DOI: 10.1002/jso.27664 ISSN: 0022-4790

Evaluating the role of level 3 axillary lymph node dissection in metastatic melanoma: Can we predict involvement?

Lily M. Parker, Kate E. Beekman, Danielle K. DePalo, Kelly M. Elleson, Jonathan S. Zager

Abstract

Background and Objectives

MSLT‐2 and DECOG‐SLT established that immediate complete axillary lymph node dissection (CLND) did not correlate with an increase in melanoma‐specific survival when compared with active ultrasound observation in patients with sentinel lymph node (SLN)‐positive disease. After those trials, there was a shift toward performing CLND only for clinically node‐positive disease. With these changes, we sought to determine the role of level III axillary lymph nodes in bulky disease and how the use of neoadjuvant therapy may impact the rate of positivity in level III axillary nodes.

Methods

We performed a retrospective chart review on all patients who underwent axillary CLND for cutaneous melanoma by one surgeon at an academic center from 2014 to 2022. These patients underwent CLND based on either having SLN+ disease or having clinically palpable or radiographically bulky disease.

Results

Of 95 patients included, there were 7 (7.3%) patients with level III positivity. One was SLN+ (1.0%), while 3 (3.1%) had bulky disease and neoadjuvant therapy, and 3 (3.1%) had bulky disease without neoadjuvant therapy. No preoperative factors were identified that predicted level III involvement. After performing CLND, the patients who had clinically palpable or radiographically bulky disease and neoadjuvant therapy had higher percent necrosis of nodes in levels I and II but not III. At 5 years, overall survival and recurrence‐free survival were improved in those without level III involvement (58% and 64%, respectively) when compared to those with level III involvement (41% and 50%), though this was not statistically significant.

Conclusions

Further study may identify better prognostic factors for level III positivity, allowing for the possibility of dissecting only levels I and II or even replacing CLND with targeted node dissections.

More from our Archive