Subcostal lymph nodes: An unusual sentinel lymph node basin in cutaneous melanoma

Abstract Background and Objectives Lymphatic drainage from subcostal nodes, along the costal groove, have not previously been characterized as sites for melanoma drainage and metastasis. This study reports a series of patients with subcostal nodes draining primary melanomas, with characterization of the sites of primary melanomas that drain to these nodes. Methods Patients who presented to our institution between 2005 and 2020 with documented cutaneous melanoma and sentinel lymph node biopsy of a subcostal node (sentinel = S), or metastases to subcostal nodes later in clinical management (recurrent = R) were included. Patient demographics, melanoma pathology, nodal features, imaging information, surgical approaches, and outcomes data were collected. Results Six patients had subcostal sentinel nodes (SNs). Primary sites included the posterior trunk and lateral chest wall. Subcostal nodes were found under ribs 10−12. Subcostal SNs had at least one dimension measuring 3 mm or less. There were no surgical complications related to removing the subcostal SN. Conclusions Melanoma can metastasize to subcostal lymph nodes and be found at the time of SN biopsy or identified at recurrence. These small nodes are fed by lymphatic channels that run in the neurovascular bundle under the ribs. When lymphatic mapping identifies a subcostal SN, it should be excised.


| INTRODUCTION
Sentinel lymph node biopsy (SLNB) is an essential procedure in the management and staging of malignant melanoma. 1,2 Preoperative lymphoscintigraphy is standard to determine lymphatic drainage and location of the sentinel lymph node(s). 3 Cervical, axillary, and inguinal lymph nodes (LN) are the most common locations for the sentinel node (SN), but aberrant and atypical sites of lymph node drainage have been described. Less commonly, SN are located in the epitrochlear, popliteal, triangular intramuscular space, internal mammary, retroperitoneal, or mediastinal nodes, and among others. [4][5][6][7] In aggregate, these sites of unusual drainage account for 5%−22% of SNs. 5,8,9 Those reports have helped to alert surgeons to the less typical locations, particularly the triangular intermuscular space nodes. 10 Costal margin LN are another uncommon site of lymph node drainage, first identified in a series of 10 patients with periumbilical primary lesions. Two of these patients "had a lymph channel passing over the right costal margin." 11 These nodes were further described as subcutaneous SN overlying the costal margin, with lymphatic drainage continuing cephalad to the internal mammary nodes. 4,11,12 In these reports of costal margin nodes, information was not provided about whether metastases were identified. To our knowledge and contrary to costal margin nodes, nodes under the rib in the costal groove have not previously been identified as potential sites of SNs nor further characterized in the literature. Management of these subcostal nodes, including the risk/benefit ratio of node resection versus the possible associated surgical complications, has not been defined.
The present report summarizes our institutional experience with identifying and managing subcostal sentinel LN in six patients. We present a single institutional experience series of LN found in the subcostal groove of lower ribs as SNs from the posterior trunk and flank. Importantly, metastatic disease was identified in 1/6 of these subcostal SNs. These data are presented including a discussion of appropriate patient management, and surgical considerations for safe removal of these nodes.

| MATERIALS AND METHODS
A retrospective review of a prospectively collected database was performed and supplemented with information collected from patient medical records, in accord with IRB # 10803. All patients with cutaneous melanoma evaluated at our institution between 2005 and 2020 were eligible. Patients with documented SLNB of a subcostal SN or with a clinical recurrence in the subcostal node were included. Patient demographics, melanoma pathology, nodal features, imaging information, surgical approaches, and outcomes data were collected. Sentinel lymph node size was extracted from the operative report unless unavailable, in which case the measurements were taken from the surgical pathology record.
Patients were examined in two subgroups: (1) patients who had a subcostal node identified at the time of SN biopsy for their original lesion without any recurrence (sentinel = S) and (2) patients with melanoma recurrence in a subcostal node (recurrent = R).
To access subcostal SNs in the operating room, the latissimus muscle is split by separating the muscle vertically in the direction of its fibers. Then, the fascial attachments along the undersurface of the rib are divided. The subcostal neurovascular bundle can be identified in the groove on the inner inferior aspect of the rib, known as the costal groove or sulcus costae. One may have to reach under the bottom edge of the rib to access the node. Thus, operating in this area poses a risk to the associated intercostal nerve, vein, and artery.
There is additional risk of either lung injury or pneumothorax. Thus, postoperative chest X-rays were obtained in each patient who underwent subcostal node excision at our institution.

| Patient population
Six patients with subcostal sentinel LN were identified and are referred to as S1−S6 (Table 1). Age ranged from 47 to 68 years. There were two females and four males; all were White.
One patient was identified who recurred in a subcostal site.

