Is the Berg Axillary Lymph Node Categorization Useful in the 3D Environment?

Abstract

The Berg muscle-based categorization of axillary lymph node location (commonly referred to as levels I, II, and III) was used extensively by pathologists and surgeons to describe the extent of axillary node dissection in breast cancer patients. However, its reproducibility with different arm positions and utility in 3-dimensional radiation treatment planning hasn't been tested. Computed tomography scans were observed in 16 patients in 2 positions: historical position (HP), ipsilateral arm abducted 90 degrees to the body axis; standard position (SP), arms above head. The volume, contents, and location of Berg lymph node levels (LNL) and the location of lymph nodes, surgical clips, pectoral muscles, and vascular structures relative to reference points were compared. From HP to SP there was no difference in LNL volumes. However, if measured from an anatomic landmark, the third thoracic vertebra (T3), LNL position varied: level I, an average of 23.1 mm anteriorly, P < 0.01; level II, 7.5 mm medially, P = 0.04; level III, 18.8 mm medially, P = 0.05. Using T3 as a reference: pectoralis major and minor muscles displaced medially (23.9 mm, P < 0.01 and 7.5 mm, P = 0.09) and anteriorly (18.2 mm, P < 0.01 and 11.2 mm, P < 0.01); axillary (18.0 mm, P < 0.01), subscapular (25.4 mm, P < 0.01), and lateral thoracic (8.4 mm, P < 0.01) vessels displaced anteriorly; axillary vessels displaced also medially (15.1 mm, P = 0.03). Disagreements in LN coverage with changes in arm position were observed in 60% (LNs) and 66% (clips) for level II. Surgeons, radiologists, and radiation oncologists should be aware that LNL coverage based on muscle boundaries varies significantly with arm position changes, making objective comparisons of information collected in different arm positions unreliable.