MRI Breast Lesion Analysis Using Deep Learning Techniques

Janhvi Chauhan; Dhaval Modi

MRI Breast Lesion Analysis Using Deep Learning Techniques

Author(s)	Janhvi Chauhan, Dhaval Modi
Country	India
Abstract	This study aimed to evaluate the effectiveness of a deep learning model in distinguishing between benign and malignant breast lesions using magnetic resonance imaging (MRI), while also characterizing various histological subtypes of these lesions. A deep learning model was developed to simultaneously detect and characterize breast lesions. The model was trained on single 2D T1-weighted fat-suppressed post-contrast MR images selected by radiologists, acquired following the administration of a gadolinium-based contrast agent. The dataset consisted of 335 MR images from 335 patients, encompassing 17 histological subtypes categorized into four groups: mammary gland tissue, benign lesions, invasive ductal carcinoma, and other malignant lesions. Model performance was evaluated on an independent test set of 168 MR images using weighted area under the ROC curve (AUC) metrics. The model achieved a cross-validation average ROC-AUC of 0.817 across a three-shuffle, three-fold setup. On the independent test set, it achieved a weighted mean AUC of 0.8. The findings demonstrate that a supervised attention-based deep learning model can effectively analyze breast MRI for lesion detection and classification. Further validation on larger and independent datasets is recommended to confirm its clinical applicability.
Keywords	Magnetic resonance imaging (MRI); Breast lesion detection; Convolution neural networks; Transfer learning; Attention model
Field	Engineering
Published In	Volume 12, Issue 4, October-December 2021
Published On	2021-12-03

View / Download PDF File

About IJSAT Fees & Payment Current Issue Publication Archive	Submit Research Paper Track Submission Status Publication Guidelines Publication Ethics Peer Review & Plagiarism	Join as a Reviewer Editors & Reviewers Reviewer Referral Program Get Reviewer Membership Certi.	Website/Journal Policies Usage Policy Content Policies Privacy Policy

Contact Us	Message on WhatsApp	+91-9687-182-185	editor@ijsat.org

International Journal on Science and Technology

A Widely Indexed Open Access Peer Reviewed Multidisciplinary Bi-monthly Scholarly International Journal

MRI Breast Lesion Analysis Using Deep Learning Techniques

Share this