Interobserver variability in upgraded and non-upgraded BI-RADS 3 lesions.

Area of Science:

• Radiology
• Oncology
• Medical Imaging

Background:

• Breast Imaging-Reporting and Data System (BI-RADS) 3 lesions are considered benign but require follow-up.
• Interobserver variability in mammographic interpretation can affect patient management.
• Accurate assessment of BI-RADS 3 lesions is crucial to avoid unnecessary biopsies or delayed diagnosis.

Purpose of the Study:

• To quantify interobserver variability in the assessment of Breast Imaging-Reporting and Data System (BI-RADS) 3 mammographic lesions.
• To determine the appropriateness of initial evaluations for BI-RADS 3 lesions that were subsequently upgraded to BI-RADS 4/5.

Main Methods:

• Retrospective review of 1,188 screen-detected BI-RADS 3 mammographic lesions.
• Comparison of 60 upgraded BI-RADS 4/5 lesions (cases) with 60 matched non-upgraded BI-RADS 3 lesions (controls).
• Independent blinded assessment of available mammographic studies by two radiologists.

Main Results:

• Significant interobserver variability was observed in the classification of BI-RADS 3 lesions, with assessments ranging from BI-RADS 0 to BI-RADS 4.
• For malignant cancers within the study group, radiologists' assessments varied considerably (BI-RADS 0-4).
• Reasons for BI-RADS 0 included incomplete views or lack of correlation, while BI-RADS 4 was due to suspicious morphology or instability.

Conclusions:

• Substantial interobserver variability exists in the assessment of BI-RADS 3 mammographic lesions.
• A notable proportion of BI-RADS 3 lesions were considered incompletely evaluated during blinded reassessment, suggesting potential inconsistencies in initial management.