Detection of Asymmetric Anomalies in Mammograms using Dynamic Time Warping for Early Breast Cancer Identification

Breast cancer is a leading cause of mortality, necessitating effective early detection techniques. This study investigates the use of Dynamic Time Warping (DTW) for assessing breast asymmetry in mammographic images to improve early cancer diagnosis. A clinical dataset of 115 patients with BIRADS cla...

Full description

Saved in:
Bibliographic Details
Main Authors: Toriz-Vazquez, Alfonso, Medina-Ramos, Luis Alberto, Bayareh-Mancilla, Rafael, Hernandez-Rodriguez, Yazmin Mariela, Eduardo Cigarroa-Mayorga, Oscar
Format: Conference Proceeding
Language:English
Subjects:
Asymmetry
Breast cancer
Correlation
Dynamic Time Warping
Electrical engineering computing
Machine learning
Mammography
Medical services
Statistical analysis
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Breast cancer is a leading cause of mortality, necessitating effective early detection techniques. This study investigates the use of Dynamic Time Warping (DTW) for assessing breast asymmetry in mammographic images to improve early cancer diagnosis. A clinical dataset of 115 patients with BIRADS classifications ranging from B1 to B5 was utilized. DTW was applied to measure differences between bilateral mammograms and quantify asymmetry. Significant variations in distances were observed across BI-RADS groups, with mean distances ranging from 1.78e+03 to 2.39e+03. This distinction indicates the effectiveness of DTW in capturing and quantifying asymmetrical patterns, particularly in cases categorized as BIRADS 1 and 2. Moreover, 36 cases with possible asymmetry were identified, demonstrating notable differences between the corresponding regions of contralateral breasts, predominantly in BI-RADS 3 to 5. Statistical analysis provided insights into the distribution and characteristics of asymmetry within each BIRADS group, underscoring the diagnostic value of BI-RADS in assessing breast asymmetry. Additionally, nipple retraction as an additional indicator enhanced the diagnostic evaluation, serving as an early indicator of breast cancer and strengthening the correlation between asymmetry and malignancy. In conclusion, DTW implementation offers a promising approach for early breast cancer detection through assessing asymmetry. Combining DTW distance analysis with additional indicators can enhance the diagnostic process, ultimately improving patient outcomes.
ISSN:2642-3766
DOI:10.1109/CCE60043.2023.10332854