Management of Malignant Melanoma of the Head and Neck Using Dynamic Lymphoscintigraphy and Gamma Probe–Guided Sentinel Lymph Node Biopsy FREE

THE SENTINEL lymph node (SLN) biopsy is revolutionizing the surgical management of primary malignant melanoma. It allows accurate nodal staging, and targets patients who may benefit from regional lymphadenectomy and systemic therapy. The SLN concept was originally proposed by Morton et al¹ and Wong et al² and is defined as the first node in the regional basin that receives lymphatic drainage from the primary tumor. The ability of the pathological evaluation of the SLN to predict the status of the entire nodal basin has been confirmed by multiple reports.^{3 - 6}

Morton et al¹ originally described injecting a vital blue dye into the dermis around the primary tumor. An incision is made over the draining nodal basin and dissection performed to identify a blue-stained lymphatic channel leading to the blue-stained SLN(s). This technique requires experience to achieve a high success rate. Krag et al⁶ and other researchers^{7 - 9} have described injecting technetium Tc 99m sulfur colloid around the primary tumor site and using a hand-held gamma probe to localize the SLN. Dynamic lymphoscintigraphy with radioactive colloid allows visualization of lymphatic channels from the injection site to the lymph nodes.^{4 ,10} This allows identification of true SLNs and distinguishes multiple SLNs from nonsentinel lymph nodes.

The lymphatic drainage of the head and neck is unpredictable, and performing an excisional lymph node biopsy in this area can be technically challenging. For these reasons, the use of SLN biopsy in the management of head and neck melanoma has not been widely accepted. This report describes our experience using dynamic lymphoscintigraphy and gamma probe–guided SLN biopsy in the treatment of melanoma.

Fifty-eight consecutive patients (47 male and 11 female) with clinical American Joint Committee on Cancer stages I and II malignant melanoma of the head and neck were treated using dynamic lymphoscintigraphy and gamma probe–guided SLN biopsy from January 1, 1994, to June 30, 1998. The mean patient age was 57.8 years (range, 11-80 years). Tumor characteristics are outlined in Table 1. The Breslow thickness ranged from 1.0 to 15.0 mm, with a mean tumor thickness of 3.82 mm. Additionally, 6 patients (10.3%) demonstrated ulcerated tumors. Statistical analysis was performed using the χ² test and the Fisher exact test.

All patients underwent cutaneous lymphoscintigraphy preoperatively using filtered technetium Tc 99m sulfur colloid (300-450 µCi; CIS-US Inc, Bedford, Mass). The radioactive tracer was injected intradermally around the circumference of the primary melanoma or biopsy site. Dynamic lymphoscintigraphy was performed with planar gamma camera imaging every 10 seconds for 10 minutes to identify focal areas of accumulation, followed by multiple 5-minute static images up to 60 minutes. In some patients, 2-hour postinjection delayed images were also obtained. A mark was placed on the skin overlying these areas to correlate with intraoperative localization. Lymphatic drainage areas in the head and neck were designated as parotid and anterior or posterior cervical triangles, depending on the location relative to the sternocleidomastoid muscle.

Measurements of radioactivity in the radiolabeled lymph nodes were made intraoperatively with a hand-held gamma probe (C-Trak; Care Wise Medical Products, Morgan Hill, Calif). This system provides an immediate and continuous variable pitch that varies directly with the incoming radiation. Sentinel lymph nodes demonstrated increased focal radiotracer uptake ("hot" spot). Counts were accumulated during a 10-second interval and recorded. Small incisions were made over areas of increased activity, which could be extended if a neck dissection would later be necessary. In the parotid region, preauricular incisions were made with skin flap elevation to expose the SLNs. Loupe magnification facilitated the dissection of individual facial nerve branches. Hot spots were removed with no attempt to dissect out individual lymph nodes. Ex vivo counts of the SLNs were obtained and compared with the nodal bed counts after removal. Vital blue dye injected at the time of surgery was used in only a few sporadic cases. All harvested SLNs were carefully labeled and, after serial sectioning, were examined histopathologically using routine hematoxylin-eosin and immunochemical staining for S100 protein and melanoma-associated antigen HMB 45. Frozen sections were not obtained. If the SLN contained tumor cells, a complete lymphadenectomy was performed at a later date.

