Evaluation of axillary ultrasound performance on Sentinel Node versus Observation after Axillary Ultrasound (SOUND) trial-eligible patients.

Abstract

The Sentinel Node versus Observation after Axillary Ultrasound (SOUND) trial suggested that sentinel lymph node biopsy could be omitted in small breast cancers with negative axillary ultrasound, despite 13.7% of preoperative axillary ultrasound being falsely negative when validated on sentinel lymph node biopsy. Our aim was to evaluate the performance of axillary ultrasound in our patient population using SOUND trial criteria. A retrospective review was performed between 2015 and 2023 of SOUND trial-eligible patients. Two subgroup univariate analyses, pathologic T classification group and molecular subgroup, were performed comparing preoperative axillary ultrasound performance metrics, demographics, and tumor characteristics. Multivariate analysis was performed to predict false-negative axillary ultrasound. 263 patients were SOUND trial-eligible, whereas only 223 had a sentinel lymph node biopsy. Overall, this study was similar to the SOUND trial in terms of demographics and tumor characteristics, as well as false-negative axillary ultrasound (13.5%). The remaining performance metrics of preoperative axillary ultrasound were sensitivity (23%), specificity (94%), positive predictive value (37%), and negative predictive value (88%). There were significant differences in specificity (P = .005) among molecular subgroups (Luminal/HER2-, HER2/neu+, and triple-negative). False-negative axillary ultrasound was significantly different (P value .015) when comparing pathologic T group (31% T2, 13% T1c, 11% T1b, and 0% T1mi/T1a). Multivariate analysis demonstrated that each 1-cm increase in tumor size was associated with 46% higher odds of a false-negative axillary ultrasound (odds ratio 1.46 per cm, P = .018). The high specificity and negative predictive value of axillary ultrasound suggests that foregoing sentinel lymph node biopsy in small breast cancers with negative preoperative AUS is reasonable, especially in patients with T1mi/T1a and T1b tumors.