Abstract
The aim of the work described here was to develop an ultrasound (US) image–based deep
learning model to reduce the rate of malignancy among breast lesions diagnosed as
category 4A of the Breast Imaging-Reporting and Data System (BI-RADS) during the pre-operative
US examination. A total of 479 breast lesions diagnosed as BI-RADS 4A in pre-operative
US examination were enrolled. There were 362 benign lesions and 117 malignant lesions
confirmed by postoperative pathology with a malignancy rate of 24.4%. US images were
collected from the database server. They were then randomly divided into training
and testing cohorts at a ratio of 4:1. To correctly classify malignant and benign
tumors diagnosed as BI-RADS 4A in US, four deep learning models, including MobileNet,
DenseNet121, Xception and Inception V3, were developed. The performance of deep learning
models was compared using the area under the receiver operating characteristic curve
(AUROC), accuracy, sensitivity, specificity, positive predictive value (PPV) and negative
predictive value (NPV). Meanwhile, the robustness of the models was evaluated by five-fold
cross-validation. Among the four models, the MobileNet model turned to be the optimal
model with the best performance in classifying benign and malignant lesions among
BI-RADS 4A breast lesions. The AUROC, accuracy, sensitivity, specificity, PPV and
NPV of the optimal model in the testing cohort were 0.897, 0.913, 0.926, 0.899, 0.958
and 0.784, respectively. About 14.4% of patients were expected to be upgraded to BI-RADS
4B in US with the assistance of the MobileNet model. The deep learning model MobileNet
can help to reduce the rate of malignancy among BI-RADS 4A breast lesions in pre-operative
US examinations, which is valuable to clinicians in tailoring treatment for suspicious
breast lesions identified on US.
Key Words
To read this article in full you will need to make a payment
Purchase one-time access:
Academic & Personal: 24 hour online accessCorporate R&D Professionals: 24 hour online accessOne-time access price info
- For academic or personal research use, select 'Academic and Personal'
- For corporate R&D use, select 'Corporate R&D Professionals'
Subscribe:
Subscribe to Ultrasound in Medicine and BiologyAlready a print subscriber? Claim online access
Already an online subscriber? Sign in
Register: Create an account
Institutional Access: Sign in to ScienceDirect
References
- Radiomics in breast cancer classification and prediction.Semin Cancer Biol. 2021; 72: 238-250
- Multimodality imaging of triple receptor-negative tumors with mammography, ultrasound, and MRI.AJR Am J Roentgenol. 2010; 194: 1160-1166
- Cancer statistics for the year 2020: An overview.Int J Cancer. 2021; (Published online April 5)https://doi.org/10.1002/ijc.33588
- Performance of machine learning software to classify breast lesions using BI-RADS radiomic features on ultrasound images.Eur Radiol Exp. 2019; 3: 34
- Applications of machine learning and deep learning to thyroid imaging: Where do we stand?.Ultrasonography. 2021; 40: 23-29
- Surgery-related shedding of breast cancer cells as determined by RT-PCR assay.J Surg Oncol. 2003; 82 (discussion 233): 228-232
- Complete breast autoamputation: Clinical image.Breast J. 2020; 26: 2265-2266
- Detection of micrometastatic disease and monitoring of perioperative tumor cell dissemination in primary operable breast cancer patients using real-time quantitative reverse transcription-PCR.Clinical Cancer Res. 2004; 10: 196-201
- Role of elastography for downgrading BI-RADS category 4a breast lesions according to risk factors.Acta Radiol. 2019; 60: 278-285
- Radiomics of US texture features in differential diagnosis between triple-negative breast cancer and fibroadenoma.Sci Rep. 2018; 8: 13546
- Combination of two-dimensional shear wave elastography with ultrasound breast imaging reporting and data system in the diagnosis of breast lesions: A new method to increase the diagnostic performance.Eur Radiol. 2016; 26: 3290-3300
- Less micrometastatic risk related to circulating tumor cells after endoscopic breast cancer surgery compared to open surgery.BMC Cancer. 2019; 19: 1070
- A review of breast density implications and breast cancer screening.Clin Breast Cancer. 2020; 20: 283-290
- The diagnostic value of MRI for architectural distortion categorized as BI-RADS category 3-4 by mammography.Gland Surg. 2020; 9: 1008-1018
- ACR BI-RADS Ultrasound. ACR BI-RADS Atlas, Breast Imaging Reporting and Data System.5th edition. American College of Radiology, Reston, VA2013
- BI-RADS update.Radiol Clin North Am. 2014; 52: 481-487
- Cancer diagnosis using deep learning: A bibliographic review.Cancers (Basel). 2019; 11: 1235
- Application of ultrasound artificial intelligence in the differential diagnosis between benign and malignant breast lesions of BI-RADS 4A.BMC Cancer. 2020; 20: 959
- Application of deep learning to predict underestimation in ductal carcinoma in situ of the breast with ultrasound.Ann Transl Med. 2021; 9: 295
- MobileNetV2: Inverted residuals and linear bottlenecks.in: 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. IEEE, Piscataway, NJ2018: 4510-4520
- Utilizing size-based thresholds of stiffness gradient to reclassify BI-RADS category 3-4b lesions increases diagnostic performance.Clin Radiol. 2019; 74: 306-313
- Classification of malignant tumours in breast ultrasound using unsupervised machine learning approaches.Sci Rep. 2021; 11: 1418
- Ultrasound positive predictive values by BI-RADS categories 3-5 for solid masses: An independent reader study.Eur Radiol. 2017; 27: 4307-4315
- Overview of radiomics in breast cancer diagnosis and prognostication.Breast. 2020; 49: 74-80
- Rapid review: Radiomics and breast cancer.Breast Cancer Res Treat. 2018; 169: 217-229
- Ultrasound elastography techniques in breast cancer.Radiologe. 2021; 61: 170-176
- Evaluating the accuracy of breast cancer and molecular subtype diagnosis by ultrasound image deep learning model.Front Oncol. 2021; 11623506
- Combination of different types of elastography in downgrading ultrasound Breast Imaging-Reporting and Data System category 4a breast lesions.Breast Cancer Res Treat. 2019; 174: 423-432
- Deep learning radiomics can predict axillary lymph node status in early-stage breast cancer.Nat Commun. 2020; 11: 1236
- Differential diagnosis of benign and malignant thyroid nodules using deep learning radiomics of thyroid ultrasound images.Eur J Radiol. 2020; 127108992
- Lymph node metastasis prediction from primary breast cancer US images using deep learning.Radiology. 2020; 294: 19-28
- Decoding the molecular subtypes of breast cancer seen on multimodal ultrasound images using an assembled convolutional neural network model: A prospective and multicentre study.eBioMedicine. 2021; 74103684
Article info
Publication history
Published online: August 30, 2022
Accepted:
June 24,
2022
Received in revised form:
May 31,
2022
Received:
March 31,
2022
Identification
Copyright
© 2022 World Federation for Ultrasound in Medicine & Biology. All rights reserved.