At least 1 SLN was detected by lymphoscintigraphy in 57 patients. The radioactive colloid did not migrate from the injection site in 1 case early in the series. Drainage to 2 or more areas occurred in 14 patients (24.6%). The average number of nodal basins mapped per patient was 1.26. Lymphatic nodal basin drainage by general site of the primary melanoma is outlined in Table 2. Seventy-five drainage areas were mapped: anterior neck, 37; parotid gland, 18; posterior neck, 18; and axilla, 2.

Sentinel lymph nodes were harvested from 72 (96%) of 75 mapped nodal basins. Three areas of high radioactivity in the parotid gland could not be isolated without risking injury to the facial nerve. One of these patients had separate drainage to the anterior neck, which contained metastatic disease. A complete neck dissection revealed a metastatic node in the parotid gland. The mean ratio of ex vivo SLN radioactive counts to nodal bed counts was 20.4 (Table 3). An average of 2.68 SLNs was harvested per patient with a mean of 2.13 SLNs per nodal basin. Sentinel lymph nodes positive for metastatic melanoma were identified in 10 patients (17.5%). Positivity by tumor thickness is detailed in Table 4.

Findings of serial sectioning and routine hematoxylin-eosin staining detected micrometastatic disease in 9 (90%) of 10 patients with positive SLNs. Immunohistochemical analysis (S100 and HMB 45) alone detected disease in 1 SLN (10%).

Ten patients underwent a therapeutic lymph node dissection. The results of complete lymph node dissections are given in Table 5. No patient with a tumor thickness less than 4.5 mm had residual positive lymph nodes at the time of complete dissection.

The mean follow-up was 15.9 months (range, 1-57 months). Analysis of disease recurrences by selected variables is given in Table 6. Relapse occurred in 10 (21.3%) of the 47 patients with negative SLN biopsy results and 7 (70%) of the 10 patients with positive SLN biopsy results (P = .005, Fisher exact test). Local recurrences occurred in 5 patients (SLN-positive, 3; SLN-negative, 2), and isolated in-transit recurrence in 2 (SLN-positive, 1; SLN-negative, 1). Regional nodal recurrence occurred in 3 patients (30%) after a positive SLN biopsy finding. One patient experienced recurrence in the opposite neck, and the other 2 patients had recurrences in regional beds as well as distant recurrence. One patient (2.1%) had an in-transit and regional recurrence after a negative SLN biopsy finding. Neither of the 2 patients whose parotid gland SLN could not be isolated developed a recurrence in the area. Six patients developed isolated distant metastases after negative SLN biopsy results (12.8%). The patient in whom SLN drainage could not be demonstrated on lymphoscintigraphy developed distant metastases.

Ten of the 22 patients with T4 tumors (>4 mm) experienced recurrence. Five (83.3%) of the 6 patients who were SLN positive had recurrence. These recurrences included 2 local, 1 regional, and 2 regional/distant. Five (31.3%) of the 16 patients who were SLN negative developed recurrences. Sites of recurrence included 1 local, 1 in-transit/regional, and 3 distant.

The lymphatic drainage patterns of the head and neck are multiple, varied, and unpredictable.^{11 - 13} O'Brien et al¹⁴ described 97 patients with head and neck melanoma who underwent lymphoscintigraphy. They found a high rate (34%) of disagreement between clinically predicted lymphatic drainage pathways and the pathways found on the basis of lymphoscintigraphy. Dynamic lymphoscintigraphy allows identification of in-transit and "second-echelon" lymph nodes like the occipital or facial lymph nodes, which potentially could be missed (Figure 1). O'Brien et al¹⁴ noted that in 22% of the cases they studied, the SLNs were identified outside the parotid and 5 main neck levels.

Wells et al¹⁵ reviewed their experience with preoperative lymphoscintigraphy, intraoperative blue dye localization, and a handheld gamma probe in 58 consecutive patients with head and neck melanoma. The mean tumor thickness was 2.21 mm; SLNs were successfully identified in 55 patients (95%), and in 6 patients (11%), the SLNs were found to contain micrometastatic disease. Completion lymph node dissection findings showed that the SLN was the only site of metastases in all 6 patients. Alex et al¹⁶ recently reported their experience with lymphatic mapping in 23 patients with head and neck melanoma (T2 and T3) using blue dye and a handheld gamma probe. Sentinel lymph nodes were successfully resected in 22 patients (96%); 3 patients (13%) had SLNs positive for melanoma, and only 1 patient with SLNs negative for melanoma had a regional recurrence. Complete lymph node dissection revealed residual nodal involvement in 1 patient. Bostick et al,¹⁷ from the John Wayne Cancer Institute, reported intraoperative lymphatic mapping in 117 patients with head and neck melanoma. Preoperative lymphoscintigraphy allowed identification of 12 patients (10%) with multiple lymphatic drainage basins. Use of blue dye alone identified SLNs in 93 (92%) of 101 basins. With the use of the intraoperative gamma probe combined with blue dye, SLNs were identified in 27 (96%) of 28 nodal basins. Overall, 14 patients (12%) had occult metastases in SLNs.

