Breast masses: computer-aided diagnosis with serial mammograms.

Area of Science:

• Radiology and Medical Imaging
• Artificial Intelligence in Healthcare
• Breast Cancer Diagnostics

Background:

• Accurate characterization of mammographic masses is crucial for timely breast cancer diagnosis.
• Computer-aided diagnosis (CAD) systems aim to assist radiologists in improving diagnostic accuracy.
• Interval change analysis, comparing prior and current mammograms, can provide valuable diagnostic information.

Purpose of the Study:

• To evaluate the impact of a CAD system employing interval change classification on radiologists' accuracy.
• To assess the CAD system's effect on differentiating malignant from benign breast masses on serial mammograms.
• To determine if CAD improves the classification of masses using the Breast Imaging Reporting and Data System (BI-RADS) categories.

Main Methods:

• Retrospective analysis of 90 temporal pairs of digitized two-view mammograms (47 malignant, 43 benign masses).
• Ten experienced readers (8 radiologists, 2 fellows) assessed masses with and without CAD assistance.
• Analysis used the Dorfman-Berbaum-Metz multireader multicase method to assess diagnostic performance metrics.

Main Results:

• The average area under the receiver operating characteristic curve (AUC) improved from 0.83 to 0.87 with CAD (P < .05).
• CAD showed a trend towards improving the partial area index above 0.90 sensitivity, though not statistically significant (P = .11).
• CAD influenced BI-RADS assessments, potentially improving correct biopsy recommendations for malignant and benign masses.

Conclusions:

• Computer-aided diagnosis incorporating interval change analysis significantly enhances radiologists' accuracy in mass classification.
• The CAD system aids in distinguishing malignant from benign masses on digitized screen-film mammograms.
• This technology shows promise for improving diagnostic decision-making in mammography interpretation.