ACR ULTRASOUND CRITERIA FOR THE DESCRIPTION OF SOLID BREAST LESIONS

Problemstellung: The American Collage of Radiology published a selection of ultrasound criteria for the description of breast lesions. This criteria are defined in detail and are illustrated in the recently published BI-RADS® Manual. The aim of our study was to evaluate the capability of these criteria to differentiate benign from malignant breast lesions.

Methoden: Between January 2004 and March 2005 solid breast lesions from 250 women were documented on the ACR form for the description of breast lesions. All lesions were confirmed by histological examination. The examinations were performed by four experienced physicians.

Ergebnisse: On histological examination we found 99 ductal, 22 lobular, 7 ductal in situ and 21 otherwise differentiated breast cancers (149 (59,6%) malignant lesions) as well as 40 fibromas and 61 mastopathic lesions (101 (40,4%) benign lesions). An ANOVA analysis revealed shape (oval or round vs. irregular, p<0,0001), orientation of the lesions axis (parallel to the skin vs. not parallel; p<0,0001), margins (circumscribed vs. not circumscribed and/or, indistinct, angular, microlobulated, spiculated; p=0,003), lesions boundary (abrupt interface vs. echogenic halo; p<0,0001) and vascularisation (not present, not assessed vs. present in lesion or adjacent to lesion) as significant criteria to distinguish between benign and malignant breast lesions. Echo pattern (hypoechoic, anechoic, hyperechoic or complex; p=0,892), posterior ultrasound features (shadowing, enhancement, combined patterns, no posterior features; p=0,117), surrounding tissue (duct changes, cooper's ligament changes, skin thickening, skin retraction, architectural distortion and edema; p=0,355) and calcifications (macro-calcification, micro-calcification in mass or out of mass; p=0,144) were not significant criteria in our analysis. The classification scheme scoring lesions from 2 (benign) to 5 (highly suspicious for malignancy) was also applied to the lesions and demonstrated in 95,1% a corrected classification.

Schlussfolgerungen: The description and classification scheme recommended by the ACR was a useful tool in our study. Nevertheless, only 5 out of 9 recommended criteria contributed significantly to the classification of breast lesions in benign and malignant tumors.