An automated mammogram classification system using modified support vector machine

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Oncology

Background:

• Breast cancer is a significant public health concern, necessitating accurate and timely diagnosis.
• Manual mammogram interpretation by radiologists is time-consuming and prone to errors.
• Computer-aided diagnosis (CADx) systems offer a potential solution to improve diagnostic efficiency and accuracy.

Purpose of the Study:

• To develop and evaluate a CADx system for automated breast cancer detection and classification from mammograms.
• To assess the system's performance in distinguishing between normal, benign, and malignant mammographic findings.

Main Methods:

• Utilized mammograms from the Mammographic Image Analysis Society database.
• Extracted 15 textural features using gray level co-occurrence matrix at various angles and distances.
• Employed a two-stage support vector machine for classification of mammograms.

Main Results:

• Achieved 100% sensitivity and specificity in detecting any abnormality (normal vs. abnormal).
• Classified abnormalities with 94.4% sensitivity and 91.3% specificity for benign/malignant distinction.
• Demonstrated high positive predictive value (89.5%) and negative predictive value (95.5%).

Conclusions:

• Automated CADx systems show significant potential in assisting radiologists with breast cancer diagnosis.
• The proposed system can enhance the accuracy and efficiency of mammogram interpretation.
• CADx systems represent a valuable tool for improving breast cancer detection rates and patient outcomes.