The Role of Interval Nodes in Sentinel Lymph Node Mapping and Dissection for Melanoma Patients

Patients

Since 1997, all patients who had clinically and radiologically N0 skin melanoma with a Breslow thickness of more than 1 mm were prospectively included in a systematic SN biopsy program. Also included were patients whose melanoma had a Breslow thickness of 1 mm or less but was also Clark IV, ulcerated, or regressed. Patients who had a local recurrence of previous melanoma that had been operated on before 1997 entered the program but were excluded from the present analysis. Patients who had melanoma of the head and neck were also excluded from the present analysis, because of less predictable SN identification (5). All patients underwent abdominal and thoracic CT to detect any suggestive lesions.

SN Biopsy

In all but 1 patient, SNs were identified using the triple-localization technique with lymphoscintigraphy, blue dye, and γ-probe. One recently pregnant patient underwent blue-dye localization only and aborted after a positive SN was found. Lymphoscintigraphy was performed on the day before surgery. Depending on the size of the scar, 2–4 intradermal injections of 10 MBq of ^99mTc-nanocolloid in 0.1 mL were made around the excision/biopsy scar (^99mTc-labeled human serum albumin nanocolloid, particles < 80 nm, Nanocoll; Amersham Health). Dynamic imaging was then performed for 15 min (128 × 128 matrix, 60 frames, 15 s/frame) to identify the draining lymphatic vessels and to follow them until they reached the draining SN. The dynamic imaging was then followed by acquisition of static planar images (256 × 256 matrix, 5 min/image) over any lymph node field that could possibly drain the primary site. The imaging was completed by static transmission imaging using a ⁵⁷Co sheet (256 × 256 matrix, 2 min/image). Images were acquired with a single- or multihead γ-camera with low-energy high-resolution collimators. The surface location of the SNs was marked after localization with a hand-held γ-probe (Scintiprobe; Pol.Hi.Tech, then neo2000; Neoprobe Corp.) or an external ⁵⁷Co pen. On the day of surgery, a total of 2 mL (i.e., 4 × 0.5 mL) of blue dye (bleu patenté violet; Laboratoire Guerbet) were injected intradermally at 4 points around the scar or the primary. Surgery was directed by the same γ-probe, with systematic exploration of all possible basins. The SNs were dissected first, and then the scar was removed along with the usual safety margins.

Histopathology

The SNs, either fresh or in formaldehyde solution, were sent directly to the Department of Pathology. The lymph nodes were bivalved, embedded in paraffin, and sectioned for hematoxylin–eosin and immunohistochemistry staining (Melan A and protein S100), but no polymerase chain reaction analysis was performed. When an interval SN was discovered, the entire local lymph node group was directly removed, with the exception of the popliteal basin. Radical selective lymph node dissection was recommended for all patients with metastatic SNs. Patients were followed up with a clinical examination every 3–6 mo and with CT regularly.

Concept and Definition

Even with an incidence of only 4.5% in the present large series, interval SNs are a relevant problem. The terminology and definition of interval SNs must be discussed. A prerequisite is to accept that the usual classic basins comprise the cervical, axillary, inguinal, and iliac lymph node chains. Possible anatomically known but unusual basins include the subclavicular, internal mammary (retrosternal), and popliteal lymph nodes, for which radical lymph node dissection has been well described. In addition, one can find isolated or small groups of lymph nodes in rare but anatomically already-defined areas such as the epitrochlear area (6), humeral area, triangular intermuscular space (7), interpectoral area (4), intercostal areas (axillary line (8) and paravertebral (9,10)), paraaortic and retroperitoneal areas (4,11), lumbar area, calf, and thigh (12). The term interval node should include the last 2 groups and seems more appropriate than the terms unusual node, in-transit node, ectopic node, node outside lymph node basins, or intercalated nodes.

Incidence and Anatomy

Preoperative lymphoscintigraphy enhances lymphatic anatomy and definitively shows all potential SN areas (3). Lymphoscintigraphy can demonstrate multiple lymph drainage basins for melanomas located around the watershed areas (midline and umbilicus level) and, moreover, can review all other lymphatic drainage to disclose any interval SNs. The incidence of interval SNs varies widely, from 3.1% to 9.8% (Table 2) (11–19). We observed interval nodes in 4.5% of cases. The wide variation in interval SN incidence can be explained by differences in definition, in centers, in types of imaging tracers (20,21), in tracer injection (22,23), and in percentage of interval SN dissection. We deliberately chose to exclude otolaryngologic melanomas (separate analysis), for which large series (4,13) have found a 4.2%–6.8% incidence of interval SNs. An important point was further highlighted by Thelmo et al. (19). High-quality analysis requires acquisition of dynamic and static lymphoscintigraphy images in order to evaluate 2 types of lymphatic channels: serial and parallel (14,19,24). In a serial type of lymphatic channel, an interval SN could be a first-tier SN whereas the distal SN in the usual basin could be a second-tier SN. A parallel type of lymphatic channel would challenge the concept that the SN is the hottest, nearest node (25). Analysis of the counts per minute and the blue dye cannot differentiate a parallel channel from a unique channel, because of the resolution of the preoperative lymphoscintigraphy (19). Finally, lymphoscintigraphy may show hot spots that do not correspond to lymph nodes: Vidal-Sicart et al. found in 2 of 61 interval SNs hot spots that were actually 1 lymphatic lake and 1 lymphangioma (17). Similarly, Roozendaal et al. found in 3 of 25 interval SNs a hot spot that was a case of lymphangioma and hot spots in 2 other patients who had no lymph node tissue on histologic examination of the hot-spot area (15).

