on our exclusive dual-view breast ultrasound dataset to enhance the model&#x0027;s ability to differentiate between malignant and benign masses. Various assembly strategies are designed to integrate the dual views into the model input, contrasting with the use of single views alone, with a goal to maximize performance. Subsequently, we compare the model against the radiologist and quantify the improvement in key performance metrics. We further assess how the radiologist&#x0027;s diagnostic accuracy is enhanced with the assistance of the model. <italic>Results:</italic> Our experiments consistently found that optimal outcomes were achieved by using a channel-wise stacking approach incorporating both views, with one duplicated as the third channel. This configuration resulted in remarkable model performance with an area underthe receiver operating characteristic curve (AUC) of 0.9754, specificity of 0.96, and sensitivity of 0.9263, outperforming the radiologist by 50&#x0025; in specificity. With the model&#x0027;s guidance, the radiologist&#x0027;s performance improved across key metrics: accuracy by 17&#x0025;, precision by 26&#x0025;, and specificity by 29&#x0025;. <italic>Conclusions:</italic> Our customized model, withan optimal configuration for dual-view image input, surpassed both radiologists and existing model results in the literature. Integrating the model as a standalone tool or assistive aid for radiologists can greatly enhance specificity, reduce false positives, thereby minimizing unnecessary biopsies and alleviating radiologists&#x0027; workload.https://ieeexplore.ieee.org/document/10666269/Breast cancer classificationdeep learningradiologist comparison
spellingShingle Ting-Ruen Wei
Michele Hell
Aren Vierra
Ran Pang
Young Kang
Mahesh Patel
Yuling Yan
Breast Cancer Detection on Dual-View Sonography via Data-Centric Deep Learning
IEEE Open Journal of Engineering in Medicine and Biology
Breast cancer classification
deep learning
radiologist comparison
title Breast Cancer Detection on Dual-View Sonography via Data-Centric Deep Learning
title_full Breast Cancer Detection on Dual-View Sonography via Data-Centric Deep Learning
title_fullStr Breast Cancer Detection on Dual-View Sonography via Data-Centric Deep Learning
title_full_unstemmed Breast Cancer Detection on Dual-View Sonography via Data-Centric Deep Learning
title_short Breast Cancer Detection on Dual-View Sonography via Data-Centric Deep Learning
title_sort breast cancer detection on dual view sonography via data centric deep learning
topic Breast cancer classification
deep learning
radiologist comparison
url https://ieeexplore.ieee.org/document/10666269/
work_keys_str_mv AT tingruenwei breastcancerdetectionondualviewsonographyviadatacentricdeeplearning
AT michelehell breastcancerdetectionondualviewsonographyviadatacentricdeeplearning
AT arenvierra breastcancerdetectionondualviewsonographyviadatacentricdeeplearning
AT ranpang breastcancerdetectionondualviewsonographyviadatacentricdeeplearning
AT youngkang breastcancerdetectionondualviewsonographyviadatacentricdeeplearning
AT maheshpatel breastcancerdetectionondualviewsonographyviadatacentricdeeplearning
AT yulingyan breastcancerdetectionondualviewsonographyviadatacentricdeeplearning