Axillary Ultrasound of Patients With and Without Breast Cancer: Sonographic Predictors of Malignancy

Considerable overlap exists between benign and malignant sonographic axillary nodal features, especially in patients without a known breast cancer (BC) diagnosis. Biopsy criteria, including cortical thickness cut- offs for patients without BC (woBC), are not defined. In this study, we aim to compare the malignancy risk of sonographic features of abnormal axillary lymph nodes in patients with and without a concurrent BC diagnosis.

Upon institutional review board approval, women who underwent ultrasound (US) guided axillary lymph node fine needle aspiration or core needle biopsy in community imaging sites from January 1, 2015 to October 26, 2021 were identified. Patients with a history of recent COVID-19 vaccination, non-nodal findings, nodal calcifications as lone finding, or no available reports of associated diagnostic study or subsequent pathology were excluded. Patients were categorized based on presence of concurrent BC diagnosis; sonographic and pathologic data were then collected. All statistical modeling was conducted using generalized linear modeling using a binary distribution.

Final analysis included 86 women woBC and 106 women with breast cancer (wBC). For woBC patients, the odds of malignancy increased 3-fold [odds ratio (OR): 4.1, 95% confidence interval (CI) (1.501, 11.055)] when an absent hilum was noted on US relative to preserved hilum; for wBC patients, the odds increased almost 27-fold [OR: 27.9 (3.573, 218.201)], p<0.001. For woBC patients, the odds of malignancy increased 19% [OR: 1.19 (0.94, 1.52)] for every unit increase in cortical thickness (mm); for wBC patients, the odds increased 73% [OR: 1.7 (1.23, 2.43)] for every unit increase, p<0.001. For woBC patients, the odds of malignancy increased 150% [OR: 2.5 (1.3, 5.1)] for every cm increase in the short axis dimension; for wBC patients, the odds increased 300% [OR: 20.95 (4.1, 105.99)] for every unit increase, p<0.001. A 3 mm cortical thickness in wBC patients had a 30% likelihood of malignancy; an 11 mm cortical thickness in woBC patients had a 30% likelihood of malignancy.

In all patients, sonographic axillary nodal features of absent hilum, increased nodal size, and cortical thickness were found to predict malignancy; however, malignancy rates were much higher in the wBC group than in the woBC group for the same imaging findings. These results suggest it may be reasonable to use higher cut-offs for criteria such as cortical thickness when determining need for biopsy in wBC vs woBC patients.