A nomogram based on conventional ultrasound and elastography for diagnosing BI-RADS category 3-5 lesions.

OBJECTIVE

To develop a nomogram using conventional ultrasound and elastography to offer a more objective and accurate Breast Imaging Reporting and Data System (BI-RADS) classification, and minimize unnecessary biopsies.

METHODS

We retrospectively analyzed 689 BI-RADS category 3–5 lesions diagnosed by ultrasound. The cases were assigned to a training group and a validation group based on the timing of the ultrasound examination. The nomogram was constructed using multivariate logistic regression analysis and evaluated through Receiver Operating Characteristic curves, calibration curves, and decision curve analysis. Conventional BI-RADS 4a lesions were re-evaluated using the nomogram.

RESULTS

The nomogram incorporated 10 predictors: age, indistinct margin, angular margin, microlobulated margin, spiculated margin, hyperechoic halo, orientation, calcification, vascularity, and elasticity assessment. The area under the curve values for the nomogram in the training and validation group were 0.906 (95% confidence interval [CI]: 0.882–0.931) and 0.921(95%CI: 0.854–0.988), respectively. The calibration curves demonstrated good agreement between the nomogram predictions and actual malignancy outcomes. The decision curve analysis indicated that the nomogram had good clinical applicability. In the training group, 13% (41/304) of the conventional BI-RADS 4a lesions were downgraded to BI-RADS 3 by the nomogram, with only one malignancy (1/41, 2%). In the validation group, 17% (7/42) of conventional BI-RADS 4a lesions were downgraded through the nomogram, with no malignancies (0/7, 0%).

CONCLUSIONS

We developed a high-performing nomogram based on conventional ultrasound and elastography. The nomogram could provide individualized malignancy risk predictions for BI-RADS 3–5 lesions using fewer indicators, and help avoid unnecessary biopsies. CLINICAL TRIAL NUMBER: Not applicable.