Bayesian probability of malignancy with BI-RADS sonographic features

Area of Science:

• Radiology
• Medical Imaging
• Oncology

Background:

• Accurate assessment of breast mass malignancy is crucial for patient management.
• Combining imaging features can improve diagnostic performance.

Purpose of the Study:

• To develop a quantitative approach for combining American College of Radiology Breast Imaging Reporting and Data System (BI-RADS) sonographic features.
• To assess the overall probability of malignancy for breast masses.

Main Methods:

• Sonograms of solid breast masses were analyzed by two blinded observers.
• A naïve Bayes model integrated BI-RADS sonographic features with patient age and mammographic data.
• Diagnostic performance was evaluated using the area under the receiver operating curve (Az).

Main Results:

• BI-RADS sonographic features showed high predictive values.
• Integrating age and mammographic features improved diagnostic performance (Az increased from 0.772-0.884 to 0.866-0.924).
• Consensus diagnosis achieved high performance (Az, 0.954).

Conclusions:

• A naïve Bayes model offers a systematic method for combining sonographic features and patient data.
• This approach aids in differentiating malignant from benign breast masses.
• The model enhances the probability assessment of malignancy.