| Primary melanoma
All primary melanomas in patients S1-S6 were located on the flank or back. The average Breslow depth was 2.2 mm (range 1.2-3.2 mm).
Two primary lesions were ulcerated (S1 and S6, Table 1). These patients underwent wide local excision of their primary melanomas with 1-2 cm margins, per National Comprehensive Cancer Network (NCCN) guidelines.

| Preoperative imaging
All patients in the SN group (Patients S1−S6) underwent preoperative lymphoscintigraphy at our institution using Technetium 99 m-labeled sulfur radiocolloid injected intradermally at the site of the primary melanoma (example in Figure 1A). Blue dye was not used, in accord with our institutional practice. 13 One or more subcostal SN was detected on each preoperative lymphoscintigraphy, allowing subcostal SN to be found before or during the index surgery.
One patient (patient S5) also underwent single-photon emission computed tomography (SPECT/CT) at the time of initial lymphscintigraphy ( Figure 1B). In this case, the subcostal SN was not initially visualized on the SPECT/CT due to the proximity of the node to the injection site. However, after the subcostal node was identified intraoperatively after wide excision of the melanoma and injection site. The radiologist was able to appreciate the node on retrospective rereview of the SPECT/CT ( Figure 1B

| SN anatomy
All subcostal SNs were located in the costal groove of the 10th to 12th intercostal ribs (Table 1), along the posterior or posterior-lateral portion of those ribs. They were found anterior (more internal) and slightly superior to the bottom edge of the rib, along the lymphatic channels that run with the subcostal artery and vein ( Figure 1B−D, Figure 2). Subcostal SNs measured 5-8 mm in maximal dimension and 2-3 mm in minimum dimension; all had at least one dimension that was 3 mm or less (Table 1, Figure 1E,F).
Intraoperative count of the Tc 99 sulfur colloid radiotracer uptake for subcostal SN ranged from 1829 to 47 980 (Table 1). One subcostal SN from Patient S2 appeared grossly pigmented on intraoperative inspection ( Figure 1C,E) and contained metastatic melanoma on the pathology report.
In two cases in the SN group, presumed SNs were evident by imaging but were not removed because of difficulty exposing them safely. In both cases, they were deep to the paraspinal muscles (patient S1, S3; Table 1). Thus, the morbidity of possible thoracotomy required for node retrieval was thought to outweigh the benefit of resection. The SN pathology is therefore unknown. In both these cases, the nodes were surveilled with CT chest. Patient S1 remained clinically free of disease for 9 years, and then presented with diffuse bony metastases. Patient S3 continued to have negative surveillance CT scans ( Table 1).
As noted above, for patient R1, the location of the presumed subcostal node metastasis was under the tip of the right 12th rib.  Table 1). One patient had 3/3 positive SN at the subcostal location (S2), in addition to a positive axillary node. Four patients with negative subcostal SNs had positive nodes elsewhere (Patients S1, S4, and S6).

| Recurrence in the subcostal nodal basin
In this group of 7 patients, three (S1, S2, R1) developed recurrences during follow-up. Patient S1 recurred 9 years following index T A B L E 1 (Continued)  also had a positive axillary node. Patient R1 underwent routine surveillance PET/CT where recurrence at the right rib was detected, and as noted above, that patient also developed recurrences near the site of the rib resection, followed by distant metastatic disease.

| DISCUSSION
This series demonstrates that small LN can be found along the neurovascular bundle in the costal groove under ribs 10−12, that lower back melanomas can drain to these nodes, and that metastases to these nodes can be found, either at the time of SN biopsy, or as a recurrence if they are not removed earlier. Thus, they should be sought when treating patients with lower back melanomas. These nodes may be identified with the aid of lymphoscintigraphy, SPECT-CT scans, or possibly blue dye. Since the primary melanomas that drain to these sites may be near the subcostal regions, the injection site signal may obscure the hot spot at the subcostal SN. However, with a high level of suspicion they can be identified ( Figure 1A,B).  management, the fact that subcostal nodes can contain metastases (either at SLNB or at recurrence), and multiple patients presented with recurrent malignant melanoma at the subcostal site (Table 1), we recommend removal of these nodes at the time of SN biopsy.

| CONCLUSIONS
For primary melanomas of the lower back, there should be a high suspicion for drainage to subcostal nodes so that they may be identified intraoperatively. Even if not evident on initial lymphoscintigraphy, they may be found intraoperatively with the gamma probe after the primary melanoma is widely excised. Subcostal SNs can be challenging to locate due to their small size and location abutting or beneath the associated rib. When lymphatic mapping identifies a subcostal SN, it should be excised along with other identified SNs, as the removal of subcostal nodes may alter staging, treatment and prognosis and possibly prevent late recurrence whose resection is more morbid.

ACKNOWLEDGEMENT
This project was supported in part by the T32CA163177 grant (to KMM, CML, and CLC), and the University of Virginia Cancer Center Support Grant P30 CA044579.