Meta-analysis: 2D Breast Ultrasound Radiomics Shows High Accuracy but Methodological Concerns

Pooled sensitivity 85% and specificity 87% (AUC 0.92) for radiomics-based differentiation of benign vs malignant breast lesions on 2D ultrasound, but likely upper-bound estimate due to high risk of bias and heterogeneity, per meta-analysis of 27 studies (9,627 lesions).

• Systematic review and meta-analysis of 27 observational studies using validation or test sets from 9,627 histopathologically confirmed breast lesions (5,733 benign, 3,894 malignant).
• Hierarchical summary ROC AUC was 0.92, with substantial between-study heterogeneity (I² 74.0–85.5%) and no evidence of publication bias (Deeks' p=0.505).
• QUADAS-2 assessment revealed frequent high or unclear risk of bias in patient selection and index test domains, indicating pooled estimates may represent upper-bound accuracy; prospective multicenter validation with standardized acquisition is needed.

RadPigeon summaries are original and for information only. They are not clinical advice.