The average number of SLNs harvested per nodal basin in this study was somewhat higher than that in previous reports. This number has been reported from 1.44 to 1.74 SLNs per basin, using blue dye and colloid,^{10 ,18 - 20} and 1.1 to 1.25 using blue dye alone.^{1 ,20} This may be the result of variation in surgical technique to locate all SLNs. Our technique has been to remove the hot spot located with the gamma probe. No attempt in vivo or ex vivo is made to dissect out individual nodes from the tissue removed. Dissection of the individual nodes using the gamma probe and the presence of blue coloration could potentially reduce the amount of pathological examination.

The overall SLN positivity in this series is similar to previous reports, considering that in almost 40% of cases the primary tumor thickness was greater than 4.0 mm. The rate of SLN positivity correlated with tumor thickness (Table 4). Complete lymph node dissection after positive SLN biopsy findings yielded residual disease in 30% of cases, which is similar to previous reports.^{15 - 16 ,18} In this series, no residual disease was found in any patients with a tumor thickness less than 4.5 mm.

Recurrence in a nodal bed after treatment of a negative SLN is an infrequent event. Kapteijn et al¹⁹ found a 3.2% false-negative rate after 93 negative SLN biopsy results.¹⁹ Gershenwald et al²⁰ reported recurrence in a mapped nodal basin after a previous negative SLN biopsy finding in 10 (4.1%) of 243 patients.²⁰ Eight of the 10 patients were found to have occult nodal metastases when the initial SLN specimens were serially sectioned and immunohistochemically stained for S100 and HMB 45. The overall false-negative rate was actually 0.8%. There was only 1 regional recurrence after a negative SLN biopsy finding in this series. This probably resulted from secondary spread to the nodal basin from an in-transit recurrence.

Drainage to the parotid region occurred in 24% of cases in this series. The performance of biopsies of SLNs in this area may present special problems. Sentinel lymph nodes in the parotid gland tend to be small and difficult to find. Dissection without adequate exposure may put the facial nerve at risk. Ollila et al²¹ reviewed their experience of parotid SLN mapping in 39 patients. Sentinel lymph nodes were successfully identified in 37 patients (94.9%). One case of temporary facial nerve paresis was reported. In this series, localization of parotid SLNs was unsuccessful in 3 (16.7%) of 18 cases. No facial nerve injuries occurred, but the high radioactive counts in the parotid gland made localizing focal activity difficult (Table 3). The addition of blue dye would be useful to provide visual information and overcome some of the limitations of high background activity in the parotid area.

Melanomas thicker than 4 mm are classified as stage III by the American Joint Committee on Cancer. Traditionally, patients with thick melanomas (T4) have not been candidates for elective lymph node dissection because of the high incidence of distant metastases. Heaton et al²² described 278 patients with T4 melanomas (median thickness, 6.0 mm). Nodal status, tumor thickness, and ulceration were all associated with survival by multivariate analysis. The SLN biopsy can be useful in these patients to provide prognostic information as well as a potential therapeutic benefit. In this study, SLN status was predictive of recurrence for T4 melanomas (83% for positive SLNs and 31% for negative SLNs) (P = .06 Fisher exact test).

Gamma probe–guided SLN localization in the head and neck region was successful in 96% of draining nodal basins. It can target regional lymphadenectomy in patients who may benefit from regional nodal dissection. The use of isosulfan blue dye in conjunction with radioactive colloid may be useful in the parotid nodal basin. Thick melanomas (>4 mm) should be eligible for SLN biopsy as nodal status is prognostic in this group.

Accepted for publication December 21, 1999.

Presented at the annual meeting of the American Head and Neck Society, Palm Desert, Calif, April 24-27, 1999.

Reprints: Grant W. Carlson, MD, The Emory Clinic, 1365B Clifton Rd NE, Atlanta, GA 30322 (e-mail: grant_carlson@emory.org).