Agreement Between an Automated Volume Breast Scanner and Handheld Ultrasound for Diagnostic Breast Examinations.

OBJECTIVES

To compare the agreement and interobserver variability of diagnostic handheld ultrasound (US) and a single volume on an automated breast volume scanner (ABVS) and to determine whether there was a significant difference if the ABVS was used by a sonographer or mammographic technologist.

METHODS

Ninety patients scheduled for diagnostic US examinations were randomized to either handheld US or the ABVS first. The AVBS was randomized between a sonographer and a mammographic technologist performing the study. The studies were blinded, randomized, and read by 2 radiologists. The lesion with the highest Breast Imaging Reporting and Data System (BI-RADS) score was used in the analysis. Final diagnoses were made by core biopsy or follow-up for 2 years. Lesions included 9 malignant and 81 benign.

RESULTS

The 90 patients had a mean age ± SD of 53.1 ± 16.3 years. The κ value for agreement between the ABVS and handheld US was 0.831 (95% confidence interval, 0.744-0.925), whereas the global agreement for a 7-point BI-RADS score was 0.488 (0.372-0.560). The agreement between the ABVS and handheld US was nearly the same when the ABVS was used by a mammographic technologist (κ = 0.858 [0.723-0.963]) or sonographer (κ = 0.803 [0.596-1.000]; P = .47). The areas under the receiver operating characteristic curves for characterization by the ABVS were 0.91 (0.84-0.96) for reader 1 and 0.91 (0.83-0.96) for reader 2; those for handheld US were 0.91 (0.84-0.96) for reader 1 and 0.83 (0.74-0.90) for reader 2, with no statistical difference. The agreement based on pathologic images was κ = 0.831 (0.718-0.944); for handheld US, κ = 0.795 (0.623-0.967); and for the AVBS, κ = 0.869 (0.725-1.000).

CONCLUSIONS

Performing a single-view diagnostic ABVS examination has good agreement with a handheld diagnostic US workup. There is no difference if the ABVS is used by a sonographer or mammographic technologist.