Related Experiment Video
Updated: Oct 22, 2026
Ultrasonographic Evaluation of Breast Cancer-related Lymphedema
Published on: January 12, 2017
Shear-wave elastography combined with ultrasound BI-RADS improves diagnostic accuracy in breast lesion
Rinzing Ongmu Bhutia1, Tsella Lachungpa2, Mingma Sherpa3
1Department of Radiology, Sir Thutop Namgyal Memorial Hospital, Government of Sikkim, Gangtok, 737101, Sikkim, India.
Background:
To compare BI-RADS and shear-wave elastography (SWE) for differentiating benign from malignant breast lesions and assess whether adding maximum elasticity (Emax) improves discrimination and biopsy decisions.
Methods:
In this diagnostic accuracy study (May 2023-August 2024), 87 adults with 93 breast lesions underwent B-mode BI-RADS and SWE (Emax, Emean, Emin). ROC analysis and logistic regression incorporating BI-RADS and Emax were performed. Patient-level bootstrap resampling with 2,000 iterations provided optimism correction.
Results:
Eleven lesions (11.8%) were malignant. Median Emax and Emean were higher in malignant lesions (200.1 vs 105.7 kPa, P = 0.003; 103.1 vs 70.7 kPa, P = 0.004). AUCs were 0.766 for BI-RADS, 0.774 for Emax, 0.767 for Emean, and 0.697 for Emin; pairwise differences were nonsignificant (all P ≥ 0.153). The model had an AUC of 0.853 (95% CI, 0.754-0.951) and optimism-corrected AUC of 0.844. AUC gain was 0.087 (0.016-0.158; P = 0.017), with corrected gain of 0.077. At a 2% risk threshold, sensitivity was 100%, but biopsy recommendations increased from 41 to 55. The BE1 rule reduced biopsy recommendations by six, avoiding five benign biopsies but missing one carcinoma.
Conclusion:
Adding Emax to BI-RADS improved discrimination, with the gain persisting after validation. However, only 11 lesions were malignant, yielding imprecise estimates. No operating point reduced biopsy recommendations without sacrificing sensitivity. The model and SWE cut-offs require external validation.
Implication Of Practice:
SWE may be used as an adjunct to BI-RADS for characterization, but Emax thresholds and the combined model should not yet guide biopsy downgrading. Larger prospective studies with external validation are needed before implementation.