The incidence of interval SNs varies regarding melanoma location. We found that patients with melanomas on the trunk (5.9%) or upper limb (7.5%) had the highest proportion of interval SNs, whereas the incidence dropped in patients with lower-limb melanomas (1.4%). This finding had already been reported in the largest published series (Table 2). These reports recommended that investigators obtain systematic lymphoscintigraphic views of all expected and possible SN locations, as we did in the present study. Lumbar melanomas have some peculiarities, because their lymphatic drainage is highly unpredictable and they have the highest incidence of interval SNs—twice that found in other locations (4,14). We observed a rare anterior intercostal lymph node on the axillary line (patient 16, Fig. 4), a finding never observed by Weinberg (8). Another interval SN (patient 14, Fig. 5) was found in the superior lumbar triangle of Grynfeltt. This triangle is known for protrusions and lumbar hernias (26). Compared with the inferior lumbar triangle of Petit, the superior triangle can exhibit a vasculonervous bundle that can be accompanied by lymphatic vessels or nodes in the same way as is the triangular intermuscular space in the shoulder. This bundle may represent the entrance of lymph flow directed to retroperitoneal and paraaortic nodes and could provide a less invasive way to surgically access these nodes. Patient 18 had exclusive lymphatic drainage that paralleled the cephalic vein, with 1 humeral SN and 3 SNs before the cephalic vein plunged under the clavicle (Fig. 7). Such drainage has already been described by Thompson et al. (3) and Haagensen (27), but the absence of axillary drainage is notable.

Surgical Treatment

The main reasons to search for and remove detected interval SNs are, first, because the risk of metastasis (0%–22%) is similar to that of other SNs and, second, because the interval SN may be the only metastatic site, as was the case in 4 (7.4%) of 54 patients in one study (16), 5 (8.5%) of 59 in another (17), 3 (10%) of 30 in a third (19), 3 (14.3%) of 21 in a fourth (14), and 11 (17.7%) of 62 in a fifth (13). Suspected interval SNs should be resected, and Uren et al. speculated that some in-transit metastases could actually be undissected/undetected interval SNs (14). For example, Statius Muller et al. observed 44 recurrences in 248 patients after a median follow-up of 38 mo; among them, 2 were recurrences in interval SNs (28). The interval SN may also be the only lymph node drainage, as in patient 18.

We explored all hot spots regarded as interval SNs. Not all previously reported studies used this systematic surgical search, with some authors considering the procedure as excessive or dangerous (paraaortic, for example). Dissection of the internal mammary lymph node (including the SN) is a well-known technique in breast cancer surgery (29,30), and laparoscopic handheld γ-probes are available. Thus, it is possible that the incidence of interval SNs would have been lower if all interval SNs had been dissected, thereby allowing the exclusion of lymphatic lakes. In the paper published in 2000 by Uren et al., only 21 of 148 interval SNs were dissected (14). Our experience yielded hot spots in interval-SN locations in 18 patients. Patient 15 deserves special mention because the location of the interval-SN hot spot was a lymph channel and there was no classic or other SN. Anatomic and physiologic studies on lymphatic circulation show that the lymphatic stream may bypass a local obstacle by using parallel, previously nonfunctional channels and an inverse flow direction. Increased lymphatic flow or an obstacle may distend ducts (31), possibly explaining the evanescent hot spots seen in some patients (15,17). Because tumor deposits have never been reported for lymphatic lakes, they must be definitively different from interval SNs (14).

Further Management of Metastatic Interval SNs

Unusual locations must be recognized and distinguished from regional or distant metastatic disease (16). When a metastatic interval SN is detected, most investigators believe that definitive treatment should include a radical regional resection of surrounding soft tissues and lymph nodes. In our surgical experience, during the initial dissection of the interval SN we have directly removed all other surrounding lymph nodes to avoid difficulty with further dissections (of the triangular intermuscular space, for example), with the exception of popliteal interval SNs. The popliteal is the principal interval node in the lower limb, varying in incidence from 4.3% to 36% of cases (3,32,33), and radical popliteal lymph node dissection is a validated surgical technique (34). Whether to perform a radical selective lymph node dissection of the distal basin or another basin when the classic SN is not metastatic is still a matter of debate. The strategy of Sumner et al. is to perform a radical selective lymph node dissection of the regional lymph node basin upstream from the interval SN (16), and Roozendaal et al. believe that a metastatic interval SN may indicate involvement of the subsequent lymph node basin (15). In contrast, having confidence in the accuracy of SN biopsy to detect micrometastatic disease, McMasters et al. recommended that negative SN basins not undergo surgery (13). Regarding these recommendations, our opinion is that they increase the risk of undertreatment and that more work is needed on interval SNs before recommendations can be made.